Insecure Mag 41

Big Data and cloud security have been some of the most discussed topics during the past
year,
and we decided it was time to dedicate an entire issue to them. We have a great lineup of industry
leaders sharing their knowledge and exploring various areas.
As youre reading this, Im about to start my yearly trip to San Francisco for the madness known
as RSA Conference, without a doubt the most signicant information security event of the year.
Im looking forward to meeting many of you, discovering new companies and seeing innovative
technologies in action. A special issue of (IN)SECURE Magazine after the event will showcase the
best of what the show had to offer and put a spotlight on several companies. See you next week
in San Francisco!
Mirko Zorz
Editor in Chief
Visit the magazine website at www.insecuremag.com
(IN)SECURE Magazine contacts
Feedback and contributions: Mirko Zorz, Editor in Chief - [email protected]
News: Zeljka Zorz, Managing Editor - [email protected]
Marketing: Berislav Kucan, Director of Operations - [email protected]
Distribution
(IN)SECURE Magazine can be freely distributed in the form of the original, non-modied PDF
document. Distribution of modied versions of (IN)SECURE Magazine content is prohibited
without the explicit permission from the editor.
Copyright (IN)SECURE Magazine 2014.
www.insecuremag.com
Most organizations are unable to
resolve a cyber attack
The lack of incident detection
and investigation puts
companies and their CISOs'
jobs at signicant risk,
according to a new Ponemon
Institute study.
In fact, when a CEO and Board of Directors
ask a security team for a brieng immediately
following an incident, 65% of respondents
believe that the brieng would be purposefully
modied, ltered or watered down.
Additionally, 78% of respondents believe most
CISOs would make a best effort guess
based on limited information, and they would
also take action prematurely and report that
the problem had been resolved without this
actually being the case.
This disconnect results from several critical
shortcomings in the current point solution
approach to cybersecurity and incident
response (IR), namely:
Lack of timely compromise detection: 86%
of respondents say detection of a cyber-attack
takes too long
Inability of point solutions to prioritize alerts
as they come in: 85% say they suffer from a
lack of prioritization of incidents
Lack of integration between point solutions:
74% say poor or no integration between
security products negatively affects response
capabilities
An overwhelming number of alerts
paralyzing IR efforts: 61% say too many alerts
from too many point solutions also hinders
investigations.
When a cyber attack happens, immediate
reaction is needed in the minutes that follow,
not hours or days, said Dr. Larry Ponemon,
chairman and founder of the Ponemon
Institute. Its readily clear from the survey that
IR processes need to incorporate powerful,
intuitive technology that helps teams act
quickly, effectively and with key evidence so
their companies and clients time, resources
and money are not lost in the immediate
aftermath of the event.
www.insecuremag.com 5
The sad state of cybersecurity
readiness
Just 17 percent of UK business leaders see
cybersecurity as a major priority, compared to
41 percent in the US, research from BT has
revealed.
The research, which assessed attitudes to
cybersecurity and levels of preparedness
among IT decision makers, highlights that UK
businesses are lagging behind their US
counterparts in crucial areas. Just one in ve
(21 percent) respondents in the UK are able to
measure the ROI of their cybersecurity
measures compared to nine in ten (90
percent) US companies. Similarly, 86 percent
of US directors and senior decision makers
are given IT security training, compared to just
37 percent in the UK.
More than half (58 percent) of IT decision-
makers globally stated that their boards
underestimate the importance of
cybersecurity. This gure increases to 74
percent in the US but drops to 55 percent in
the UK. The difference in levels of
preparedness correlates with attitudes to
threats. Non-malicious insider threats (e.g.
accidental loss of data) are currently the most
commonly cited security concern globally,
being reported as a serious threat by 65
percent of IT decision makers. In the UK this
falls to 60 percent and is followed by malicious
insider threats (51 percent), hacktivism (37
percent) organized crime (32 percent), nation
states (15 percent) and terrorism (12 percent).
In the US the proportion of IT decision makers
who see non-malicious insider threats as a
severe threat increases to 85 percent and is
followed by malicious insider threats (79
percent), hacktivism (77 percent), organized
crime (75 percent), terrorism (72 percent) and
nation states (70 percent).
400Gbps NTP-based DDoS attack
hits Cloudare
Matthew Prince, CEO of content delivery
network Cloudare, has conrmed that one of
its customers has been targeted with a very
big Network Time Protocol (NTP) reection
attack - "bigger that the Spamhaus attack
from last year."
He didn't name the customer, but he has
shared that the attack reached the level of
over 400 gigabits per second, that it probably
caused congestion on some peering
exchanges (mostly in Europe), that (based on
sampled data) it misused just over 4,500
miscongured NTP servers, and that the
customer initially wanted to pay with a stolen
credit card.
Despite the recommendation issued by US-
CERT about updating public-facing NTP
servers to a ntpd version that doesn't allow
attackers to use them for NTP amplication
attacks, there are still many vulnerable ones
out there.
"The attack relies on the exploitation of the
'monlist' feature of NTP, as described in
CVE-2013-5211, which is enabled by default
on older NTP-capable devices. This command
causes a list of the last 600 IP addresses
which connected to the NTP server to be sent
to the victim," explains US-CERT.
Intrinsic-ID enhances its Saturnus
cloud security solution
Intrinsic-ID has released a new version of
Saturnus (www.intrinsic-id.com/saturnus), its
device-unique cloud security solution, that
gives users total control over the protection of
their data. The new version includes
enhanced usability features that make using
the cloud safer without compromising
exibility, performance or ease of use. As part
of the launch of this new version, the company
is offering a money back guarantee.
Saturnus is a hardware/software solution
based on proprietary and patented Hardware
Intrinsic Security (HIS) technology. The USB
token contains a smart-card chip embedded
with HIS, the strongest technology to generate
and store security keys. On top of this, the
Saturnus software application runs on the
client device that will process the data. While
working in the cloud, the Saturnus token is
connected to the USB port of a device. The
hardware-based security is augmented by a
second step, a username and password-
based login system. This combination is called
two-factor authentication and is based on two
unrelated factors: something known to the
user (username and password) and
something the user has (the hardware token).
The result of two-factor authentication based
on HIS is incredibly strong protection for all
data in the cloud.
The USB token provides security based at the
client site. When the USB token is connected,
les are encrypted before they leave the
device on the way to the cloud. This
encryption is performed by using symmetric
key cryptography, which means les are
encrypted and decrypted using the same key.
Since encryption and decryption are only
performed on the client side and within the
hardware of the token, the key never leaves
the user and is therefore completely secure
from malicious use.
The GUI enables an intuitive use of the
application. All functions are visible on the
application screen and les can be dragged
into the secure Saturnus environment.
Infosecurity Europe to feature over
350 exhibitors
Infosecurity Europe (bit.ly/infosec-2014) is
Europes number one Information Security
event. Featuring over 350 exhibitors, the most
diverse range of new products and services,
an education program and over 12,000
visitors from every segment of the industry, its
the most important date in the calendar for
infosec professionals across Europe.
Take the chance to hear about new and
existing products, services and solutions as
exhibiting companies take to the stage to
demonstrate the capabilities of their
information security solutions. Pose your
questions directly to the solution providers and
nd the answers youve been looking for.
The Business Strategy Theatre seminars
feature case studies and best practices for
addressing the challenges and issues facing
management, CEOs and other board level
directors. Benet from the opportunity to:
Discover how to tackle the key business
challenges and issues impacting how an
organization protects itself against the latest
threats.
Gain rst-hand experience from vendors and
end-users, sharing practice experience and
real life learning.
Access learning that can be applied directly
to your business.
IE 0-day used in watering hole attack
tied to previous campaigns
An Internet Explorer zero-day vulnerability
(CVE-2014-0322) is actively exploited in the
wild in a watering-hole attack targeting visitors
to the ofcial website of the U.S. Veterans of
Foreign Wars, FireEye researchers warned.
"Its a brand new zero-day that targets IE 10
users visiting the compromised website a
classic drive-by download attack. Upon
successful exploitation, this zero-day attack
will download a XOR encoded payload from a
remote server, decode and execute it," they
explained.
"We believe the attack is a strategic Web
compromise targeting American military
personnel amid a paralyzing snowstorm at the
U.S. Capitol in the days leading up to the
Presidents Day holiday weekend. Based on
infrastructure overlaps and tradecraft
similarities, we believe the actors behind this
campaign are associated with two previously
identied campaigns (Operation DeputyDog
and Operation Ephemeral Hydra)," they
added.
This new campaign has been dubbed
"Operation SnowMan," and the similarities
with the aforementioned earlier campaigns are
many: exploitation of an IE zero-day, delivery
of remote access Trojan (Gh0st RAT),
"watering hole" exploit delivery method,
related C&C infrastructure, the use of a simple
single-byte XOR encoded (0!95) payload
obfuscated with a .jpg extension.
"The exploit targets IE 10 with Adobe Flash. It
aborts exploitation if the user is browsing with
a different version of IE or has installed
Microsofts Experience Mitigation Toolkit
(EMET)," they shared, and pointed out that
installing EMET or updating to IE 11 are
perfect mitigation measures.
It is believed that the same actors have likely
orchestrated all these campaigns. So far, the
targets were US government agencies,
defense companies, IT and law rms, NGOs,
mining companies, so it's safe to say they
were cyber espionage campaigns aimed at
stealing condential information.
Websense researchers say they have
discovered the use of this same vulnerability
as early as January 20, 2014 (FireEye
detected the exploit on February 11), and that
the targets were the visitors to a fake site
mimicking that of the French aerospace
association GIFAS, which includes contractors
and rms in both the military and civilian
aircraft industry.
Google offers ve grants to women
in security to attend HITB2014AMS
Google is offering ve grants
to women in security to attend
the Hack In The Box
Amsterdam conference in
May.
These grants include a VIP
ticket to the conference on the
29th and 30th of May, an exclusive invite to
the HITBSecConf Speakers Reception on the
28th, an invite to the Girl Geek Dinner
Amsterdam on the 29th and an invite to the
HITB Post Conference Reception sponsored
by Microsoft on the evening of the 30th.
Winners of the grant will also receive up to
1000 EUR towards travel costs (to be paid
after the conference).
To nd out more about the Google Women in
Tech Travel and Conference Grant program,
see here - www.google.com/edu/students/
women-in-tech-conference-and-travel-grants/
As crimeware evolves, phishing
attacks increase
The number of phishing campaigns increased
by more than 20 percent in the third quarter of
2013, with crimeware attacks evolving and
proliferating, according to the APWG.
The total number of unique phishing websites
observed rose to 143,353 in Q3, up from Q2s
119,101. The increase is generally attributable
to rising numbers of attacks against money-
transfer and retail/e-commerce websites.
During the same period, there was an 8
percent decline in the number of brands
targeted by phishers, as the number of brands
targeted fell from an all-time high of 441 in
April 2013 to 379 in September 2013.
Attack vectors continued to evolve, placing
social media at forefront of crimewares
vanguard in the quarter. In the 3rd quarter of
2013, we also saw a change in the phishing
themes used by malware authors. An
emphasis on social media-themed subjects,
such as Invitation to connect on LinkedIn,
was used to entice users who would be used
to seeing such subjects, said APWG
contributor Carl Leonard of Websense
Security Labs.
Encryption use continues to grow
Use of encryption continues to grow in
response to consumer concerns, privacy
compliance regulations and on-going cyber
attacks and yet there are still major challenges
in executing data encryption policy, according
to a Ponemon Institute study.
Key ndings:
Steady increase in the deployment of
encryption with 35% of organizations having
an enterprise wide encryption strategy
Most organizations deploy encryption to
lessen the impact of data breaches
The number one perceived threat to
sensitive data is employee mistakes rather
than external attack
Two biggest challenges faced by
organizations executing a data encryption
policy are knowing where sensitive data
resides and managing the actual technology
Key management identied as a major issue
by more than half of organizations
Organizations with the highest security
posture are now three times more likely to
have a formal encryption strategy than those
with the lowest security posture.
The results of the study show there has been
a steady increase in the deployment of
encryption solutions used by organizations
over the past nine years, with 35% of
organizations now having an encryption
strategy applied consistently across the entire
enterprise compared with 29% last year.
The survey also indicated that only 14% of
organizations surveyed do not have any
encryption strategy compared with 22% last
year.
Mobility is the weakest security link
Surveying more than 750
security decision makers and
practitioners, a CyberEdge
Group report found that more
than 60 percent had been
breached in 2013 with a quarter
of all participants citing a lack of
employer investment in
adequate defenses.
Key ndings include:
Concern for mobile devices. Participants
were asked to rate on a scale of 1 to 5, with
5 being highest their organizations ability
to defend cyber threats across nine IT
domains. Mobile devices (2.77) received the
lowest marks, followed by laptops (2.92) and
social media applications (2.93). Virtual
servers (3.64) and physical servers (3.63)
were deemed most secure.
The BYOD invasion. By 2016, 77 percent of
responding organizations indicate theyll have
BYOD policies in place. 31 percent have
already implemented BYOD policies, 26
percent will follow within 12 months, and
another 20 percent will follow within two years.
Inadequate security investments. Although
89 percent of respondents IT security budgets
are rising (48 percent) or holding steady (41
percent), one in four doubts whether their
employer has invested adequately in cyber
threat defenses.
Malware and phishing causing headaches.
Of eight designated categories of cyber
threats, malware and phishing/spear-phishing
are top of mind and pose the greatest threat to
responding organizations. Denial-of-service
(DoS) attacks are of least concern.
Ignorance is bliss. Less than half (48
percent) of responding organizations conduct
full-network active vulnerability scans more
frequently than once per quarter, while 21
percent only conduct them annually.
Careless employees are to blame. When
asked which factors inhibit IT security
organizations from adequately defending
cyber threats, low security awareness among
employees was most commonly cited, just
ahead of lack of budget.
What do government security pros
think?
Tripwire and the Government
Technology Research Alliance
(GTRA) announced the
results of a U.S. government
cybersecurity survey that
evaluated the attitudes and
responses of 111 security and
compliance professionals from
U.S. government agencies and contractors.
Cybersecurity continues to be one of the top
priorities of senior executives in the federal
government, said Ron Ross, fellow at
National Institute of Standards and
Technology (NIST). Studies, such as this one,
bring together important data points that help
decision makers assess trends and take part
in an ongoing dialog that will help us craft
effective solutions to our difcult and
challenging cybersecurity problems.
Key ndings include:
60 percent believe the new NIST framework
will improve security.
55 percent believe government IT security
has improved due to the administrations
policies.
46 percent say they have seen reductions in
risk due to continuous monitoring efforts.
43 percent of IT security and compliance
employees consider poor governance and the
dysfunctional Congress the biggest security
threat we face.
45 percent of respondents believe funding is
the greatest challenge their agency faces in
successfully implementing cybersecurity
programs; only 37 percent believe they have
adequate resources to properly implement
policy; and when asked what federal security
leaders should do to connect security to the
agency mission, the second-most popular
response was more funding.
How Edward Snowden's actions
impacted defense contractors
A new ThreatTrack Security study sheds light
on the attitudes of a very exclusive group of
IT and security managers - those employed by
U.S. defense contractors - at a time when
national cybersecurity is under scrutiny.
75% of the respondents indicated that the
Edward Snowden incident has changed their
companies' cybersecurity practices in one of
the following ways:
55% say their employees now receive more
cybersecurity awareness training
52% have reviewed or re-evaluated
employee data access privileges
47% are on higher alert for anomalous
network activity by employees
41% have implemented stricter hiring
practices
39% say their own IT administrative rights
have been restricted.
63% of the survey respondents hold either
secret, top secret or condential clearances.
However, of those who have access to or
store condential information, 27% do not hold
such clearances. This represents a potential
privileged access problem wherein contractor
employees without such clearances may have
easy access to sensitive government data.
Lack of skills hindering appsec
programs
An ongoing shortage of skills in application
security is severely hampering the
implementation of effective Appsec programs,
according to SANS. The 2014 Application
Security Programs and Practices survey
queried 488 IT and security professionals
about the current and future state of
application security in their organizations.
"One thing that stands out this year is the
increase in number of organizations with a
formal application security program in place.
Approximately 83% of respondents (up from
66%) have an Appsec program in place, and
more than 37% have a program that has been
operating for more than ve years," says
SANS Analyst Frank Kim. "This indicates that
a lot of progress is being made, but it also
highlights that there is much more to do."
In the survey, more than 35% of respondents
test the security of their business-critical
applications on an ongoing basis, up from
23% in last year's survey. And, encouragingly,
only a small percentage (fewer than 3%) of
respondents left application security to chance
and did not test at all. The survey found that a
lack of qualied staff and lack of skills are
seen as the major inhibitors to instituting
Appsec programs.
"This year's survey provides valuable and
surprising insights into the challenges that
organizations face today in implementing a
successful Appsec program," says SANS
Analyst Jim Bird. "It's not only funding and
getting management buy-inthere are other,
more fundamental problems, including a
shortage of skills, that are preventing people
from taking care of security where it makes
the most difference, upfront in design and
development."
Windows, IE, Java are most
vulnerable
When compared with the numbers from the
previous year, 2013 has seen an increase in
reported security vulnerabilities and, what's
more, the number of critical vulnerabilities has
also risen - although it's considerably smaller
than in 2009.
GFI researchers have combed through the
details provided by the US National
Vulnerability Database (NVD), and have
discovered that in 2013, an average of 13 new
vulnerabilities were reported each day,
bringing the total to 4794 - 447 more that in
2013.
50 percent of the aws were found in products
of only 10 vendors out of 760. The numbers
are both a testament to the number of
different offerings these big rms have and to
their popularity, which naturally points to the
conclusion that they are more often targeted
by hackers and analysed by security
researchers for security aws.
Oracle has topped the list not only because of
Java vulnerabilities, but also because of
hardware aws found in the company devices.
Still, Microsoft can't sigh a sigh of relief, as the
company has had a huge rise in "high
severity" vulnerabilities when compared to
2012 numbers.
Critical vulnerabilities found in its various
operating systems made Microsoft occupy 8
of the rst 9 spots on the list of most targeted
OSes in 2013. Finally, Microsoft's Internet
Explorer, Oracle's Java and Google's Chrome
have ended up occupying the rst three spots
(respectively) on the list of most targeted
applications.
USA still the global spam king
SophosLabs revealed the Dirty Dozen top
spam-relaying nations, as it published the nal
Spampionship league table of 2013. Once
again, it was the USA which earned the
leagues top spot, generating 14.5 percent of
the total spam volume sent during the last
quarter of the year, giving it a clean sweep of
top nishes for 2013. However, the gap to
second place narrowed, with China re-
emerging as a major player in the spam
sending Dirty Dozen, leaping from 4.6 percent
to 8.2 percent, while Russias spam
contribution edged up from 3.0 percent in Q3
to 5.5 percent in Q4.
"The most obvious message from the Dirty
Dozen charts is that the problem of zombied
computers spewing spam is a truly global
one, says Sophos Senior Security Analyst,
Paul Ducklin. Every region of the world is
strongly represented, with the exception of
Africa.
Spammers don't send spam themselves: they
use botnets, or "zombie armies", of malware-
infected computers to distribute their spam for
them, almost always without the owners of the
infected computers being aware.
Turning to the Spampionship table of spam-
relaying countries by population, the numbers
indicate the average spamminess per
person compared to the USA. Results show
things have stayed pretty stable, as Belarus
retained its top spot, with the average
computer there over 10 times more likely to
send spam than if it were in the USA.
Amidst the rapid growth of cloud computing, in multiple studies over the past
several years security and privacy are commonly cited as top concerns. But
a look at the actual experiences of cloud customers nds that often, those
concerns are misplaced.
The existence of misconceptions about cloud
computing is not necessarily surprising. The
industry is still evolving and the range of cloud
services continues to grow. Even the denition
remains unclear to many people. Try asking
several folks to explain the cloud and see
what you get.
Among consumers in particular, the cloud is
still a bit of a puzzle. Web hosting company
Webfusion surveyed over 1,000 people in the
UK in 2013 and found that only 34 percent of
them claimed to understand what cloud com-
puting means. Even smaller percentages cor-
rectly identied services like Dropbox, iTunes,
and Gmail as cloud services. A companion
survey conducted in the US turned up similar
results.
It might seem understandable for consumers
to have challenges describing a somewhat
intangible, technical concept like cloud com-
puting. But what about businesses?
Typically, business leaders seem aware of the
benets touted by cloud providers -- reduced
capital costs, economies of scale, time sav-
ings, and exibility. However, organizations
that are considering cloud computing also
voice a number of apprehensions.
To better understand the concerns that are
acting as barriers to cloud adoption, and to
see whether those barriers matched the expe-
riences of actual cloud customers, in June of
2013 Microsoft commissioned an independent
study by comScore (bit.ly/1lorQzR), which
focused on SMBs.
Respondents were not aware of the connec-
tion to Microsoft. Nor did comScore ask the
respondents about specic offerings. The ob-
jective was to learn about their experiences
with cloud services, regardless of the vendor.
Looking rst at SMBs that had not yet adopted
the cloud, the study found:
60 percent cited concerns around data secu-
rity as a barrier to adoption
45 percent said they were concerned that
using the cloud would result in a lack of con-
trol over their data privacy
42 percent expressed concerns about the
reliability of the cloud.
Ensuing surveys have produced similar re-
sults. In fact, the headline Security concerns
are still holding back cloud adoption recently
appeared on Help Net Security (bit.ly/1drdyfp).
Perhaps we shouldnt be surprised. Plenty has
been written about these barriers to adop-
tion. However, these perceptions are actually
strongly refuted by the realities (and benets)
reported by companies that use cloud serv-
ices. From the comScore survey, among
SMBs actually using the cloud:
94 percent said they had experienced secu-
rity benets in the cloud that they didnt have
with their former on-premises technology ap-
proach. Benets included more consistent
system updates, better spam email manage-
ment and up-to-date antivirus protection
62 percent said that their levels of privacy
protection increased as a result of moving to
the cloud
75 percent said they experienced improved
service availability since moving to the cloud.
Clearly, the benets of the cloud outweigh the
concerns. Improved reliability, security and
privacy protections help give back both time
and money. For example, the study showed
that, as a result of moving to the cloud:
70 percent of SMBs were able to invest more
in product development and innovation, de-
mand creation and expansion into new mar-
kets
50 percent of SMBs have pursued new op-
portunities because of the time they saved
managing security.
But even while cloud services are taking on
much of their customers security work, its
important for businesses to remember they
still have some responsibilities. For example,
cloud customers will still need to manage the
security of their client devices ensuring up-
to-date antivirus software, and educating em-
ployees on the importance of using strong
passwords.
The chart on the following page illustrates the
mix of security responsibilities between cus-
tomer and provider, depending on the service
model deployed.
Employees in particular play an important role
in protecting an organizations data and other
assets. Knowing how to spot phishing scams,
and other types of social engineering is im-
perative. Employees should be trained to be
alert for and to avoid bogus links in email and
on suspicious web sites.
More and more people are also using their
smartphones and other personal devices to
access company data and systems remotely.
To help organizations and their employees
learn to defend against online fraud and other
cybercrimes, Microsoft has published an
Internet Security at Work Toolkit
(bit.ly/1eNaXwL), with tips, fact sheets, videos
and other information. IT Pros should consider
downloading and sharing those resources
across their organization. Specic guidance
(bit.ly/1ehbdj3) on recognizing and avoiding
scams that come through email or web sites is
also available.
For businesses that havent yet adopted cloud
services, a good way to begin is by assessing
the organizations current level of prepared-
ness with the Cloud Security Readiness Tool
(bit.ly/M4O7WV), released by Microsofts
Trustworthy Computing Group in 2012.
Knowing that the vast majority of customers
have experienced security improvements after
moving to the cloud should help ease con-
cerns among potential adopters. Its time we
busted the myth of the insecure cloud, and
remove that perceived barrier, once and for
all.
Jeff Jones is a 25-year security industry professional that has spent the last several years at Microsoft helping
drive security progress as part of the Trustworthy Computing initiative. In the role of Director, Jeff draws upon
his security experience to work with enterprise CSOs and Microsoft's internal security teams to drive practical
and measurable security improvements into Microsoft process and products.
For many years, Jim Reavis has worked in the information security industry
as an entrepreneur, writer, speaker, technologist and business strategist. Jim
is helping shape the future of information security and related technology in-
dustries as co-founder, CEO and driving force of the Cloud Security Alliance.
How has the cloud security landscape
changed in the last ve years? How ma-
ture is the cloud today?
Cloud computing has matured dramatically
over the past ve years, and the use cases
are quite broad. Five years ago, many IaaS
offerings were very simple, with perhaps ve
or so options.
Today many of those same providers have
literally hundreds of offerings, management
tools and product add-ons, and third party
providers have added many new security so-
lutions. Many of these offerings today are
purpose-built for securing cloud providers,
where ve years ago they were often legacy
security products tweaked for the cloud.
Some of the most interesting innovations are
identity as a service, cloud aware encryption,
cloud application control and log manage-
ment.
Cloud adoption is at an all-time high, yet
those that are not using it are still saying
the biggest obstacle is security. What can
service providers do to earn customer
trust? What should customers be on the
lookout for?

Certainly the Snowden issue, which I will dis-
cuss later, is a factor. However, the majority of
customers I talk to say the main issue is com-
pliance and governance. It is about showing
proof of security, rather than technical security
itself, which is a nuance to the issue that
many surveys miss. Solving this is mostly an
educational issue with the various players, I
think.
Policy makers and auditors need to under-
stand how the cloud really operates. Security
professionals need to understand that the
risks are not black and white - you may actu-
ally reduce risks by using a public cloud pro-
vider that has better rewalls and backup
systems. The big thing providers need to do to
increase trust is to be transparent with their
security and governance practices. We think
the CSA Security, Trust and Assurance Regis-
try Program (cloudsecurityalliance.org/star)
provides the global model for trust in the
cloud. It is a control framework that is mapped
against key security standards and require-
ments, it allows providers to publish their se-
curity practices for all to see, includes 3rd
party certication where needed and will in
the future provide continuous monitoring.
Last year you launched the Software De-
ned Perimeter Initiative, a project to de-
velop an architecture for creating highly
secure and trusted end-to-end networks
between any IP addressable entities, allow-
ing for systems that are highly resilient to
network attacks. Are you satised with the
response? What enterprises are working
with you on the development?

We are quite pleased with the response so
far, and a few very large enterprises have im-
plemented prototypes and pilots with positive
results. We have much more work to do in or-
der to publish mature specications and sim-
plify the adoption. SDP is a big idea that says
we are going to need to change our view of
how we implement IP networks to decrease
the global discovery and visibility of comput-
ers. The IP-controlled thermostat in my home
is my business alone.
What trends do you expect to see in the
next 12 months? Do you expect to see a
notable increase in cloud security
automation?

We see the growth in consumer-owned mo-
bile devices and new generations of the Inter-
net of Things creating a force multiplier that
will lead to even more aggressive cloud adop-
tion.
As organizations lose control of the endpoint
device, they will have fewer options to prevent
cloud usage, although solutions are coming to
market to help them manage this usage. We
do see a lot more automation in the entire
lifecycle of acquisition, implementation and
management of cloud services. Whether it is
called cloud brokering or by another name,
we see a lot of this automation coming from
intermediaries.
AS ORGANIZATIONS LOSE CONTROL OF THE
ENDPOINT DEVICE, THEY WILL HAVE FEWER
OPTIONS TO PREVENT CLOUD USAGE
After Edward Snowden's revelations,
there's been a growing trend of organiza-
tions moving or considering moving their
data to cloud providers outside the US.
What's your take on this situation?
We continue to monitor this highly dynamic
situation. According to the survey we con-
ducted in the immediate aftermath of the reve-
lations, 10% of customers outside of the
United States had stopped a cloud project us-
ing a US-based service, while another 56%
said the news would negatively impact their
future adoption of US-based cloud services.
I would say that in considering all of the data
we have analyzed over the past several
months, this trend is not growing and has
somewhat stabilized for the following reasons:
Some of the revelations have uncovered
surveillance programs that have been con-
ducted without the knowledge of the provider,
such as via tapping ber outside of a data
center. This has created an opinion that there
is perhaps less complicity than was originally
imagined and that in some cases the provider
could even be a considered a victim of sur-
veillance.
Generally speaking, customers seem to feel
that the larger US cloud providers are taking
positive actions to provide assurances that
their information is being safeguarded from
unwarranted government access. Many have
announced greater efforts at end to end en-
cryption. Some are even taking legal action
against the US federal government. To be
clear, the customers are telling us there is
work yet to be done, and are forcefully push-
ing for greater transparency from the provid-
ers in their relationships with governments
and in their own management of customer in-
formation.
Still other revelations have shown that coun-
tries besides the US are conducting similar
options to acquire large consumer data sets.
Some of these countries collaborate with the
NSA. Very recently, EU Justice Commissioner
Viviane Redin was highly critical of European
hypocrisy related to the Snowden revelations,
given their own spying activities.
The mature conclusion that is being arrived at
is that an enterprises information must be
safeguarded against a whole host of threats:
domestic, international, private sector, public
sector, criminals, hacktivists, etc.
The general consensus is that organizations
need to take a holistic approach to increase
the baseline of their security, which includes
governance issues related to provider selec-
tion; technical architecture improvements,
such as increased encryption and logle
monitoring and several other practices. More
and more, the CSA best practices are being
cited and used as the foundation of this in-
creased security baseline.
Mirko Zorz is the Editor in Chief of (IN)SECURE Magazine and Help Net Security (www.net-security.org).
This article is an attempt to perform an analysis of cloud security by taking
into consideration two aspects. The rst one is related to the different cloud
security models (IaaS, PaaS, and SaaS), and we will attempt to assess some
of the security risks for each. The second is related to the uncertainty of such
assessments and the ways to deal with these uncertainties, in particular
based on Service Level Agreements (SLAs), insurance, and certications.
The basic premise for our analysis is that you,
as a representative (CSO/CIO/) of a com-
pany, are considering migrating to the cloud,
but are concerned about the security implica-
tions of such a move.
It should be noted that our analysis is not in-
tended for life-critical applications, or any in-
dustry where a security breach is considered
completely unacceptable - our analysis is for a
typical business application in a typical busi-
ness enterprise, where there is a need to nd
the right balance between solution costs and
solution risks.
Table 1 on the following page illustrates, more
or less, a typical split of security responsibili-
ties between a cloud service provider (CSP)
and the customer for different cloud service
models.
IaaS and security
In the Infrastructure as a Service (IaaS)
world, a CSP usually provides you with a
number of virtual servers and other virtual ap-
pliances that are essentially managed by
yourself. One can reasonably argue that with
IaaS, most of the security is still handled by
the customer. If an organization is switching a
part of its infrastructure to IaaS, it still needs to
congure rewalls, IPSs, IDSs, etc. In this set
up, the customer is usually also responsible
for OS patching and application security. In
theory, the CSP may provide help with setting
the system up, but in practice that help is
Table 1.
rather limited, so the balance of responsibili-
ties usually looks like this: the CSP provides
the (usually virtualized) servers, the customer
does the rest.
This approach can be either a blessing or a
curse, depending on your point of view. If
you're looking to delegate to the cloud as
much as possible, it is not exactly good news:
you still need to run your own IT/security de-
partment (in a cloud/IaaS setup, IT/security
department can be more easily outsourced to
the cloud, but this is beyond the scope of this
article). On the other hand, if you're more
concerned about migration to the cloud being
smooth, out of all cloud migration scenarios
IaaS requires the least possible change, with
the logical part of your infrastructure kept
more or less intact - or at least not that af-
fected as it would be when migrating to the
other cloud models.
In a sense, IaaS is the closest you can get to
managing your own infrastructure. From many
points of view - including the security one - it
is very close to how traditional hosting ISPs
function, with the main differences being the
following:
1. You get on-demand scaling and the asso-
ciated reduction in costs, which is usually the
main reason why enterprises are switching to
the cloud in the rst place.
2. In exchange for reduced costs, you need to
deal with a set of cloud-specic security risks
which include at least two rather broad cate-
gories:
Category 1: Attacks coming from some other
customer of the same CSP. Anybody, includ-
ing a hacker from some distant country, can
(attempt to) crack the security that separates
virtual machines - yours and the attacker's - in
the cloud.
Category 2: Attacks via cloud administrator.
How big are these risks? In practice, we feel
that the risk for attacks cracking the barrier
between two virtual machines is not that big.
We've taken a look at CVE vulnerabilities for
the last 3 months, and have found that only
about 20 of 1300 vulnerabilities (about 1.5 %)
are virtualization-related. It is also interesting
to note that out of these 20 vulnerabilities, at
least half are directly related to UI, so they're
sitting somewhere in between the two risk
categories mentioned above. Of course, the
number of known vulnerabilities shouldn't be
confused with real risks, since a single vulner-
ability can be enough to defeat security com-
pletely, but it still provides some information
about the attack surface. Lets say that this
risk is not too high, and can be assessed with
some level of certainty.
Unfortunately, the risk of being hit with an at-
tack via cloud administrator is much more dif-
cult to assess. Assuming that cloud adminis-
trators are running desktops/laptops (and they
usually are), we need to recall that any system
is only as strong as its weakest link. So, to
break into your cloud (and to access your
sensitive data) it is not necessary for the at-
tacker to break into a neighboring VM via a
weakness in the hypervisor.
Instead, it is sufcient to break into the
desktop/laptop that the cloud administrator
uses to manage the cloud, including your VM.
Normally, administrators have the option to
make an image of your VM while it is running.
Such an image is rarely encrypted, and is
even more rarely encrypted with a key that is
not available to the same cloud administrator.
Also, such an image can be made without you
noticing it, but even worse that that: even if
you're using encryption, such an image will
contain all your keys in unencrypted form. All
this combined together, it means that whoever
controls the desktop/laptop of the cloud ad-
ministrator (the one who can create an image
of your VM) should be considered as a person
who can access your data.
To make things worse, it is not only the cloud
administrator who has such powers, but also
any hacker that is sitting half-a-world away but
has managed to convince the cloud adminis-
trator to open a malicious e-mail attachment
several months ago. And as the March 2011
RSA breach (tinyurl.com/o6s3o3k) has dem-
onstrated, such spear-phishing attacks are
extremely difcult to prevent even in the most
security-conscious environments.
Compare it to the situation with more tradi-
tional ISP hosting: while the ISP administrator
can physically take your HDD and compro-
mise your data, taking out your HDD cannot
be done remotely, and this makes a world of
difference. With traditional ISP hosting, a re-
mote attacker would need to bribe the ISP
admin to obtain access to your sensitive data.
This is theoretically possible, but not that
likely. With a CSP, all the remote attacker
needs is to spear-phish the laptop of the CSP
administrator, and that is a much more realis-
tic scenario.
The likelihood of such spear-phishing attacks
being successful depends heavily on the se-
curity practices of a specic CSP, and may be
at any point of the spectrum between security
as solid as rock and has holes large enough
to let USS Enterprise through.
In practice, it means that while risks in the rst
category can be assessed with at least some
level of condence and without a deep analy-
sis of the security practices of a specic cloud
provider, risks related to attacks via cloud ad-
ministrator depend on the practices of a spe-
cic CSP, and this leads to uncertainty in se-
curity assessments.
PaaS and security
In the Platform as a Service (PaaS) world,
you normally get an application platform that
your developers use for developing new appli-
cations the migration of existing applications
without rewriting them is usually not an option.
PaaS usually provides APIs to access plat-
form services, such as network and storage.
Unfortunately, for PaaS-based systems, secu-
rity is even more difcult to assess than for
IaaS-based ones. When using PaaS, you in-
herit all the security risks from IaaS, and face
additional ones specic to PaaS.
With PaaS, separation between customers is
handled by the PaaS provider. Database (or
any other storage) separation is normally a
responsibility of the CSP, too.
It leads us to the following categories of addi-
tional PaaS-specic risks:
Category 3: Risks related to holes in PaaS
API-based separation - unless the PaaS pro-
vider uses VMs to separate customers. APIs
(especially higher-level APIs normally used in
PaaS) are notoriously buggy and insecure, so
there is considerable uncertainty when it
comes to security assessments of these risks.
Category 4: Risks related to database ac-
cess separation. Data access separation is
tricky for each way to do it right there are at
least several dozen ways to do it wrong. While
we're shouldnt automatically assume that all
PaaS providers are doing it wrong, we still
cannot rule it out; once again, it means that
risk assessment is not possible without a de-
tailed security analysis of the specic CSP.
SaaS and security
If you're dealing with the SaaS cloud model,
things can be a bit simpler for you. Usually,
SaaS CSPs are providing a very specialized
environment for a very specic task, and there
is little pressure to go down the slippery
slope of moving more and more data into the
cloud without understanding risks associated
with such a move. If the data that you have
moved into the cloud is not sensitive, you're
perfectly ne.
If, on the other hand, that data is sensitive,
from a security perspective most of the risk
categories listed above still apply.
Summary of risk categories
Now let's summarize, in one table, all the
CSP-related risks mentioned above. While the
list of the risks represented by those four
categories is by no means exhaustive, it is
sufcient to illustrate the magnitude of security
uncertainties faced by enterprises planning a
migration to the cloud.
Table 2.
While uncertainty estimates in Table 2 are
subjective, we've discussed them with a few
people from the security community and have
found that there is not much disagreement on
them. It should also be noted that for the fol-
lowing discussion we won't rely on specic
uncertainties, but only on the assumption that
for all cloud service models, there is substan-
tial security uncertainty related to the specic
security practices of a specic CSP.
When we take a look at Table 2, we can see
that even for IaaS, which has the lowest over-
all CSP-related uncertainty compared to the
other cloud service models, security uncer-
tainty is high. In other words, without going
deep into the details of a specic CSP, we
cannot be sure how good their security is.
One (especially someone representing a
CSP) may say: Hey, you're speaking about
security breaches affecting a CSP, but the
same types of breaches can happen in your
own environment. So, what's the difference?
Granted, most of the risk categories men-
tioned above also apply to your own environ-
ment, but there is one fundamental difference:
in your own environment you can nd out how
risky it is (it will be very expensive and labor-
intensive, but still possible). In a CSP envi-
ronment it is much more difcult, up to the
point of sometimes being next to impossible.
In a nutshell, this how difcult is to nd out
how secure your system is is exactly what
security uncertainty is about.
The importance of security certainty
The importance of being certain about the se-
curity of your system cannot be overesti-
mated. Security is a very special entity com-
pared to other technological issues busi-
nesses need to deal with. If the system does
not do what is expected, or is not performing
well, it is usually highly visible so the problem
can be addressed right away. When it comes
to security, you never know that the system is
broken until it is too late. And given the scale
of potential consequences arising from a se-
curity breach - which in worst case can easily
cost hundreds of millions or even billions - tak-
ing uninformed security risks is the last thing a
CTO/CIO/CSO should do.
Dealing with uncertainty: Audits, SLAs, in-
surance, and certications
In business, there are four well-known ways to
deal with similar uncertainties:
1. Perform a security audit of the CSP your-
self. Fortunately or not, in most cases this
won't y (unless, of course, you're that big that
you have your own cloud).
2. Rely on the guarantees that a CSP pro-
vides in its Service Level Agreement (SLA).
3. Rely on insurance (provided by a third-party
insurance company)
4. Rely on a third-party security audit (ideally
by a reputable authority).
Items 1, 3 and 4 on the list are rather obvious,
but item 2 one may require a bit of explana-
tion.
Service Level Agreements (SLAs) are a well-
known way for service providers to provide
certain guarantees for difcult-to-estimate as-
pects of their service. In a sense, it is the pro-
verbial putting your money where your mouth
is. If an ISP says I guarantee that your net-
work will be up 99.99% of time, or it is money
back for this month, it increases the cus-
tomer's condence that the ISP is able to
achieve this target. Also, it creates a very spe-
cic incentive for the ISP to make sure that
the network is indeed up and running at least
99.99% of time. These two sides of a SLA
combined explain the high popularity of SLAs
and SLA guarantees in the world of ISPs,
mitigating to a great extent the inherent uncer-
tainty of ISP service quality.
Security guarantees in CSP SLAs
You still want to migrate to the cloud, and
there is no reliable third-party security audit.
What kind of security guarantees you need
from the CSP's SLA?
Let's assume that we are considering whether
a migration to cloud makes business sense,
and let's assume that the decision is purely
money-based (which implies, among other
things, that the potential of a single breach is
not prohibitively expensive for the company).
A bit of math
Let's dene the following:
S: Amount that could be saved due to transfer
to the cloud during a certain time period T (S
can be expressed, for example, in dollars/
year)
L: Amount of loss if a security breach happens
(per security breach; L can be expressed, for
example, in dollars)
!: Rate of security breaches - a frequency of
security breaches expressed in terms of a
number of security breaches per time period
T. Somewhat similar to the failure rate from
engineering. Under our conditions, 1/! can be
seen as average time between security
breaches (somewhat similar to MTBF)
": Losses due to security breaches happened
during time period T; " = L* !
The mathematical expectation of security
breach losses " can be calculated as:
E["] = E[L*!] = L* !
For the benet arising from moving to the
cloud to be positive, it is required that S >
E["], that is, S > E["] = L * ! or, respectively, L
should be less than S / !.
If compensation C is provided (for example, in
SLA, or by a cyber-insurance company) in
case of a security breach, then losses includ-
ing compensation L1, if security breach hap-
pens, become L1 = L C and "1, security
breach losses including compensation, will
have a mathematical expectation of
E["1] = E[(L-C)*!] = (L-C) * !
Respectively, for the benet arising from mov-
ing to the cloud to be positive, it is required
that S > E["1] = (L-C) * ! or L-C < S / ! [**]
The value of ! depends heavily on the secu-
rity procedures adopted by the CSP. To ac-
count for the worst-case scenario, one may
assume that value of ! may happen to be ar-
bitrary high; then, taking a limit as ! ap-
proaches innity, [**] effectively becomes
L C < 0, [***] which, in simple words, means
that the compensation per breach (from the
CSP provider or from the insurance company)
should be greater than the losses from the
breach.
Back to business - SLAs
The formula above [***] can be summarized
as follows: if considering the worst-case sce-
nario in presence of an arbitrarily high security
uncertainty, the move to the cloud will be cost-
efcient only if the compensation in case of
security breach (guaranteed by the SLA or by
an insurance company) will be more than cost
of security breach to the company. Or in even
simpler form: with the conservative assump-
tions above,
In the absence of reliable third-party certica-
tions, putting data in the cloud only makes -
nancial sense if the cost of the data trans-
ferred into the cloud is less than the compen-
sation provided by the SLA or by cyber-
insurance in case of a security breach.
It is interesting to note that the clause in the
SLA you should be looking for doesn't depend
on the type of the CSP (being IaaS, PaaS or
SaaS); it is more a question of if the CSP will
be willing to provide a compensation guaran-
tee good enough for your migration.
In practice, we wouldn't expect CSPs provid-
ing SLA guarantees themselves in case of a
security breach on the order of millions; the
reason being that a CSP breach is likely to
affect many (if not all) of its customers, and
paying more than few months worth of their
fees would be way too risky for the CSP. Still,
having clauses in the SLA that guarantee you
a few months worth of your CSP fees can al-
low you to move the least sensitive part of
your expensive infrastructure to the cloud.
On the other hand, there is a possibility that
the CSPs may obtain cyber-insurance cover-
ing security breaches, from independent in-
surance companies; this may allow them to
substantially increase the guarantees in the
SLAs.
Back to business: Cyber-insurance
As discussed above, CSPs themselves are
not likely to provide compensation on the or-
der of millions in case of a security breach be-
cause of risks being too high for a CSP. How-
ever, there are companies out there whose
whole business revolves around taking that
risk for others: insurance companies (in our
case cyber-insurance companies).
In theory, we could expect two different insur-
ance models for our case. In the rst model, a
cyber-insurance company could insure cus-
tomers against security breaches directly. In
the second one, a cyber-insurance company
may insure a CSP against security breaches,
so that the CSP is able to put higher guaran-
tees into their SLA. As it is usual with group
insurance, we could expect that the second
model would be more cost-efcient for the
customer.
In practice, there are indeed companies that
provide cyber-insurance against security
breaches. One such company is CloudInsure,
which has been reported to provide insurance
with compensations up to $1 million (and in-
surance costs in the range from $5K to $10K
per year). It has also been reported that
CloudInsure is ready to insure CSPs (as de-
scribed in second insurance model above),
though as of now it is not clear if any CSP is
including their (or anybody else's) insurance
against security breaches in their SLAs. In any
case, even with current pricing, cloud insur-
ance may easily allow migration of the data
worth on the order of $1M to the cloud (though
you need to account for insurance cost when
calculating cost efciency of such a move).
The maze of certications: Compliance
and risks as two sides of the same coin
When you ask any sizable CSP about their
security certications, they will show you a
very long list: HIPAA, SAS70, SSAE16 SOC1/
SOC2, FIPS 140, ISO 27001, PCI DSS, and
so on. The sheer number of items in this al-
phabet soup served by the CSP can be really
impressive. Now, to understand the real value
that these certications have for you as a CSP
customer, we need to dene what your goal
with certication is. There are two possible
reasons why you may want a certication from
your CSP compliance and risk assessment.
It should be understood that, strictly speaking,
being secure does not imply compliance with
standards, and vice versa. While usually
measures aimed at reaching compliance do
help security (and therefore reduce risks), and
often measures aimed to improve security
help to reach compliance, the nal states of
being secure and being compliant are not
100% the same, and separate analysis of
compliance and security may be necessary to
ensure that the system is both compliant and
secure.
Certication and compliance
The rst reason for you to make sure that your
cloud provider does have certication is that
you need to be compliant with a certain regu-
lation. For example, if you need to store medi-
cal data and you're located in the US,
chances are that you need to be compliant
with HIPAA data protection requirements, with
no way around it. In a similar way, if you want
to process credit cards, you likely need to be
compliant with PCI DSS. It means that if
you're moving respective portion of your data
and/or processing to the cloud, you will very
likely need your CSP to be compliant with an
appropriate regulation and have appropriate
certication. We should note that this aspect
of certication is not exactly the focus of the
article, so we won't concentrate on it too
much. One word of caution though even if
your CSP is compliant with a certain regula-
tion, it doesn't mean that by using their serv-
ices you will be automatically compliant, too.
Just one example your CSP can be an IaaS
cloud provider, and can be PCI DSS compliant
in a sense that they handle physical security
according to PCI DSS, and that their own sys-
tems are secure according to PCI DSS re-
quirements; however, as they're IaaS, they
can't make sure that your web server sits in
the DMZ, or that your rewall (provided by
IaaS) is congured according to PCI DSS; it
means that even if a CSP is perfectly PCI
DSS compliant, you as a merchant can be
easily non-compliant with all the potentially
unpleasant consequences. Or if we express it
in mathematical terms as a rule of thumb,
CSP compliance is usually a necessary condi-
tion for you to be compliant, but not a suf-
cient one.
Certication as a way to assess security
risks
The second reason to choose a CSP with a
certication is to reduce security uncertainty
and therefore to allow you to move more sen-
sitive applications and/or data to the cloud.
Here, however, some understanding of the
real meaning of various certications is nec-
essary. There are several problems with un-
derstanding and comparing the certications
presented by most cloud providers. The very
rst word of caution is that you should never
judge a CSP for risk accessing purposes
based on the number of the certications they
have. Certications should be taken one by
one, and the merits for risk assessment
should be evaluated for each certication.
Security-unrelated standards and certica-
tions
There are many standards and certications
out there which are popular, but that have little
to do with security. One such example is SAS
70 (in fact, the misuse of SAS 70 as a security
certication was/is so frequent that Gartner
has even issued a press release titled Gart-
ner Says SAS 70 Is Not Proof of Security,
Continuity or Privacy Compliance).
Paper-only compliance and under-
specication
A big problem with many certications is that a
big part of them can be complied with without
making any changes to the system, but
merely by producing paperwork. While paper-
work can have its own merits (for example,
very few people will argue about the useful-
ness of security policy), certications that as-
sure very little beyond pure paperwork don't
really provide much in terms of risk assess-
ment. In other words, if the only thing a CSP
needs to do to comply is to produce a pile of
papers, the certication is not worth much for
practical purposes (it may still be needed for
compliance though).
A close cousin of paper-only compliance is
under-specication. It happens when the
standard to be complied with does provide
some value beyond paperwork, but the set of
security controls within is not sufcient to pro-
vide comprehensive coverage. For example, if
a certain standard species that passwords
need to be changed on regular basis, it is a
good thing, but if the standard doesn't require
using rewalls to protect certain parts of the
CSP infrastructure, it may allow fully compliant
organizations to not use rewalls at all. Taking
into account the fact that a system is only as
secure as its weakest link, this is clearly not
enough to provide meaningful improvement in
assessed risks.
Lack of cloud specics
The cloud environment has its own unique
challenges. For example, while Category 2
risks described earlier are not exactly unique
to the cloud, in practice there is a substantial
increase in Category 2 risks in cloud environ-
ments.
In particular, consolidation of control on ad-
ministrative PCs used to control hundreds of
CSP clients makes these PCs much more at-
tractive attack targets, which in turn makes
attacks on them much more likely and much
more damaging. We strongly feel that to pro-
vide a meaningful capping of risks, such is-
sues need to be addressed in the standards
and certications.
Navigating the alphabet soup of existing
certications
Let's take a look at existing security-related
certications from the point of view of risk as-
sessment in general, and caveats listed above
in particular.
SAS 70
SAS 70 (also known as Statement on Audit-
ing Standards No. 70 report) is one example
of a statement that is widely (and wrongly) de-
clared as a security certication. French
Caldwell, research vice president at Gartner,
has said about SAS 70: SAS 70 is basically
an expensive auditing process to support
compliance with nancial reporting rules like
the Sarbanes-Oxley Act (SOX). [...] Chief in-
formation security ofcers (CISOs), compli-
ance and risk managers, vendor managers,
procurement professionals, and others in-
volved in the purchase or sale of IT services
and software need to recognize that SAS 70 is
not a security, continuity or privacy compli-
ance standard.
Another Gartner research vice president, Jay
Heiser, has added that Given that SAS 70
cannot be considered as proof that an offered
IT service is secure, it should be a matter of
suspicion when a vendor insists that it is.
In 2011, SAS 70 has been retired in favor of
SSAE16.
SSAE 16 (SOC1/SOC2)
SSAE16 (Statement on Standards for At-
testation Engagements) is a standard that
effectively replaced SAS 70. SSAE 16 audit
reports come in three different avors Serv-
ice Organization Control 1 (SOC 1), SOC 2,
and SOC 3. Out of these, SOC 1 is a more or
less direct replacement of SAS 70, and has
been criticized by the security community be-
cause its controls are (like those in SAS 70)
self-dened, and therefore merely having a
SOC 1 report does not provide enough infor-
mation about security properties of the sys-
tem. SOC 2 reports, however, were received
more favorably, but with a reservation that
even SOC 2 is still underspecied with respect
to security details.
It is worth noting that both SAS 70 and SSAE
16 are standards prepared by AICPA, which
stands for American Institute of Certied Pub-
lic Accountants and, as we understand, the
reports are meant to be prepared by CPAs
(Certied Public Accountants). We have no
doubts that it has severely limited technical
details in these statements (one cannot rea-
sonably expect a CPA to know security at the
level of a (ISC)2 CISSP).
HIPAA
While HIPAA has never been intended to cer-
tify the security of a cloud provider, it is still
used by many CSPs as a part of alphabet
soup of certications they have. When looking
at HIPAA regulations, it's easy to notice that
there are two different sets of rules Privacy
Rule and Security Rule. HIPAA Privacy Rule
has very little to do with security at all; HIPAA
Security Rule has certain rules regarding se-
curity, but they're very few and too far be-
tween to provide substantial coverage for a
cloud provider environment. It doesn't mean
that you won't need a HIPAA-compliant CSP:
if you need to store PHI data in the US, you
certainly need a HIPAA-compliant provider.
Still, having HIPAA compliance helps risk as-
sessment of the CSP in a rather limited way.
ISO 27001
ISO 27001 (or more precisely, ISO/IEC
27001:2005 or newer version 27001:2013)
is a standard that aims to provide a framework
for information security management systems
(ISMS). In the industry, ISO 27001 has a long-
standing reputation of being vague and un-
derspecied (its cousin 27002 is better de-
ned, but as far as we know, there is no ISO
27002 certication). With the release of
27001:2013 there was an improvement in this
regard (now normative Annex A contains spe-
cic controls from 27002:2013 to be imple-
mented), however, it still looks pretty much
underspecied and lacking cloud specics for
the purposes of the risk assessment in the
cloud. While covering pretty well such areas
as organizational security, human resources
security, and physical security, the important
elds of network security and encryption are
grossly underspecied (just as one example,
we weren't able to nd in ISO 27001 or ISO
27002 any reference to a minimal strength of
encryption algorithms which are allowed to be
used).
Based on this, our subjective opinion is that
the value of ISO 27001/27002 for practical
purposes such as risk assessment is rather
limited, unless it is complemented by other
means (such as the Cloud Control Matrix).
FIPS 140
FIPS 140 (Federal Information Protection
Standard publication 140) is an interesting ex-
ample of a standard that is fully specied,
though, taken alone, it still cannot be used as
a reliable metric of CSP security.
Most of the standards mentioned before were
listing too few technical details, while staying
at a level which is too high to provide neces-
sary coverage. In contrast, FIPS publications
tend to be very down-to-earth, and have all
the necessary technical details.
The problem with the application of FIPS
standards (including, but not limited to, FIPS
140) to CSPs is that the scope of FIPS is lim-
ited to cryptography. As it is well-known in the
security industry, having good cryptography is
only one of many prerequisites to having a
secure system, so using only FIPS 140-
compliant algorithms is clearly not enough to
achieve meaningful system-wide security.
Moreover, in many cases even combining
FIPS 140 with higher-level standards like ISO
27001 may be insufcient to guarantee secu-
rity: the problem here is that between the
high-level view of ISO 27001 and the low-level
view of FIPS 140 there is still a substantial
uncovered gap related in particular to security
protocols used within the system.
As a big portion of security attacks comes ex-
actly from breaches in the protocols (see, for
example, recently developed BREACH attack
on TLS), there is still a need for a more com-
prehensive security standard covering the
whole stack of technology used, from high
level to low level, including everything in be-
tween.
PCI DSS
PCI DSS is, in our opinion, a very nicely writ-
ten standard, that doesn't suffer from the
under-specication that most of previously
discussed ones do. While it is reasonably de-
tailed, its still possible to comply with it in
practice.
The main problem we see with it for the pur-
pose of CSP risk assessment is that PCI DSS
has a very narrow scope. Everything that is
not related to credit card numbers is out of
PCI DSS scope, so if you're absolutely inse-
cure but do not process credit cards at all,
technically you're still compliant with PCI DSS.
On the other hand, if PCI DSS certication for
a CSP can be interpreted as it is safe to
process credit card anywhere on this CSP
(and our point of view is that this is the only
reasonable interpretation of PCI DSS claims
for the CSP, but it is better to ask your specic
CSP about their interpretation) it will say
quite a lot about security processes of a spe-
cic CSP. While still lacking cloud-specic de-
tails, it is probably one of the most compre-
hensive existing security standards.
Our subjective opinion is that when we see
that certain CSP is PCI DSS certied (assum-
ing that interpretation of being certied above
stands), it makes us a bit more assured about
this CSPs security practices; still, due to the
limited scope of PCI DSS, for the purposes of
a risk assessment of a CSP we'd clearly pre-
fer to rely on some certication with much
wider scope.
New kids on the block: CCM and STAR
Probably understanding the limitations of the
existing certications, several years ago the
Cloud Security Alliance (CSA) has started the
development of a new industry guidance
framework, the Cloud Control Matrix (CCM).
In September 2013, CCM 3.0 has been re-
leased.
The CCM specically aims to provide security
guidelines for CSPs to follow, and to provide a
framework for assessment of security con-
trols. It covers many of the gaps of the previ-
ous standards; one of the most important ad-
vantages of the CCM is that it is cloud-
specic, and as such, it addresses cloud-
specic issues.
Just as one example: based on the cloud
service model and the cloud being public or
private, CCM suggests which controls might
fall under CSP responsibility versus the con-
suming organization; this delineation of re-
sponsibility in the cloud is one major item
desperately missing from previous standards.
While we feel that there is still a long way
ahead for the CCM (in particular, covering
cloud-specic attack vectors, such as those
described in the described four categories),
we strongly feel that the CCM represents a
desperately needed move in the right direction
(moreover, as far as we know, CSA's efforts
are the only efforts in this area).
The STAR (CSA Security, Trust and Assur-
ance Registry) is an initiative parallel to the
CCM, and is partially based on it. Up until re-
cently, only STAR layer 1 (self-assessment)
was available; as with any self-assessment,
its value for risk assessment was rather lim-
ited. However, in September 2013 it has been
announced that STAR certication layer 2,
based on independent third-party assess-
ments and audits, is available from BSI (Brit-
ish Standards Institute). The STAR certica-
tion is based on ISO 27001, but is aided with
certain controls from the CCM (currently from
CCM 1.4, though in the future support for later
versions of CCM is planned), which allows it
both to deal with under-specication and to
provide cloud-specic controls.
As of now, STAR layer 2 certication is proba-
bly the best single thing a CSP can show you
to demonstrate that they really care about se-
curity. At this time, STAR layer 2 certication is
still so very new that no CSP has it yet; how-
ever, we could expect rst STAR layer 2 certi-
cations by mid-2014.
Conclusion
We have performed an analysis of security
uncertainties inherent to cloud service provid-
ers (CSPs) and some possible ways to over-
come these uncertainties. We have found that
while compensations for security breaches (in
SLAs and/or by cyber-insurance) can allow
enterprises to push some of the less sensitive
data to the cloud, further reduction of security
uncertainty is still necessary. Such an uncer-
tainty reduction can be achieved via cloud-
specic efforts like the CCM and the STAR.
We feel that such efforts are absolutely nec-
essary to deal with security uncertainty, which
severely limits information being moved into
the cloud.
This article has been prepared based on cloud security research project conducted by authors at OLogN
Technologies AG, Liechtenstein.
Sergey Ignatchenko is a Security Researcher with security experience in various elds going back to 1996.
Among other things, he has designed a comprehensive and regulations-compliant security solution for one of
G8 stock exchanges, and a security architecture for a very large system with 100+ millions of users, 10+ thou-
sands of RSA SecurID tokens, and capable of processing billions transactions per day. His systems have
passed numerous security audits with ying colours. He can be reached at [email protected]
Dmytro Ivanchykhin is an Information Security Consultant with interests spanning areas from mathematical
aspects of cryptography (with an emphasis on abstract algebra and Galois elds) to applied cryptography and
secure protocols. He can be reached at [email protected]
The advent of cloud services can be referred to as one of the most serious
historical changes in the data storage and access model. Ease of deployment,
scalability and economic efciency spurred a mass migration of businesses
from the standard CAPEX model to the cloud. In the rush of implementing
these new business practices, the crucial point of observing the privacy of
and controlling access to in-house data has somewhat fallen by the wayside.
Ironically, recent spy scandals revealed
through the efforts of Edward Snowden and
the Wikileaks project made it clear that the
question of privacy and data security has
never been as important as it is today.
Who is spying on me?
Revelations in information security lead us to
realize that the world of information is much
more dangerous than it appeared before. The
NSA is developing wide-scale global surveil-
lance projects (tinyurl.com/okdes6f), and is
already capable of silently capturing Internet
communication. At the same time, intelligence
bureaus are working on establishing back-
stage contacts with rst-tier online service
providers such as Microsoft, Google and
Amazon. Despite the providers efforts to re-
fute the collaboration claims by stating that no
mass uncontrolled disclosures of private data
had taken place, the public continues to rail
against the connection and to call for the
companies to respect their customers right to
privacy.
On the other hand, it is important to remember
that while international cloud legislation is still
quite immature, data centers and the data
they store are normally subject to the laws of
the country in which they physically reside.
Data privacy and protection laws in that
country may differ signicantly from that of
your homeland, and theres always the possi-
bility of natural disasters, revolutions, and lo-
cal wars creating a chaotic situation that can
result in your data being accessed by people
that should not have access to it.
The underworld
Aside from global surveillance threats, we
should keep in mind that the data centers of
major cloud providers are attractive targets for
hacker gangs, since a single successful
break-in attempt can result in access to an
impressive amount of potentially valuable in-
formation. We are, therefore, expecting a
sharp rise in the number of professional at-
tacks targeting cloud services in the near fu-
ture, and believe that a certain percentage of
them will prove to be successful. The underly-
ing technology is not going to be the only tar-
get; a bribed or blackmailed service providers
employee might be as helpful in gaining ac-
cess to the data as a rootkited server, and at a
much lower cost for the attacker.
What is worse is that its really tempting for
cloud service providers to hush up the hacks.
Making this information public affects the pro-
viders reputation, and privacy regulators may
apply tangible nancial sanctions for privacy
legislation violations. Finally, there is usually
little or no direct evidence of the theft, and
there is always a chance that the attackers
wont brag about it publicly. All this means that
there is no real incentive for the service pro-
viders to announce the theft until it becomes
evident from information provided by other,
indirect sources.
One should not underestimate the risks of
data theft from the cloud service providers fa-
cilities due to technical or human error. History
shows us that even the largest and most re-
spected companies make mistakes in imple-
menting information security measures. The
need to provide data reliability and accessibil-
ity makes the service providers job difcult.
The customers data is usually stored in highly
redundant form, across a variety of sites with
synchronization links between them. For an
attack to be successful, its enough for just
one of those sites or links to contain a aw.
And sometimes there is no need to attack the
working infrastructure at all. Hard drives are
known to eventually fail - can you be entirely
sure that none of the failed hard drives,
scrapped and sent to a landll by the service
provider, wont end up in the wrong hands at
some stage of the disposal process?
Cloud: A safe or a cloakroom?
What I was trying to point out above is that
storing sensitive data in the cloud poses many
risks. So what, you ask, show me a busi-
ness with no risks at all. That makes sense.
As a matter of fact, the presence of risks and
their potential impact are of little importance
for a business. The only thing that matters is
the business ability to adequately mitigate
those risks.
Imagine that you went to a bank in the morn-
ing and deposited $1000 in cash in your bank
account. Later that day, two men in balaclavas
forced the cashier to hand them all the cash
including your $1000. Does it mean that the
bank will withdraw that $1000 from your ac-
count and leave you with nothing? No, be-
cause the bank is responsible for the money
you deposit with them. They take the respon-
sibility of storing your money safely, and take
on all the risks associated with this commit-
ment. The exact methods they use to deal with
the risks building underground bunkers, hir-
ing security staff, or transferring risks to insur-
ance companies - are not something you need
to care about.
Cloud services have that in common with
banks - the only difference is that its not your
money you hand to them, but your megabytes.
They take on the responsibility of providing the
maximal levels of data availability and protec-
tion. But what happens if your data is stolen or
destroyed? The scope of the cloud providers
responsibility is written down in their service-
level agreement (SLA), and Ill bet you wont
be happy when you nally open and read it:
You are responsible for taking your own
steps to maintain appropriate security, protec-
tion and backup of Your Content, which may
include the use of encryption technology to
protect Your Content from unauthorized ac-
cess...
NEITHER WE WILL BE RESPONSIBLE
FOR ANY COMPENSATION, REIMBURSE-
MENT, OR DAMAGES ARISING IN
CONNECTION WITH: ... ANY UNAUTHOR-
IZED ACCESS TO, ALTERATION OF, OR
THE DELETION, DESTRUCTION, DAMAGE,
LOSS OR FAILURE TO STORE ANY OF
YOUR CONTENT OR OTHER DATA.
Basically, cloud service providers disclaim any
responsibility for disclosing your data to any
other party, as well as for losing or damaging
it. They also suggest that you take your own
steps to maintain appropriate security and pro-
tection of your data. Therefore, mitigating the
corresponding risks becomes your own re-
sponsibility. A good solution would be to trans-
fer the risks to specialized insurance compa-
nies. Unfortunately, due to the novelty of the
cloud concept and apparent difculties in as-
sessment of nancial equivalents of such risks
as data loss or revelation, insurance compa-
nies struggle to offer any product suitable for
the cloud realities. I have high expectations
about insurance companies introducing proper
security quality standards for cloud services,
as it is difcult to imagine a better incentive for
a service provider to improve the quality of
their product than the cost of insurance premi-
ums paid by their customers.
And while insurance companies are working
hard on their new cloud-specic products (we
hope), users of cloud services must take care
of the security of their data themselves.
Whats in the box?
The majority of cloud service providers have
their own security measures in place. Lets
summarize what is on offer and what kind of
protection those measures actually provide.
The majority of providers offer the following
security instruments: secure TLS-driven data
transfer, user authorization and, sometimes,
server-side encryption (SSE). The scheme be-
low shows a typical route of data between the
customers environment and the cloud.
Picture 1. Data ow between the customers premises and the cloud storage.
The TLS protocol secures data transfers be-
tween two peers, or, in our case, between the
customers computer and the cloud front end.
The (correct) use of TLS gives the customer
the ability to establish the genuineness of the
service endpoint and protect the data ex-
changed from passive or active interception.
Each side of the protocol works like a black
box, which takes plain data from the applica-
tion layer and sends encrypted output to the
network. When receiving data, the black box
applies a reverse permutation to the encrypted
data and passes the decrypted data up the
stack. With the help of user authorization the
service identies a particular account holder
and conrms his identity. The authorization is
typically based on establishing the fact of
ownership of an authorization token. The most
widely used types of authorization tokens are
access keys, digital certicates and pass-
words. It is important to keep in mind that any
entity in possession of an authorization token
can send authorized requests to the service
on behalf of the legitimate account holder.
Therefore it is really important to keep authori-
zation tokens safe.
Server-side encryption (SSE) is an additional
layer of security adopted by certain cloud
service providers. Before sending the cus-
tomers data to the data centre, the provider
encrypts it with a strong symmetric cipher. If
an attacker gets access to the data centre,
they will be unable to decrypt the data without
having the encryption key.
The most critical point of this scheme is that
the key and the data encrypted with it reside in
the same environment, the cloud providers
network. Even though it may be stored in a
completely isolated subnet whilst the data re-
mains in the data centre, the key still meets
the data at point B where the encryption or
decryption is performed. Besides, the data in
unencrypted form is fed to the input of the TLS
server endpoint A, which normally is a front-
end web server, subject to the relevant attack
risks. In case an adversary gets access to one
of those points, they will be able to read the
customers data in the clear without any need
for the encryption key.
It should be noted that if your authorization
token is leaked, server side encryption wont
prevent data theft. With the token in his pos-
session, an adversary can easily impersonate
a legitimate user and get access to the data
through the services public REST or SOAP
interface.
Finally, server-side encryption wont protect
the data from the eyes of intelligence agen-
cies.
Ready. Steady. Go.
I hope I persuaded you that data of any posi-
tive signicance must always be covered with
an extra layer of protection in addition to the
security measures provided and adopted by
the service providers. One of the simplest yet
effective methods of building such an extra
layer is to adopt client-side encryption (CSE)
(in contrast to server-side encryption offered
by the service providers).
The main idea behind CSE is that the cus-
tomer encrypts their data before sending it
over to the service provider, and decrypts the
data after downloading it back. With CSE, the
customer keeps the encryption key securely in
his own environment rather than entrusting it
to the service provider. This way, the cus-
tomers data goes through the A and B points
in customer-encrypted form and cant be re-
covered by an adversary who controls those
points.
Even if a lucky attacker or intelligence person-
nel gain full control over the providers compu-
tational infrastructure, they will be unable to
recover the data.
Picture 2. A modied data ow with CSE in force.
Even if an adversary manages to steal the vic-
tim customers authorization token and imper-
sonates them to the service, the data returned
by the service will still be in encrypted form,
and they will be unable to recover it without
obtaining the CSE encryption key.
A variety of encryption keys can be used with
the CSE scheme, from generic symmetric en-
cryption keys and asymmetric key pairs up to
passwords and customers biometric informa-
tion. This exibility is achieved by separating
customers encryption keys (user keys) from
object encryption keys (session keys).
The data is rst encrypted with a randomly
chosen session key (SK). The session key is
then encrypted with a customers user key (for
example, a public RSA key). The encrypted
data together with the encrypted session key
is then sent to the cloud service provider.
When reading the data back, the customer
rst decrypts the session key with their user
key, and then uses the session key to decrypt
the data.
The session keys-based scheme has a num-
ber of advantages, the most substantial of
which is its ability to adopt user keys of virtu-
ally any nature and the uniqueness of per-
object encryption keys. The latter makes it im-
possible for an attacker who recovers a single
session key to decrypt the rest of encrypted
objects, as they are encrypted with different
session keys.
Picture 3. The CSE from the inside.
Aside from encryption, the customer may ac-
company protected objects with MDP (modi-
cation detection and prevention) records. An
MDP record is basically a message digest
computed by the writer over the unencrypted
data and encrypted with it before sending it
over to the cloud service. When reading the
data back, the customer computes a message
digest over the decrypted data and compares
it to the message digest attached to the en-
crypted object. If an MDP-ed object was al-
tered while residing in the cloud or in transit,
the message digests wont match.
In certain scenarios it makes sense to use
strong digital signatures instead of basic MDP
records to provide for a higher level of modi-
cation detection in multi-user cloud environ-
ments. In particular, strong signatures might
be useful for addressing proof of authorship
and non-repudiation tasks.
Another attractive side of the CSE is that one
can actually encrypt data with several different
user keys at the same time. Objects encrypted
in such way can be decrypted with any of the
keys used for encrypting them. This feature
can effectively be used to set up exible
access-rights schemes or establish secure
cloud-driven document ow within the organi-
zation.
Wait a minute, an attentive reader will say,
does it mean that the protection of my client-
side encrypted data wholly depends on the
encryption key? Correct. The encryption key
is a single point of access of the CSE scheme,
just like it is in any other properly designed
encryption scheme. Thats why adequate pro-
tection of encryption keys is a task of the
highest importance.
Per-object session keys are not an easy target
for attackers.
In fact, the only way for an attacker to recover
the session key is to guess it by trying different
keys one by one. This task can be really
tough; provided that the session key is of suf-
cient length and a cryptographically strong
RNG is used for generating it, guessing the
session key is an infeasible task. Thats why
the attacker is most likely to concentrate on
gaining access to the user key, which might be
much easier to achieve.
Operating systems and third-party vendors
provide a variety of mechanisms for securing
user keys, from protected operating system
areas to dedicated cryptographic hardware.
Some only give the key away after authenti-
cating the user who requests it; the others do
not export keys at all, instead performing the
decryption on board on behalf of the request-
ing user. In either case it is the responsibility
of the customer environment administrator to
dene and adopt proper key usage and stor-
age policies in order to minimize the risk of
revealing the keys to unauthorized parties.
Now that you understand the basics of client-
side encryption, we can consider the task of
integrating the mechanism into existing cloud-
driven infrastructures. As I described above,
this is mainly a matter of adopting appropriate
key management policies, techniques and
procedures, and those should be chosen
based on the role of the cloud storage service
in your infrastructure and the types of logical
links between the local environment and the
service.
The one-to-one links arise where access to
the storage is restricted to one or a xed small
number of the customers computers. These
links can often be employed in scenarios
where the cloud is used for storing large
amounts of data in a centralized way, mainly
as a replacement for classic relational data-
bases. Typical examples of one-to-one links
are those used in personal backup and syn-
chronization tools and links between the
worker and storage roles within computa-
tional cloud environments.
Where the number of one-to-one links is
small, or where there is no need for an extra
database abstraction layer, the encryption
keys can be stored directly on the computers
that access the storage. In larger organiza-
tions, access to the storage can be organized
through a small set of cloud gateways that
will be responsible for storing the keys and will
optionally provide a standard interface (e.g.
ODBC) for outer data storage.
Picture 4. Classic and cloud-driven infrastructures.
The many-to-one links arise in environments
characterized by widely decentralized access
to the cloud storage and volatile structure, and
are typical for scenarios where the cloud is
used for sharing or synchronizing data be-
tween different company users or depart-
ments. As the set of users approved to access
the storage is constantly changing with time,
encryption keys require a higher level of pro-
tection and manageability. This is where
enterprise-level hardware security modules
may be of great help. HSMs can be effectively
used for storing encryption keys securely and
enforcing individual access rights to them for
every user who needs to access the cloud.
Every change in a particular users position
can be easily put in practice by altering the
corresponding key access record on the HSM.
Picture 5. Cloud-driven infrastructure with advanced key management rules in force.
The CSE scheme may effectively be used for
managing access rights in environments with
complex information access rules by adopting
multiple encryption keys procedures. Each
logical group of users (e.g. developers,
colonels, board members) is assigned a
dedicated encryption key. Every object is then
encrypted with keys of all groups whose
members are allowed to access it. A multiple
keys approach can be exibly integrated into
existing corporate group policy rules.
Dollars and cents
How much will it cost to deploy the CSE
scheme in a live environment? The nal bill
will depend on the size of your infrastructure
and the chosen key management strategy.
Extra trafc and cloud storage capacity
costs are oating somewhere around zero. A
protected object is only up to a kilobyte larger
than a matching unprotected object, so from a
point of view of trafc and space economy it
virtually does not matter whether you work
with protected or unprotected objects.
Compensation of productivity decrease
caused by extra computational burden on the
customers resources. In practice, the burden
is imperceptible. Modern processors provide
encryption speeds of up to 8-10 MB/sec, much
faster than an average speed the cloud stor-
age providers are capable of accepting or giv-
ing away the data with. This means that a
computer that encrypts outgoing data on-the-
y will be idling for a certain amount of time
waiting for the preceding portion of encrypted
data to be delivered to the cloud front-end.
Expenses related to key management can
make a signicant share of the migration
budget, especially if many-to-one links are
present and HSMs are adopted to manage the
keys. The costs in this case directly depend on
the size and complexity of the infrastructure,
the type of encryption keys involved, and the
expected frequency and number of read/write
operations. The exact HSM to use should be
chosen based on the key management re-
quirements, and should be capable of han-
dling the estimated operation load.
The cost of implementing the CSE protocol
may pose another unpleasant surprise. The
current labour rate of qualied software devel-
opers, especially those with good knowledge
of cryptography, is not among the cheapest on
the software market, and a proper implemen-
tation of the CSE takes a signicant amount of
time. Happily, the market offers some out-of-
the-box software products that can help adopt
the CSE in faster, cheaper and user-friendlier
way.
Popular SecureBlackbox middleware from El-
doS offers cloud storage access components
that come with built-in support for CSE. The
product supports the majority of popular cloud
services, including S3, Azure, SkyDrive, Drop-
box and Google Drive. Besides the encryption
components themselves, SecureBlackbox of-
fers out-of-the-box support for a variety of en-
cryption key types, from easy-to-use
password-based keys to cryptographically
strong symmetric and asymmetric keys resid-
ing on hardware security devices.
The one-off asking price for the product is
fairly competitive and equals to 3-4 hours of
work of a qualied security expert - a good
bargain, taking into account that integrating
SecureBlackbox-driven CSE into an existing
infrastructure is an easy-to-do task that only
requires common programming skills.
Client-side encryption is an inexpensive and
easy-to-integrate method that addresses the
risk of private data revelation, topped up with
useful access control features. Even if
adopted in its simplest form, CSE provides
that extremely durable extra layer of protection
that your remotely deposited information
needs in todays world.
Ken Ivanov, PhD, is a Director and a Chief Security Expert at EldoS Corporation (UK), a provider of security
and le system solutions for software vendors. For more than 12 years Ken helped businesses, individuals
and governmental agencies all over the world to adopt information security measures. Being really passionate
about security and privacy, as well as liberal values, Ken is concerned about the increasing inuence of inter-
national intelligence services. Thats why the question of privacy for data residing in the cloud is of particular
importance to him. He can be contacted at [email protected].
Pim Tuyls is the Founder and CEO of Intrinsic-ID. Tuyls initiated work on
Hardware Intrinsic Security within Philips Research in 2002. As a principal
scientist, he managed the cryptography cluster in Philips Research in which
the initial research was carried out. Later he transferred this work to Intrinsic-
ID and headed the technology development.
Based on your experience, what strategic
errors do SMBs make when using/moving
to the cloud? Does security play a big role
in their decision-making?
SMBs mainly think about the benets of the
cloud: scaling and elastic IT-enabled capabili-
ties, higher productivity, leveraging a mobile
workforce, to mention a few. But they don't
make a proper risk analysis since they are not
aware of the possible consequences.
Breached customer data might lead to a law-
suit. By storing data unprotected in the cloud
specic laws and regulation might be violated.
Recently, we encountered a few cases where
SMBs were not aware that their back-up sys-
tem stores their data unprotected in the cloud.
This can lead to huge nes that could have
been avoided if an appropriate pre-study was
done and security measures had been taken.
Currently, security does not play a big role in
their decision-making process, but awareness
is growing. The revelations made by Edward
Snowden last year have clearly awakened
many people and organizations. In-house se-
curity expertise is usually missing in small or-
ganizations. It is important that they partner
with a security expert in time to make sure that
their transition to the cloud goes smoothly and
leads to success instead of a big headache.
What advice would you give to organiza-
tions that have mission-critical data in the
cloud? What critical steps should they take
in order to ensure ongoing security?
My recommendation to organizations with
mission-critical data stored in the could is to
conduct a risk and vulnerability prole to un-
derstand what kind of data they store in the
cloud; what type of cloud the data is stored in
(private versus public), and what the conse-
quences are if the data in the cloud gets ex-
posed. In all cases, we recommend imple-
menting the best possible security solution
that the budget will allow.
Once organizations understand what is at risk
they need to take measures to nd and im-
plement a security solution. We recommend a
solution that uses a two-factor authentication
system based on an independent hardware
root of trust, which will ensure that the data
stored in the cloud is properly encrypted be-
fore it leaves the device. By independent root
of trust we mean that the security keys should
not be inuenced nor known by any other
party, and should be generated within the de-
vice itself.
Early in the process of moving to the cloud,
companies should evaluate cloud security so-
lutions that are best suited to their needs and
in line with their budget. In most cases,
software-only solutions are not adequate be-
cause they have several weaknesses and the
highest protection is usually achieved with
hardware-based systems that offer client-side
protection. Client-side protection means that
the keys are in the hands of the customer and
cannot be learned by a trusted party.
SINCE ITS A CLOSED SYSTEM,
A PRIVATE CLOUD OFFERS
TEMPTING SECURITY ADVANTAGES
While security is still the most signicant
obstacle to cloud adoption, companies
also wonder if they should use a private or
a public cloud. What criteria should they
use to decide?
There are various ways to look at this, but cur-
rently we see the following main aspects.
What we see more and more in the market is
that companies tools have to be compatible
with those that their employees use in their
private life. Employees are accustomed to us-
ing their own mobile devices and even their
own (most often) public clouds. In that sense,
using a public cloud is attractive and compa-
nies often decide to use those.
However, this opens the door for all kind of
attacks on the companies network since there
is no control of the software installed on the
devices of employees. Such a policy has to be
handled with care and security solutions need
to be installed on the users devices.
Since its a closed system, a private cloud of-
fers tempting security advantages. For an out-
sider, its much more difcult to attack such a
cloud. But even in that case, a company still
depends on the trustworthiness of the owner
or administrator of the private cloud. Neverthe-
less, I think that the main consideration should
be that they use a system that guarantees that
they have full control of their data and abso-
lute privacy.
Even if a company is inclined to opt for a pri-
vate cloud, they should choose one that has a
security system in place (maybe via a partner)
one that ensures full protection of the com-
pany data so that even if/when the cloud pro-
vider is hacked, the privacy of the company
data is guaranteed.
Edward Snowden's revelations have im-
pacted on, among other things, the way
organizations think about their data in the
cloud. As a European company, have you
seen an uptick in business or concern?
What have been the reactions among your
clients?
Absolutely! We see that many organizations
are beginning to take data privacy very seri-
ous, especially in Europe and the US. They
have started projects to implement data pro-
tection systems or are setting up projects to
analyze their vulnerabilities. Some projects
aimed at further integrating the cloud have
been delayed till a clear picture of the vulner-
abilities and risks are in place.
OUR VISION IS TO ELIMINATE DATA
SECURITY AND USER AUTHENTICATION
ISSUES IN THE CONNECTED WORLD
This year almost all organizations, both large
and small, that we are in contact with are con-
sidering how to improve their security against
attacks and accidents.
Clearly, small companies are moving faster
with implementations of systems to protect
their data in the cloud, while large organiza-
tions need more time to build a vulnerability
and risk prole and depend more on consen-
sus building.
How does Intrinsic-ID enable organizations
to protect their cloud data? What makes
you stand out in the marketplace?
What Intrinsic offers customers is peace of
mind knowing that they are getting the
highest-level of security solution for the price
and in a way that is easy to implement. We
offer our customers a full solution to authenti-
cate to the cloud and protect data in the cloud
based on an independent root of trust that is
unique. It makes sure that nobody outside the
company can access the companys data.
Our solutions consist of two components:
software that runs on PCs or mobile platforms
and a physical token that implements the in-
dependent root of trust. This root of trust is
based on HIS technology or the electronic n-
gerprint of a chip. Since all chips are unique
due the deep-sub micron process variations,
this guarantees unique, very high quality keys
to start with: the keys have full entropy or ran-
domness and are therefore almost impossible
to guess even for governmental organizations.
On top of this, our approach makes sure that
our clients have full control over their security
and don't have to be nervous about the fact
that a backdoor might be in place that allows
organizations to listen in or break in to their
systems. Our secure cloud product not only
delivers security but also allows our customers
to work securely in the cloud. Therefore we
have implemented a secure sharing mecha-
nism: it allows users to share a document with
one or more colleagues without disclosing any
information to outsiders.
Finally, our solutions have been developed by
a team with in-depth security expertise in all
the aspects of a security system. The team
consists of hardware security specialists,
software security experts and system security
architects. This combination of expertise al-
lows us to build a strong, integrated security
product.
What are your plans for the rest of the
year? What areas are you focusing on?
Our vision is to eliminate data security and
user authentication issues in the connected
world, and we will continue to work to simplify
and improve methods for achieving that. We
intent to further extend our secure cloud offer-
ing to other platforms and enhance it with
more features that allow it to work together
with various mobile and cloud services, thus
providing a complete solution to our custom-
ers.
Mirko Zorz is the Editor in Chief of (IN)SECURE Magazine and Help Net Security (www.net-security.org).
Adequate protection of your digital data can be very straightforward. Just like
your own house, the best way to secure it is with a unique physical key that
you keep in your own hands.
Our virtual world keeps expanding and the
anytime, anywhere cloud is growing. More
and more, employees integrate their own de-
vices and favorite apps in their daily work.
Traditional security measures are no longer
enough, and coming up with good security so-
lutions for the cloud has been challenging.
CIOs realize that there is no single solution to
curb the increasing risks, and that a layered
approach is needed.
Encryption systems in the cloud
An age-old method for preventing sensitive
information falling into the hands of unauthor-
ized people is encryption. Also, concerning the
cloud, protection by encryption has become
an increasingly important activity, especially
when it comes to the mitigation of the risk of
breaches.
More and more, cloud providers, but also spe-
cialized companies, offer encryption/
authentication tools for data in or on the way
to the cloud. In Figure 1 I illustrate three cate-
gories of solutions: server-based, gateway-
based and client-based encryption systems.
a. Server-based encryption
In a server-based encryption system, the
server encrypts all data stored on the server.
This implies that a (master) key is hidden on
the server, too. Such a system offers a basic
level of protection, but it does not solve the
problem of secure cloud storage.
On the one hand, this system assumes im-
plicit trust in the owner of the server encryp-
tion key. On the other hand, when the encryp-
tion key is compromised in an attack, all the
data of all the users of the system are ex-
posed simultaneously.
Figure 1. Encryption/authentication solutions for the cloud can be divided in three categories, according to who
controls the security keys.
b. Gateway-based encryption
Gateway-based systems install a gateway
within the company. All company trafc that
goes to the outside world will be encrypted
with a key that is stored in the gateway.
The advantage of this system stems from the
fact that enterprise unique keys are used to
protect the data stored in the cloud. A suc-
cessful attack on the system of one enterprise
will not automatically translate to a successful
attack on the data of a neighboring enterprise.
But, it will reveal all the data of the enterprise
victim. Furthermore, there is still an implicit
assumption of trust in the owner of the keys
that are stored on the gateway.
c. Client-based encryption
Client-based encryption systems encrypt the
data at the client side. This means that the
keys are generated at the client side and all
les are encrypted on the client device before
being sent to the cloud. No trust has to be
given to another party for the generation and
the management of the keys.
Within this category there are software- and
hardware-based systems. In software-based
systems, the cryptographic keys are derived
from passwords that have to be provided and
remembered by the user.
Due to this human component (creating and
remembering a password with high random-
ness is difcult) the cryptographic keys will not
contain a lot of entropy. When cryptographic
keys are not sufciently random, there is a risk
that these keys will be compromised using
brute-force (guessing) attacks.
In hardware-based systems, the keys are
generated from physical randomness of the
hardware that the user owns: a smartcard,
USB stick or Micro SD card. This guarantees
the highest entropy level of the keys, i.e. they
cannot be brute-forced. A hardware-based cli-
ent side encryption system encrypts the data
at the client side with the strongest keys with-
out having to trust another party.
This solution gives control back to the user
and has a clear parallel with securing our
home in the physical world.
Strongest (unclonable) key in the hand of
the user
Cryptographic keys used in electronic devices
are traditionally stored in non-volatile memory
(typically secure EEPROM or E-fuses). How-
ever, this approach is sensitive to specic se-
curity issues, like the eavesdropping and tam-
pering of the keys.
Instead of storing keys in non-volatile memory,
it is nowadays possible to generate and store
secure keys based on unique physical proper-
ties of the underlying hardware.
This approach is called Hardware Intrinsic Se-
curity (HIS) and makes use of the concept of
Physical Unclonable Functions (PUFs). The
principle of a PUF can best be described as
biometrics for electronic devices and is illus-
trated in Figure 2.
The different threshold voltages for the tran-
sistors in SRAM result in a cell-unique start-up
behavior. Start-up values establish a unique
and robust ngerprint that is turned into a se-
cure secret key. Keys are not present in the
hardware when the device is not powered.
The fact that the keys are only generated
when they are required minimizes the window
of opportunity for an attacker to compromise
them.
The strength of this kind of hardware-based
key generation and storage can be combined
with a username and password-based login
system. This combination is called two-factor
authentication. This means that security is
based on two unrelated factors - something
you know and something you have.
Given that both of these factors are required
to access cloud data, it will become even
more difcult for an attacker to gain unauthor-
ized access to or manipulate the data. Two-
factor authentication provides incredibly
strong protection for all data in the cloud.
d. Secure data storage in the cloud
Secure data storage in the cloud can be
achieved by using symmetric key cryptogra-
phy based on, for example, AES. In symmetric
key cryptography les are encrypted and de-
crypted using the same key. I strongly rec-
ommend a key that is securely generated from
and stored in the hardware. Since encryption
and decryption are only performed on the cli-
ent side, this key never leaves the users
hardware and is therefore completely secure
from malicious use.
Due to the client-side encryption, les inter-
cepted during transmission or illegally ac-
cessed in storage are unusable.
Figure 2. The principle of Physical Unclonable Functions (PUFs): secret keys are extracted from the biomet-
rics of an electronic device.
Only the legitimate user (password) that is in
possession of the key (hardware) can down-
load the les stored in the cloud onto his de-
vice, after which they can be automatically
decrypted in the background.
e. Secure sharing of data in the cloud
Secure le sharing can be done using asym-
metric cryptography, also known as public-key
cryptography. In that case, a public/private
key-pair has to be generated during the instal-
lation of the hardware for each user. Although
different, the two parts of such a key pair are
mathematically linked. The public key is used
to encrypt les, while the private key is used
to decrypt les. While the private key never
leaves the hardware token of the user, the
public key should be stored in a public key da-
tabase (see Figure 3).
Since this public key cannot be used to de-
crypt data stored in the cloud, there is no se-
curity risk involved for the user when publicly
storing this key. Only with the securely-stored
private key can data from the cloud be de-
crypted.
Before sharing a le with another user in the
cloud, the public key of the receiving user
must be retrieved from the public key data-
base. The le is encrypted with the public key
before it is stored in the cloud in a location
that the recipient can access. After the recipi-
ent has been notied that a shared le is
available, the le can be downloaded from the
cloud onto the recipients device. On this de-
vice, where the recipients private key is avail-
able, the downloaded le can be decrypted
and used. Since the private key of the recipi-
ent is the only key that mathematically
matches the public key used for encryption,
the recipient is the only person that can de-
crypt this le. In case attackers are able to in-
tercept the le, they will not be able to decrypt
it because they lack the private key.
Figure 3. The ow of information for sharing encrypted data in the cloud.
The future
The cloud is a blessing for the modern econ-
omy, but recent events have also made it very
clear that it can easily turn into a nightmare if
security issues are not properly managed. In
order to avoid disasters and enable the
clouds true potential, security has to be an
integral aspect of any cloud-based system.
A strong two-factor authentication implemen-
tation as described in this article provides the
strongest security of data in the cloud without
compromising exibility, performance or ease
of use. Data can be safely available anytime,
anywhere from multiple devices.
HIS-enabled chips and tokens have the poten-
tial to offer top-level security and key man-
agement exibility to protect a whole new
class of applications. Apart from securing
payment transactions and provisioning media
content, these applications may include
machine-to-machine, smart grid, track-and-
trace and many others as they emerge from
the rapidly developing Internet of Things.
Dr. Pim Tuyls is the CEO of Intrinsic-ID, a provider of security IP cores and applications based on patented
Hardware Intrinsic Security (HIS) technology.
Prior to Intrinsic-ID, Pim was at Philips Electronics where he initiated work on Hardware Intrinsic Security as a
principal scientist within Philips Research. Several of Pim's papers relating to secure implementations of
Physical Unclonable Functions (PUF) technology have been published at security conferences.
Thousands of FTP sites
compromised to serve malware and
scams
Some 7,000 FTP sites and servers have been
compromised to serve malware, and its
administrators are usually none the wiser,
claim Hold Security researchers.
FTP sites function as online le caches and
are accessible remotely. Users who have the
required login credentials can upload and
download les from them, but other users can
also retrieve certain les hosted on such a
server if given a specic link that leads to the
le (and without needing to provide login
credentials). It is this latter capacity that
makes login credentials to FTP servers a
prized haul for cyber scammers, as they
upload malware and malicious links to the
server, then embed direct links to them in
spam emails delivered to potential victims.
Access to an FTP server can also be
occasionally leveraged by the attackers to
compromise connected web services.
"The victim companies hosting exploited FTP
sites are spread across the spectrum from
small companies and individual accounts with
ISPs to major multi-national corporations,"
noted the researchers.
"Hackers planted PHP scripts armed with
backdoors (shells) and viruses in multiple
directories hoping that these directories map
to web servers of the victim companies to gain
control of the web services. They also
uploaded HTML les with seamless re-directs
to malicious sites."
Alex Holden, the company's CISO, has
shared that among the compromised le
transfer servers are also some that belong to
The New York Times and UNICEF. Affected
organizations have been notied of the
problem and some have already moved to x
it. It is unknown who stole the FTP
credentials, and who is using them, but
judging by the complexity of some of the
passwords, it's natural to assume that they
haven't been guessed, but stolen via
information-stealing malware.
PoC mobile malware records swipes
on touch screen smartphones
A security researcher has developed proof-of-
concept malware capable of capturing
screenshots and nger swipes on mobile
devices, and is set to demonstrate his creation
at the RSA Conference in San Francisco.
As users inexorably switch from computers to
mobile phones and tablets for their online
shopping and banking needs, malware
developers will surely set their sights on
creating nancial malware adapted to the new
platforms.
Neal Hindocha, senior security consultant for
Trustwave, has proved that the feat can be
done, but his proof-of-concept malware does
not lend itself to industrial scale data
collection.
You see, this "touchlogger" is capable of
logging the coordinates of any swipe or touch,
and of taking screen captures, but it can't (yet)
distinguish between general use and use for
online payment and banking purposes.
Going through the swipe logs and screenshots
to search for relevant information is a very
time-consuming and labor-intensive task that's
difcult to automate, so using it against
specic targets seems more logical and more
efcient than infecting a huge number of
devices.
Hindocha has managed to make the PoC
code work on jailbroken iOS and rooted
Android devices, but shared with Forbes that
it's possible get it to work on non-rooted
Android devices as well - possibly by infecting
the device when it's plugged into a PC.
The malware might be aimed at capturing
nancial information, but since it records every
touch, an attacker would also discover
additional relevant information such as the
user's security code for unlocking the device,
image and alpha-numeric passwords, etc.
Java-based malware hits Windows,
Mac and Linux
Kaspersky Lab researchers have recently
analyzed a piece of malware that works well
on all three of the most popular computer
operating systems - the only thing that it
needs to compromise targeted computers is
for them to run a awed version of Java.
The Trojan is written wholly in Java, and
exploits an unspecied vulnerability
(CVE-2013-2465) in the JRE component in
Oracle Java SE 7 Update 21 and earlier, 6
Update 45 and earlier, and 5.0 Update 45 and
earlier.
Once the malware is launched, it copies itself
into the users home directory and sets itself
to run every time the system is booted. It then
contacts the botmasters' IRC server via the
IRC protocol, and identies itself via a unique
identier it generated.
The malware's main reason of existence is to
make the infected machine ood specied IP
addresses with requests when ordered to via
a predened IRC channel. The botmasters
simply have to dene the address of the
computer to be attacked, the port number, the
duration of the attack, and the number of
threads to be used in it.
At the time of analysis, the botnet formed by
machines "zombied" by this particular Trojan
was targeting a bulk email service.
Sophisticated cyber-espionage tool
uncovered
Kaspersky Lab discovered The Mask (aka
Careto), an advanced Spanish-language
speaking threat actor that has been involved
in global cyber-espionage operations since at
least 2007. The primary targets are
government institutions, diplomatic ofces and
embassies, energy, oil and gas companies,
research organizations and activists. Victims
of this targeted attack have been found in 31
countries around the world from the Middle
East and Europe to Africa and the Americas.
The main objective of the attackers is to
gather sensitive data from the infected
systems. These include ofce documents, but
also various encryption keys, VPN
congurations, SSH keys (serving as a means
of identifying a user to an SSH server) and
RDP les (used by the Remote Desktop Client
to automatically open a connection to the
reserved computer).
The campaign was active for at least ve
years until January 2014. During the course of
Kaspersky Labs investigations, the C&C
servers were shut down. 380 unique victims
between 1000+ IPs have been counted.
The complexity and universality of the toolset
used by the attackers makes this cyber-
espionage operation very special. This
includes leveraging high-end exploits, an
extremely sophisticated piece of malware, a
rootkit, a bootkit, Mac OS X and Linux
versions and possibly versions for Android
and iOS. The Mask also used a customized
attack against Kaspersky Labs products.
Among the attacks vectors, at least one
Adobe Flash Player exploit (CVE-2012-0773)
was used. It was designed for Flash Player
versions prior to 10.3 and 11.2.
The malware intercepts all the communication
channels and collects the most vital
information from the infected system.
Detection is extremely difcult because of
stealth rootkit capabilities.
Fully functional trojanized FileZilla
client steals FTP logins
Trojanized versions of the popular FileZilla
FTP client are being offered to unsuspecting
users via hacked websites with fake content.
"Malware installer GUI is almost identical to
the ofcial version. The only slight difference
is version of NullSoft installer where malware
uses 2.46.3-Unicode and the ofcial installer
uses v2.45-Unicode. All other elements like
texts, buttons, icons and images are the
same," Avast researchers warn.
"The installed malware FTP client looks like
the ofcial version and it is fully functional! You
cant nd any suspicious behavior, entries in
the system registry, communication or
changes in application GUI."
It's interesting to note that one of the
malicious versions has been compiled way
back in September 2012, and is still detected
by just a couple of commercial AV solutions.
Another one dates back to September 2013,
and is also poorly detected.
"We assume that the stolen FTP accounts are
further abused for upload and spread of
malware. Attackers also can download whole
webpage source code containing database
log in, payment system, customer private
information etc," the researchers pointed out.
Researchers uncover months-old
POS malware botnet
With the Target and Neiman
Marcus breaches, the topic of
malware that collects card
data directly from Point-of-
Sale devices has received
renewed interest. The PoS
malware used in the former has
been identied as a modied version of the
BlackPOS malware, but there are other similar
ones currently in use out there.
RSA researchers have discovered the entire
server infrastructure used in a global PoS
malware operation that targets retailers in the
US, Russia, Canada and Australia, and have
managed to access part of it. The malware in
question is the ChewBacca Trojan, which is
capable of logging keystrokes and scraping
the memory of PoS systems and the card
magnetic stripe data they contain.
"RSA observed that communication is handled
through the TOR network, concealing the real
IP address of the C&C server(s), encrypting
trafc, and avoiding network-level detection,"
they noted. "The server address uses the
pseudo-TLD '.onion' that is not resolvable
outside of a TOR network and requires a TOR
proxy app which is installed by the bot on the
infected machine."
"The ChewBacca Trojan appears to be a
simple piece of malware that, despite its lack
of sophistication and defense mechanisms,
succeeded in stealing payment card
information from several dozen retailers
around the world in a little more than two
months," the researchers noted.
"Retailers have a few choices against these
attackers. They can increase stafng levels
and develop leading-edge capabilities to
detect and stop attackers (comprehensive
monitoring and incident response), or they can
encrypt or tokenize data at the point of
capture and ensure that it is not in plaintext
view on their networks, thereby shifting the
risk and burden of protection to the card
issuers and their payment processors."
Android bootkit infects 350,000
devices
The rst ever Android Trojan with bootkit
capabilities has been discovered and
analyzed by Dr.Web researchers, who warn
that the malware is already operating on some
350,000 mobile devices around the world. The
malware - dubbed Oldboot - resides in the
memory of infected devices and launches
itself early on in the OS loading stage, they
say, and believe that the Trojan is being
distributed via modied rmware. To ensure
persistence, the attackers have inserted one
of the Trojan's components into the boot
partition of the le system, and have altered
the script that is tasked with initializing the OS
components.
"When the mobile phone is turned on, this
script loads the code of the Trojan Linux-
library imei_chk, which extracts the les
libgooglekernel.so and GoogleKernel.apk and
places them in /system/lib and /system/app,
respectively," the researchers explained.
"Thus, part of the Trojan Android.Oldboot is
installed as a typical application which further
functions as a system service and uses the
libgooglekernel.so library to connect to a
remote server and receive various commands,
most notably, to download, install or remove
certain applications."
Even if other elements of the Trojan are
removed successfully, the modied script will
restart the installation process by triggering
the imei_chk each time the device is rebooted.
Beware of malicious specialized
software keygens
Masquerading malware as key generators for
popular games is a well-known malware-
delivery tactic, but it's not often that you see
malicious keygens for other types of software.
Nevertheless, it happens occasionally. Trend
Micro researchers warn that malware peddlers
have lately been targeting with this approach
professionals working in a variety of
industries. They have spotted fake generators
for specialized (and expensive) engineering
(Aveva) and automotive repair (AllData)
software, multimedia tools (Bigasoft),
benchmarking software (Geekbench),
software for chemists and biologists
(CambridgeSoft), computing software
(Wolfram Mathematica) and, yes, some
games.
Unfortunately, the offered executables are not
what they seem. Once installed, they pave the
way for other malicious software to be
installed on the compromised computer, and
lately that software is often a fake AV variant.
"Fake antivirus software has declined
signicantly from its heyday several years ago
(in part due to crackdowns on their payment
systems)," the researchers pointed out. "Since
then, it has been overshadowed by rst police
ransomware and then in more recent months
by CryptoLocker."
Mac Bitcoin-stealing Trojan lurks on
download sites and GitHub
CoinThief, the recently discovered Bitcoin-
stealing Trojan that targets Mac users, has
been spotted being offered on several
download websites such as CNET's
Download.com and MacUpdate.com, as well
as masquerading as precompiled binaries in
several GitHub projects.
The malware's initial variant installs browser
extensions for Safari and Google Chrome that
monitor all web browsing trafc, looking
specically for login credentials for many
popular Bitcoin websites as well as Bitcoin
wallet sites and login credentials. These
newer variants have already been made to
include also a browser extension for Firefox
("Pop-Up Blocker 1.0.0").
"The malware is being distributed disguised
as price tickers for Bitcoin ("Bitcoin Ticker
TTM for Mac") and Litecoin ("Litecoin Ticker"),
which have been available on download.com
since early December. According to the
download stats, the malware has been
downloaded 57 times," SecureMac
researchers noted. Fortunately, the two
websites have already reacted and removed
the malware.
In a Reddit thread initiated by Nicholas
Ptacek, lead developer at SecureMac, the
developer of Bitcoin Ticker TTM has noted
that his original app was never open source,
so it seems like his app was never trojanized,
and that only its and his name was used to
trick users into downloading the malware.
Ptacek also shared that the malware is being
distributed on GitHub in the BitVanity and
StealthBit projects and wrote in details about
how to remove the malware from the system if
you have been infected.
Still, it would be probably wrong to assume
that the malware is not still being distributed
on other download sites and under different
names, so be careful when downloading
anything, and check for the malicious
extension.
The popularity of Bitcoin and other digital/cryptographic currency cannot be
denied. Different users like using it for different reasons, but many agree that
the question of keeping their stash safe is something that occasionally keeps
them up at night.
With the currencys rising popularity, different
services started popping up to help solve that
dilemma. Still, some chose to keep their digital
money on their own devices. But what hap-
pens if these devices break down, get cor-
rupted, or get accidentally erased?
Chris Bross, Senior Enterprise Recovery En-
gineer at well-known data recovery rm
DriveSavers, has recently been involved in a
few cases where customers needed their digi-
tal currency recovered.
In the past, we have typically been asked to
recover user-created data les like photos,
QuickBooks, videos, etc. Now we are being
asked to recovery digital currency, he shared
with (IN)SECURE Magazine. We did not par-
ticularly start offering recovery services to Bit-
coin users, customers just came to us begging
for help.
Usually, the value of the data on any device is
gauged by the user to whom it belongs. The
recovery process can be relatively expensive,
but when Bitcoins are at stake, the value of
the recovery makes it even more worthwhile.
So far, the company has had 8 requests for
digital currency recovery, 7 for Bitcoin and 1
for Litecoin.
In general, hardware failures of the storage
devices are more common, but in many of
these cases the users were completely at fault
because they deleted the data, and in two
cases, even overwrote it, he explained.
These two cases ended with the experts re-
covering the wallet.dat les, but the encrypted
data was unfortunately corrupted beyond re-
pair.
Having said that, no situation is entirely hope-
less and we always want to attempt the data
recovery, as early as possible after the data
loss event has occurred, to prevent the user
from making a poor choice that leads to
additional loss," he notes.
While the company usually deals more with
companies needing to recover data, all of their
requests for Bitcoin recovery so far have been
from individuals.
"In almost all cases, the user was keeping
only a single copy of their wallet.dat le on a
single storage device," says Bross. "Three of
the eight total cases were stored on a SSD
(sold state drive), while the others were on
HDDs. In all cases, the users did not trust the
available cloud backup or other options to rep-
licate their critical Bitcoin data."
In one particular case where the customer
stored his digital wallet on his Microsoft Sur-
face Pro tablet, performed a number of steps
that led to data loss AND disabled the com-
puter's ability to boot up, they had to develop a
new method to be able to create an image of
the SSD without physically removing it from
the tablet.
"Since it was an SSD, there were concerns
about BGC (background garbage collection)
processes that would erase and sanitize the
missing data as a normal function of system
maintenance. We needed to intercept and halt
those processes to mitigate any additional
loss of data," he shared.
"Do you also offer the service of tracing and
recovering stolen Bitcoin?" I asked, and re-
ceived a negative answer. "Although, we do
offer forensic data recovery services for any
case that may end up in litigation or prosecu-
tion," Bross noted.
Finally, I asked him to share some tips on how
to keep one's digital currency safe.
"Print it on paper. Ironically, in these modern
times of digital everything, paper and ink are
more reliable and last longer than magnetic or
solid state storage," he offered.
"In addition, keep multiple copies of the les
on other secured and encrypted media that
you personally control yourself. If you choose
to store this data in the cloud, be well aware of
potential cloud security vulnerabilities and
risks."
Zeljka Zorz is the Managing Editor of (IN)SECURE Magazine and Help Net Security
(www.net-security.org).
Talk of Big Data seems to be everywhere these days. A simple web search
for the phrase returns countless vendors offering solutions to the challenge,
myriad events discussing the topic, and blogs, articles, and forums lled with
talking points.
Amidst this sea of information, organizations
are starved for practical knowledge they can
leverage internally to make some sense of the
Big Data puzzle in order to meet their chal-
lenging operational requirements.
This article aims to provide some practical
guidance that organizations can leverage to
begin to navigate their way through the Big
Data challenge.
Data value vs. data volume
Distilled to its essence, Big Data is about two
symbiotic, but somewhat diametrically op-
posed components: collection and analysis.
Both are equally important, but the more data
one collects, the more difcult analysis be-
comes due to the volume and variety of data.
Much of this article will discuss ideas relating
to the analysis component of this challenge,
but a brief discussion regarding collection is
helpful to frame the discussion. As the readers
of this article are aware, the proper instrumen-
tation of an enterprise network for network se-
curity monitoring soon results in both an over-
whelming variety of data sources and a nearly
unmanageable volume of total data.
As one might expect, different data sources
carry different value for security operations.
For example, proxy logs, with their rich meta-
data and relatively compact size provide high
value to security operations.
Conversely, logs from routers or switches,
though helpful for network operators, are sig-
nicantly more voluminous but provide less
value for security operations purposes.
When architecting a log collection strategy,
most organizations do not assess the value of
each data source to security operations and
instead resort to collecting every data source
they can obtain from the enterprise. This ap-
proach produces a few cascading effects:
More data requires more storage, which
adds cost.
In place of additional storage, organizations
typically cut the retention period of logs (e.g.,
from six months to three months), which hin-
ders the ability to perform historical forensics
The additional volume of data creates a
larger volume of noise, which in turn makes
identifying anomalous trafc more difcult
Performance slows down (e.g., queries re-
turn less quickly), which in turn reduces pro-
ductivity and efciency.
Additionally, regardless of the actual length of
the retention period, most organizations apply
a uniform retention period. In other words, the
policy is that after N days (whatever N is), all
data rolls off to archive and ceases to be ac-
cessible as part of day to day operations.
Another benet of assessing data value is that
organizations can better understand which
data sources should be retained longer than
others.
When attempting to manage enormous volumes of data effectively, its important to
streamline and optimize workow at every opportunity.
Streamlining workow
When attempting to manage enormous vol-
umes of data effectively, its important to
streamline and optimize workow at every op-
portunity. It is most effective to facilitate the
organizations human resources (most often
the scarcest) working smarter, not harder.
How to achieve this in practice requires more
details than this article allows for, but some
general guidelines are provided here:
Develop efcient processes for all functions
(e.g., incident response for a malicious code
incident)
Automate within the process where possible
(not for automations sake)
Enrich data where possible to further opti-
mize the process (not for enrichments sake)
Focus analysts on one centralized work
queue/alerting stream
Enable/facilitate rapid and efcient
investigation/analysis with the end goal of
resolution (no analysis for analysis sake).
Integrating actionable intelligence
Intelligence can greatly add to the both the de-
tection and response capabilities of the enter-
prise. There are a lot of sources (whether free
or by subscription) referring to themselves as
security intelligence sources. In order for in-
formation to be intelligence, it must be timely,
reliable, high delity, and actionable.
In order for information to meet these re-
quirements, it must have context and a use
case. Organizations use the context and use
case to best leverage the intelligence. For ex-
ample, integrating intelligence regarding mali-
cious email attachment MD5 hashes into the
alerting stream requires a much different ap-
proach than integrating intelligence regarding
command and control (C2) URL patterns/
substrings.
If the intelligence comes at a relevant time
(timely), does not produce high volumes of
false positives (reliable), is of high quality
(high delity), and produces relevant alerting
(actionable), it can be leveraged according to
its specic context and use case.
Communal presence
The security community is a relatively small
and tight-knit community built almost entirely
on trust. Many of the relationships and the
most trusted information-sharing exchanges
within the industry are built on a strong foun-
dation of trust. These exchanges require a
give and take sharing information and meth-
ods, as well as receiving information and
methods. The information and methods ex-
changed this way are most often of extremely
high quality. A strong presence in the commu-
nity, particularly by the organizations leader-
ship can greatly aid in obtaining practical,
timely, and high quality information and meth-
ods to assist with the Big Data challenge.
Conversely, leadership that maintains a strong
presence in the community is more likely to
motivate the organization to be an active par-
ticipant in trusted information sharing ex-
changes.
Remembering the user
While systems get infected, users create, edit,
use, and share data. In many sophisticated
attacks, the adversary is after the organiza-
tions sensitive, proprietary, and/or condential
data. With the volume of data (legitimately)
entering and leaving the network on a daily
basis, it is quite difcult to identify anomalous /
malicious activity within this data. Most con-
temporary analysis techniques are IP (or do-
main) centric. If the vantage point (for some
analytical techniques) is shifted to a user-
based perspective, additional possibilities
emerge. When examining human beings, as
opposed to machines, it becomes much easier
to analyze activity (both historically and in
near real-time) to identify trends and depar-
tures from normal behavior. This enables the
enterprise to build richer analytical techniques
and alerting.
Summary
The challenge of Big Data is maturing from a
marketing buzzword to a reality confronting
large enterprises. The volume of data brings
new challenges, but it also brings new tech-
nologies, opportunities, and capabilities. Big
Data reminds us to consider both aspects of
collection and analysis, where one necessi-
tates the other. As new technologies emerge,
it is important to remember that they will help
with the challenge, but they alone cannot
solve the problem. Smooth and seamless in-
tegration of people, process, and technology
is just as important as always.
Joshua Goldfarb (www.yourcyberanalyst.com) is a cyber security analyst with over a decade of experience
building, operating, and running SOCs. Before joining nPulse Technologies (www.npulsetech.com) as its CSO,
Goldfarb worked as an independent consultant, applying his analytical methodology to help enterprises build
and enhance their network trafc analysis, security operations, and incident response capabilities to improve
their information security postures. Earlier in his career Goldfarb served as the Chief of Analysis for US-CERT.
While Apache Hadoop and the craze around Big Data seem to have exploded
out into the market, there are still a lot more questions than answers about
this new environment. One of the biggest concerns, second perhaps only to
ROI, is security.
This is primarily due to the fact that many
have yet to grasp the paradigm shift from tra-
ditional database platforms to Hadoop. Tradi-
tional security tools address a separation of
duties, access control, encryption options,
and more, but they are designed for a struc-
tured, limited environment, where data is
carefully collected and cultivated.
Hadoop, on the other hand, is an environment
with limited structure, high ingestion volume,
massive scalability and redundancy, designed
for access to a vast pool of multi-structured
data. Whats missing is new security tools to
match.
Another challenge with securing Hadoop
comes from the rapid expansion of the envi-
ronment itself. Since its initial development,
new tools and modules have been coming out
not only from Apache, but nearly every other
third-party vendor as well. While security is
tested and implemented for one module, three
more have come out and are waiting for the
same treatment. This makes it very difcult to
create an overall security architecture for the
entire Hadoop ecosystem as it continues to
grow. However, some security tools have
been released over the last few years, includ-
ing Kerberos, which provides strong authenti-
cation. But Kerberos does little to protect data
owing in and out of Hadoop, or to prevent
privileged users such as DBAs or SAs from
abusing the data. While authentication re-
mains an important part of the data security
structure in Hadoop, on its own it falls short of
adequate data protection.
Another development was the addition of
coarse-grained volume or disk encryption,
usually provided by data security vendors.
This solved one problem (protecting data at
rest) but considering one of the primary goals
behind Hadoop is using the data, one might
suggest that it provided little in the grand
scheme of Big Data security. Sensitive data in
use for analytics, traveling between nodes,
sent to other systems, or even just being
viewed is subject to full exposure.
Up until recently, Big Data technology vendors
have often left it to customers to protect their
environments, and they, too, feel the burden
of limited options.
Today, vendors such as Teradata, Horton-
works, and Cloudera, have partnered with
data security vendors to help ll the security
gap. What theyre seeking is advanced func-
tionality equal to the task of balancing security
and regulatory compliance with data insights
and big answers.
The key to this balance lies not in protecting
the growing ecosystem, or blanketing entire
nodes with volume encryption, but targeting
the sensitive data itself at a very ne-grained
level, with exible, transparent security. Apply-
ing this security through a comprehensive
policy-based system can provide further con-
trol and additional options to protect sensitive
data, including multiple levels of access to
various users or processes. Once secured,
the data can travel throughout the Hadoop
ecosystem and even to outside systems and
remain protected.
The options for ne-grained data security in
Hadoop now include encryption (AES or
format-preserving), masking, and Vaultless
Tokenization.
Typically, encryption is the least desirable op-
tion, as standard strong encryption produces
values that are unreadable to the tools and
modules in Hadoop, format-preserving en-
cryption is typically much slower than masking
or Vaultless Tokenization, and both require
complicated cryptographic key management
across tens or even hundreds of nodes.
Masking was developed for non-production
systems and testing, and has found a home in
Hadoops early, experimental phase. Individ-
ual data elements are either replaced with
random values or generalized so that they are
no longer identiable. It is fast, produces val-
ues that are readable to systems and proc-
esses, and requires no key management.
However, because masking was designed for
non-production, it is usually not reversible,
and is therefore not ideal for any situations
where the original data may be needed some-
time after the data is masked.
Vaultless Tokenization, similar to masking,
also replaces data elements with random val-
ues of the same data type and length. It is
also much faster than format-preserving en-
cryption, virtually eliminates key management,
and is transparent to processes. The added
benet comes from the ability to perform both
one-way protection and reversible security.
This provides ideal protection for test/dev en-
vironments and can also allow retrieval of the
original data when required by authorized
users or processes.
THE OPTIONS FOR FINE-GRAINED DATA SECURITY
IN HADOOP NOW INCLUDE ENCRYPTION, MASKING,
AND VAULTLESS TOKENIZATION.
Due to the read-only nature of the Hadoop
environment (les cannot be updated, you
can only create a le, read it and delete it),
application of these ne-grained protection
methods requires a unique approach.
This is typically performed in one of two ways.
The rst is a secured gateway, situated in
front of Hadoop, which parses incoming data
to identify sensitive data elements, and ap-
plies the selected protection method before
passing the data on to Hadoop. The second is
a secured landing zone, which may be a node
or partition within Hadoop that is protected
with coarse-grained encryption. Files arrive in
the landing zone, and are then parsed by one
of the processing applications in Hadoop
(MapReduce, Hive, Pig, etc.), identifying and
protecting sensitive data elements before in-
gesting the data into the main Hadoop cluster.
This method utilizes the massively parallel
processing of Hadoop to efciently protect
data.
In the next ve years, the creation of data by
more and more people and devices will con-
tinue to drive the companies towards Hadoop
and other Big Data platforms. The require-
ments for handling extreme levels of volume,
velocity, variety, and veracity will only in-
crease, and Big Data will assume more and
more critical business functions.
As the environment becomes more estab-
lished, usability and enterprise integration will
improve, new data exchange protocols will be
used, and a set of security tools will be stan-
dardized and made native to platforms.
Laws and regulations relating to privacy and
security will also continue to increase, and
security will become an even more vital com-
ponent in Big Data. Companies will be unable
to harness the massive amounts of machine-
generated data from the Internet of Things
without implementing comprehensive data
security - rst in the area of industrial envi-
ronment (power grids, etc.) and later on con-
sumer use (healthcare, etc.). Security will be
viewed not only in terms of loss-prevention,
but value creation, enabling compliant data
collection, use, analysis, and monetization.
Big Data security will evolve, becoming in-
creasingly intelligent and data-driven in its
own right. We will see more tools that can
translate security event statistics into action-
able information. Data security policies will be
intricately designed, and likely multi-layered,
utilizing a combination of coarse- and ne-
grained security methods, access control,
authentication, and monitoring.
In the exciting near future, the data is only
getting bigger, but we must not allow it to out-
grow security.
Ulf T. Mattsson is the CTO of Protegrity. Ulf created the initial architecture of Protegritys database security
technology, for which the company owns several key patents. His extensive IT and security industry experi-
ence includes 20 years with IBM as a manager of software development and a consulting resource to IBM's
Research and Development organization, in the areas of IT Architecture and IT Security. Ulf holds a degree in
electrical engineering from Polhem University, a degree in Finance from University of Stockholm and a mas-
ter's degree in physics from Chalmers University of Technology.
Information security continues to be an increasing problem because of the
volume, velocity and variety associated with Big Data. This article summa-
rizes the nature of information security risks and describes how better infor-
mation stewardship based on information centric security is essential to
manage these risks.
The term Big Data is as much a reection of
the limitations of the current technology as it is
a statement on the quantity, speed and variety
of data being generated. Big Data needs to be
understood as data that has greater volume,
variety or velocity than can be comfortably
processed using the technology that we al-
ready have.
Big Data comes from a number of sources,
both internal and external. Many organizations
have accumulated large amounts of data that
is not being exploited.
There is an even larger amount of data that is
held in publicly available sources, like gov-
ernment databases, and social media. In addi-
tion to this, the inbuilt instrumentation of smart
systems generates a massive amount of still
untapped data.
To realize its potential value Big Data needs to
be transformed into Smart Information, which
can then be used to improve planning and in-
crease efciency as well as to create new
kinds of products.
Information security challenges
The underlying information security chal-
lenges of malice, misuse and mistake apply
equally to Big Data. Big Data techniques can
also be used by criminals to improve their ex-
ploits, provide insight that facilitates security
breaches, and aggregate data to assist with
identity theft. Big Data can be misused
through abuse of privilege by those with ac-
cess to the data and analysis tools; curiosity
may lead to unauthorized access and informa-
tion may be deliberately leaked. Mistakes can
also cause problems - corner cutting could
lead to disclosure or incorrect analysis.
There are three major risk areas that need to
be considered:
Information lifecycle: Big Data turns the
classical information lifecycle on its head. The
provenance of the data may be doubtful, the
ownership of the data may be subject to dis-
pute, the classication of the information dis-
covered may not be feasible until after analy-
sis. For all of these reasons the compliance
requirements and needed controls cannot
easily be predetermined.
Data provenance: Big Data involves absorb-
ing and analyzing large amounts of data that
may have originated outside of the organiza-
tion that is using it. If you dont control the
data creation and collection process, how can
you be sure of the source and the integrity of
the data? How do you know that you have the
right to use the data in the way that is being
planned? These points are dened very
clearly in a UK report on the use of smart me-
tering for power consumption by utility com-
panies.
Technology unknowns: The technology that
underlies the processing of Big Data was con-
ceived to provide massively scalable process-
ing rather than to enforce security controls.
While this is not a new phenomenon in the IT
industry, there has not been sufcient time for
the inherent vulnerabilities and security weak-
nesses to become manifest.
Looking after property that is not your own is
called stewardship. Information stewardship is
not a new term; it has been in use since the
1990s and covers a wide range of challenges
involved in managing information as a key or-
ganizational asset. These challenges include
the management of the whole information life-
cycle from ownership to deletion, as well as
aspects like business value, data architecture,
information quality, compliance and security.
The basic objectives of information security for
Big Data are the same as for normal data: to
ensure its condentiality, availability, and in-
tegrity.
To achieve these objectives certain processes
and security elements must be in place. There
is a large overlap with the normal information
security management processes, however
specic attention is needed in the following
areas:
Everyone is responsible: The unstructured
nature of Big Data means that it is difcult to
assign the responsibility to a single person.
Everyone in an organization needs to under-
stand their responsibility for the security of all
of the data they create or handle. This means
creating a culture of security.
Verication of data source: Technical
mechanisms are needed to verify the source
of the external data used (for example, digital
signatures).
System integrity: Good oversight and control
over the integrity of the systems used for
analysis is needed, and that also goes for
privilege management and change control. Be
careful to validate conclusions if you cant
explain why the results make sense, they
probably dont. Always build in a way to check
dont let Big Data lead you to stupid conclu-
sions.
Secure processing: Measures to secure the
data within the analysis infrastructure are
needed to mitigate potential vulnerabilities and
to secure against leakage. These could in-
clude disk level encryption and a high level of
network isolation. Big Data should be secured
in transit, preferably by using encryption, but
at minimum by using SSL/TLS. If the cloud is
being used to process Big Data, you must
know how to verify that it is secured.
Access management: Access to the analysis
infrastructure, the data being analyzed, and
the results should be subject to proper IAM
controls.
Audit: Logging and monitoring of activities on
the analysis infrastructure is crucial to proper
auditing.
Mike Small (CEng, FBCS, CITP) is a fellow of the BCS, a member of the London Chapter of ISACA Security
Advisory Group, and a senior analyst at KuppingerCole. Until 2009, Mike worked for CA where he developed
CAs identity and access management product strategy. He is a frequent speaker at IT security events around
EMEA.
Traditional business intelligence has generally targeted structured data that
can be easily parsed and analyzed, but advances in analytics methods now
allow the examination of more varied data types. New analytics tools and
methods are expanding the possibilities for how enterprises can derive value
from existing data within their organizations and from freely available external
information sources, such as Software as a Service (SaaS), social media and
commercial data sources.
These advances allow enterprises to make
better business decisions and increase com-
petitive advantage. But this renaissance of
analytics capability can also introduce addi-
tional technical and operational risk, so enter-
prises must weigh the technical and opera-
tional risk against the business risk that is as-
sociated with failure to adopt Big Data analyt-
ics.
As with any potential investment that is in-
tended to bring about an improvement in ef-
cacy or efciency of business activities, sev-
eral key elements must be well understood to
enable systematic strategic planning:
Anticipated returns and potential impacts to
competitiveness through adoption
Potential impact to the current operational
ecosystem
Opportunity cost for the investment (i.e.,
what else the enterprise might have invested
in instead)
Loss of value for investments already made.
Objective and systematic analysis of these
factors becomes increasingly challenging as
the industry hype that surrounds a new tech-
nology or business trend increases: hype can,
in some cases, create unfounded pressure to
adopt, or create barriers to adoption in others.
For technologists, the potential technical, op-
erational and compliance risk that is associ-
ated with maintaining and operating on a large
volume of potentially sensitive data is very ap-
parent; however, the business-relevant factors
that provided the initial impetus for adoption of
these analytics tools and methods may be
less apparent.
However, understanding the business case -
the rationale for adoption, the anticipated re-
turn that the business hopes to achieve, and
the competitiveness impact to the business if
the enterprise chooses not to adopt while its
competitors do adopt - is equally important.
For information security, audit and governance
professionals, lack of clarity about the busi-
ness case may stie organizational success
and lead to role and responsibility confusion.
Understanding the business case
Big Data refers to large, quickly growing or
varied types of information (high volume, high
velocity, and/or high variety information assets
that require new forms of processing to enable
enhanced decision making, insight discovery
and process optimization).
Big Data analytics is the application of emerg-
ing statistical, processing and analytics tech-
niques for Big Data for the purpose of advanc-
ing the business applying statistical models
and techniques to business information to de-
rive conclusions that are benecial to that
business.
Big Data analytics is particularly appealing to
many enterprises because, in many cases,
they have already made some investment in
both business analytics and the collection of
large data sets, on which analysis can be ap-
plied.
This means that the foundation from which to
draw new conclusions, explore new ways of
doing business and open up new avenues of
competitive advantage may already be in
place.
What is this competitive advantage speci-
cally? Some data suggests a direct correlation
between the use of Big Data analytics and
protability. For example, one study cites an
increase in overall protability of six percent
as a direct result of using Big Data effectively.
That metric, while appealing in the abstract,
lacks sufcient underlying context and level of
detail to be able to understand precisely how
that correlation is made. Specic case studies,
by contrast, provide a clearer picture of how
increases in competitiveness are achieved
and why these analytics techniques provide
value. These studies show the imaginative
uses to which pre-existing data are leveraged
as a result of better analytics techniques, and
the transformative impacts that are achieved.
What makes improvements for Big Data ana-
lytics particularly compelling from a business
perspective is that the data already exists.
Data exists about customers, such as pur-
chases they make and their receptiveness and
responsiveness to marketing efforts. During
the normal course of business, many enter-
prises collect large volumes of data about their
customerstheir habits, preferences, the
specics of individual transactions, fraud his-
tory, etc. When analyzed, this data allow en-
terprises to make changes and, subsequently,
measure the performance of those changes.
This allows those enterprises to dynamically
shift inventory and / or pricing in response to
consumer demand.
Furthermore, data analytics allows enterprises
to create better-targeted marketing cam-
paigns, to better measure the efcacy of those
campaigns, and to launch new products and
service offerings in response to customer de-
mand. From a business standpoint, invest-
ment in Big Data analytics is compelling be-
cause it leverages the otherwise latent or un-
used resource of already-collected data.
Not every enterprise will be equipped to make
use of Big Data analytic techniques. Some en-
terprises may be missing key skills in their ex-
isting personnel, or they may be missing criti-
cal portions of the technological ecosystem.
Also, the technical ecosystem may not be laid
out in a way that allows the techniques to op-
erate; and enterprises may lack the processes
to gain access to data and make use of the
intelligence they collect as a result of the ap-
plication of these methods.
Its important to nd out the exact current
situation, because investments made before
readiness is fully achieved may be inefcient,
suboptimal in terms of the results they pro-
duce or, as a worst case, may represent need-
less expense.
From a process standpoint, existing silos
should be evaluated to determine whether in-
dividual business units, departments and per-
sonnel are willing and able to share informa-
tion and act on information received. This con-
sensus needs to happen so that analysis can
be performed (disparate sources of data may
need to be consolidated to operate on them)
and so that the derived conclusions can be put
to productive use. Enterprises need to con-
sider that these areas may not share informa-
tion currently and may have a history (depend-
ing on the culture) of competitiveness, an-
tagonism, or resistance to outside inuence.
These cultural barriers can impede open and
collaborative exchange of important data ele-
ments and act as a barrier to adaptation in re-
sponse to conclusions drawn. This consensus
among silos can extend beyond the depart-
ment level and down to the level of individual
personnel. For example, key stakeholders
may not know precisely where key data ele-
ments reside or how to access those data
elements.
TECHNICAL AND OPERATIONAL RISK SHOULD
CONSIDER THAT CERTAIN DATA ELEMENTS MAY
BE GOVERNED BY REGULATORY OR
CONTRACTUAL REQUIREMENTS
Likewise, Shadow IT (technology adopted
without the direct oversight or, in many cases,
awareness of the IT organization) can compli-
cate information sharing because technology
adopted without centralized oversight may
represent a signicant repository of critical in-
formation, and lack of central awareness of
the information may limit the ability to include
it in the scope of analysis.
Because of these factors, some degree of or-
ganizational self-awareness is required to
think through supporting processes and iden-
tify potential problem areas before enterprises
undertake signicant investment in Big Data
analytics.
Lastly, the technology implementation plays a
role in determining organizational readiness.
In many cases, new tools are required to sup-
port the analysis to be conducted, and capa-
bilities for data storage and computation may
need to be evaluated to ensure sufciency.
Moreover, sufcient data on which to operate
need to exist and be accessible to analysts.
Data sources must be identied, which in-
volves locating structured data (e.g., data or-
ganized in a relational database) and unstruc-
tured data (e.g., data stored ad hoc on a le
system or in a loose collection.) Identifying
data sources can likewise involve data in a
variety of different formats, including video,
audio, images and text. Computational re-
sources may need to be expanded to enable
operation and analysis of these data.
Information security and audit practitioners
that are evaluating a Big Data analytics initia-
tive in their enterprise need to weigh the man-
agement and mitigation of the technology risk
of adoption against the business risk to the
enterprise if they choose to not adopt Big Data
analytics.
Technical and operational risk should consider
that certain data elements may be governed
by regulatory or contractual requirements and
that data elements may need to be centralized
in one place (or at least be accessible cen-
trally) so that the data can be analyzed. In
some cases, this centralization can compound
technical risk. For example:
Amplied technical impactIf an unauthor-
ized user were to gain access to centralized
repositories, it puts the entirety of that data in
jeopardy rather than just a subset of the data.
Privacy (data collection)Analytics tech-
niques can impact privacy; for example, indi-
viduals whose data is being analyzed may feel
that the revealed information about them is
overly intrusive.
Privacy (re-identication)Likewise, when
data is aggregated, semi-anonymous informa-
tion or information that is not individually iden-
tiable information might become non-
anonymous or identiable in the process.
These risk areas can cause some practitio-
ners to be understandably wary. However,
analytics efforts can be used to offset risk by
applying the tools and techniques to security-
event information, to transaction information
for the purpose of detecting fraud, or to other
technical information for risk-reduction pur-
poses.
Stockpiles of security-relevant information,
such as user and system activity, can be
logged and examined the same way as more
business-facing data can be logged and ex-
amined. The same analytics techniques and
tools that streamline and increase the quality
of business processes can likewise streamline
and increase the quality of other risk-
mitigation processes.
Tools purchased and analytics techniques that
are acquired to help enable business-facing
efforts can, with planning, be adopted by in-
formation security and risk management areas
to help advance their goals as well.
The security and audit practitioners consid-
eration of risk can and should be holistic.
If an enterprise elects not to employ these
techniques, there is a risk to the business, be-
cause competitors will capitalize on the oppor-
tunity.
This result could have ramications just as se-
rious to the enterprise as a security or privacy
breach or dreaded business continuity implica-
tions.
A holistic view of the risk in an enterprise
should seek to account for all sides of the risk
equation, including the following:
Business value of adoption
Business risk of non-adoption
Technical/security/privacy risk that may in-
crease depending on the implementation used
to support the Big Data analytics approaches
at the technical level
Possible risk-offsetting benets of the tech-
nology at the technical level
A number of business dynamics make the ap-
plication of new and better analytics appealing
to enterprises.
By looking at how these analytics techniques
are transforming enterprises in real-world sce-
narios, the value becomes apparent as enter-
prises start to realize dramatic gains in the ef-
ciency, efcacy and performance of mission-
critical business processes.
The business case is made even more com-
pelling by the fact that most enterprises al-
ready have in place the foundation for analysis
in the form of more data than they can cur-
rently use productively.
Most enterprises already retain a large
amount of data, such as information about
their customers, metrics about the perform-
ance of internal business processes, data
about information systems and their technol-
ogy ecosystem, transactional information
about sales and marketing and numerous
other data items about how they do business.
While some new areas of technical risk may
arise as a result of more voluminous and con-
centrated data, the business consequences of
not adopting Big Data analytics may outweigh
the technology risk.
Understanding this business case can help
security, audit and governance practitioners in
two ways: It helps them to understand the mo-
tivation and rationale driving their business
partners who want to apply Big Data analytics
techniques within their enterprises, and it
helps balance the risk equation so that techni-
cal risk and business risk are addressed.
Norman Marks is a member of ISACAs Emerging Business and Technology Committee.
As data grows exponentially, and with it the ability to use it effectively, how
can organizations ensure that the analysis of data sets containing sensitive
information doesnt result in a costly data breach? If an organization can ac-
cess this data and analyze it, then who else can access it?
Organizations are struggling to understand
how the collection of extremely large and
complex structured and unstructured data sets
can be protected. And it is clear some tradi-
tional approaches to database and application
security are not well suited to Big Data de-
ployments.
Imagine a supercomputer available in sec-
onds thanks to cloud computing where many
data sources are merged for analysis from
across an enterprise. It is easy to see how
sensitive information could be exposed or
created in such a scenario.
There is also a risk that human error could
combine the data or allow access by unau-
thorized users. Its fair to say that by its very
makeup the aggregation of multiple informa-
tion sources Big Data is sensitive informa-
tion. Stakeholders expect this information to
be secure in the hands of organizations that
use it. So, if Big Data is so large and complex,
how can it be protected?
The answer, as Google executive chairman
Eric Schmidt recently said, is to encrypt eve-
rything. Admittedly, Schmidt was responding
to the issue of government surveillance, but
he is nonetheless right. Encryption does pro-
tect your data, big or small, when a breach
occurs. Encryption ensures that your data re-
mains secure, regardless of where it resides
in the data center, the cloud, a mobile device
or even in the hands of hackers from enemy
organizations, nation states or malicious in-
siders.
The caveat here is that the encryption algo-
rithms need to be secured and the keys to de-
cipher the encryption need to be protected. In
addition, Big Data processes expect data in
real world formats so a complete data protec-
tion strategy has to include format-preserving
approaches such as tokenization.
Organizations must rst understand what to
protect (encrypt or tokenize) and then must
learn how to manage the encryption keys
used for this protection. Specically, how are
the keys generated? Where are they stored?
What is the strength (or size) of keys? How
often are keys changed? These are just some
of the questions that will help you build a
strong encryption strategy for protecting data.
A key management solution needs to allow a
comprehensive data protection strategy. Seg-
regating encryption and tokenization tech-
nologies and managing them individually is a
recipe for failure.
A security strategy is only as strong as its
weakest link. Solutions must also protect data
efciently. This means that the impact of the
encryption on business and cost must be
minimal. The solution should, preferably, be
transparent, especially to the end users who
are less likely to be concerned with security
and compliance issues.
Most users simply want to get their job done
quickly and without technical issues. So as
part of the protection approach and best prac-
tices, authentication and access controls need
to be seamlessly integrated to ensure that
only the correct users are accessing the right
information.
WHILE EVERYTHING IS BIGGER AND FASTER
WITH BIG DATA, ULTIMATELY ENCRYPTION
AND TOKENIZATION REMAIN CRITICAL
ELEMENTS OF PROTECTING SENSITIVE DATA
At the forefront of Big Data protection is the
Cloud Security Alliance. The group is speci-
cally working to address the security and pri-
vacy issues magnied today by the velocity,
volume, and variety of Big Data, such as
large-scale cloud infrastructures, diversity of
data sources and formats, streaming nature of
data acquisition and high volume inter-cloud
migration.
The CSA draws attention to this issue when it
states: Securing Big Data stores: this docu-
ment focused on using Big Data for security,
but the other side of the coin is the security of
Big Data.
The CSA highlights the way Big Data differs
from regular data: Big Data is differentiated
from traditional technologies in three ways:
the amount of data (volume), the rate of data
generation and transmission (velocity), and
the types of structured and unstructured data
(variety) (Laney, 2001).
While everything is bigger and faster with Big
Data, ultimately encryption and tokenization
remain critical elements of protecting sensitive
data, whether in transit across high speed
networks, stored encrypted in large volumes
of data at rest, or tokenized during processing.
Condential data is everywhere in every or-
ganization. Therefore a comprehensive and
holistic data protection strategy includes en-
cryption, tokenization, authentication and ac-
cess management, with the understanding
that key management is critical.
With great power comes great responsibility,
and that power must be used wisely to ensure
that the analytics are put to good use, that the
data will remain in the control of those people
who will do that, and that its protected from
those who may have other intentions.
Prakash Panjwani is Senior Vice President & General Manager, Data Protection Solutions at SafeNet
(www.safenet-inc.com).
RSA Conference 2014
www.rsaconference.com/helpnet
Moscone Center, San Francisco, CA, USA
24 February - 28 February 2014
InfoSec World Conference & Expo 2014
www.infosec-world.com
Disney's Contemporary Resort, FL, USA
7 April - 9 April 2014
Infosecurity Europe 2014
www.infosec.co.uk
Earls Court, London, UK
29 April - 1 May 2014
HITBSecConf2014 Amsterdam & HITB Haxpo
conference.hitb.org
De Beurs van Berlage, Amsterdam, The Netherlands
29 May - 30 May 2014

Insecure Mag 41

Uploaded by

Document Informationclick to expand document information

Copyright:

Available Formats

Insecure Mag 41

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Insecure Mag 41

Uploaded by

Copyright:

Available Formats

Big Data and cloud security have been some of the most discussed topics during the past

You might also like

Insecure Mag 41

Uploaded by

Document Informationclick to expand document informationMagazine related to computer security

Document Informationclick to expand document information

Copyright:

Available Formats

Insecure Mag 41

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Insecure Mag 41

Uploaded by

Copyright:

Available Formats

Big Data and cloud security have been some of the most discussed topics during the past

You might also like