Cloud Storage

Module 3
• Cloud storage means saving your data with an online service instead
of on your own computer.
• You access this data over the Internet, just like other cloud services.
• There are several advantages to cloud storage over traditional
methods.
• If your data is in the cloud, you can reach it from anywhere with
Internet access, which is great for people who travel a lot.
• Workers don't need to use the same computer or carry physical
storage devices to access their data.
• Additionally, if your organization has branch offices, everyone can
access the data from the cloud provider.
The Basics
• There are many cloud storage systems, each designed for specific
needs. Some focus on storing email or photos, while others can
handle any type of data. Providers range from small-scale operations
to massive facilities.
• At its simplest, a cloud storage system requires just one data server
connected to the Internet. Users upload files to the server, which
saves the data. To retrieve it, they access the server through a web
interface, which either sends the files back or lets them work with the
data directly.
• Most systems, however, use many servers to ensure reliability. Data is
stored on multiple machines for redundancy, so even if one server or
power supply fails, the data remains accessible.
• Many clients choose cloud storage not just for extra space but for
safety. It ensures their data is secure even if their physical location is
compromised.
Storage as a Service
• Storage as a Service (SaaS) refers to renting storage space from a third-
party provider, ideal for users who lack the budget, resources, or expertise
to manage their own storage systems.
• This service has grown popular due to the complexity of modern backup,
replication, and disaster recovery needs, especially for small and medium-
sized businesses.
• The main benefit of SaaS is cost savings.
• Users pay based on the amount of storage or data transferred, avoiding the
expense of building and maintaining storage infrastructure.
• Data is backed up to the provider’s servers using client software over a
wide-area network (WAN).
• In case of data loss, users can retrieve their data from the provider.
Providers
• There are countless cloud storage providers available online, with
new ones emerging regularly. While many offer general-purpose
storage, some focus on specific types of data. Below are examples of
specialized cloud providers, with details on major players to follow.
• Examples of Cloud Storage Providers
• Google Drive: Users can store, share, and collaborate on files, including
documents, spreadsheets, and presentations, with seamless integration across
devices.
• Web Email Providers: Services like Gmail, Outlook, and Yahoo Mail store emails
and attachments on their servers, enabling access from any Internet-connected
device.
• Google Photos and Apple iCloud Photos: Specialized in hosting and organizing
digital photos and videos, with features like AI-based search and automatic
backups.
• YouTube and Vimeo: Host millions of user-uploaded videos, offering features for
sharing, streaming, and monetization.
• AWS and Microsoft Azure: Provide robust cloud solutions, including storage for
applications, websites, and databases, catering to both businesses and
individuals.
• Social Media Platforms: Platforms like Instagram, Facebook, and LinkedIn store
user-generated content, including photos, videos, and documents, on their
servers.
• Dropbox and OneDrive: Offer versatile cloud storage for any type of digital data,
with tools for collaboration and synchronization across devices.
Security in Cloud Storage
• To protect data, cloud storage systems employ several security techniques:
• Encryption: Data is encoded using a complex algorithm, requiring an encryption
key for decoding. Although cracking encrypted data is theoretically possible, it
requires immense computing power, making it highly impractical for most
hackers.
• Authentication: Users must create a username and password to access the
system, ensuring only authorized individuals can log in.
• Authorization: Access levels are defined, allowing only specific users to access
certain data. For instance, a junior employee might have limited access, while IT
administrators have full access.
• Despite these measures, concerns about vulnerabilities remain:
• Hackers may find ways to bypass security and access sensitive data.
• Internal threats, like a disgruntled employee misusing their credentials, could
lead to data alteration or destruction.
• Maintaining robust security protocols and monitoring systems is critical to
addressing these risks.
• Reliability in Cloud Storage
• Reliability is a key concern for cloud storage users. An unreliable
system can lead to data loss or inaccessibility, making it a major
liability. Trust in a provider also hinges on the company's financial
stability and ability to maintain operations.
• Redundancy: Most providers mitigate reliability issues by storing data
across multiple servers to prevent loss in case of a crash. However,
the risk of outages or access problems still exists.
• Reputation: A provider’s reliability directly impacts its reputation.
Users avoid providers perceived as unreliable, and such companies
struggle to survive in a competitive market.
• Maintaining a strong track record of reliability and stability is critical
for providers to build trust and retain clients.
• Advantages of Cloud Storage
• Cloud storage is becoming a preferred solution for organizations due to its
flexibility and reliability.
• Data Accessibility: Data resides on the web across multiple storage systems
rather than a single hosting site. Providers balance server loads and move data
between datacenters, ensuring data is stored close to where it’s most needed for
quick access.
• Disaster Recovery: Storing data off-site protects it during disasters like fires or
hurricanes. This ensures business continuity, reducing downtime and potential
losses.
• Backup Solutions: Cloud storage eliminates the risk of losing critical information
when local backups are destroyed.
• Flexibility in Integration: Solutions vary in complexity. Some use agent-based
applications that automatically transfer files to the cloud, while others require
manual uploads through a web interface.
• Providers like Amazon S3 are widely known, but enterprises might choose
vendors offering tailored features like service level agreements (SLAs) and direct
customer support to suit their specific needs. These features are critical for
businesses relying heavily on cloud storage.
Cautions for Using Cloud Storage
• Adopting cloud storage requires a careful and strategic approach due to its evolving
nature.
• Start Gradually: Use cloud storage for noncritical purposes initially, avoiding full reliance
until its maturity and reliability are proven.
• Compatibility Issues: Large enterprises may face challenges with vendors like Google or
Amazon due to the need for application rewrites and limited portability. Vendors like
3tera may offer better compatibility with platforms like LAMP, Solaris, Java, or
Windows.NET.
• Key Concerns: Price and reliability remain major concerns. Vet vendors thoroughly to
ensure they provide quality service and clear data recovery policies, as mistakes can lead
to permanent data loss.
• Testing Before Commitment: Many businesses adopt a "test and try" approach by
experimenting with a few services to assess compatibility with existing IT systems.
• Legal Considerations: Storing copyrighted material such as music or videos may not
always be feasible due to licensing restrictions.
• Vendor Accountability: Understand the provider's terms regarding data maintenance,
recovery, and accountability in the event of data loss or security breaches.
• Redundancy is Key: A robust strategy involves multiple redundant systems, including
local backups, offsite backups, synchronization, and archiving.
• Being selective and implementing layered solutions helps mitigate risks and ensures
secure, reliable storage.
• Outages in Cloud Storage
• Organizations must be aware of the risks of storing data on the Internet, as
outages can disrupt access to critical information.
• Example of Outage: In February 2008, Amazon S3 experienced a major outage,
causing numerous client applications to go offline. Although no data was lost due
to Amazon's redundancy practices (storing multiple copies in different locations),
clients were unable to access their data during the outage.
• In July 2024, a faulty update from cybersecurity firm CrowdStrike caused a global
IT outage, impacting numerous organizations across various sectors. The incident
led to disruptions in airlines, media companies, and financial institutions,
highlighting the vulnerabilities associated with cloud-based services.
• Lessons Learned: While providers may respond to such issues by improving
systems, outages highlight the potential for disruptions.
• Key Takeaway: Cloud storage can be beneficial, but organizations should proceed
cautiously and ensure they have contingency plans in place to mitigate downtime
risks.
• Reliable access and redundancy are essential considerations when choosing a
cloud storage provider.
• Theft in Cloud Storage
• Storing data in the cloud comes with potential security risks, including
unauthorized access and theft:
• Data Breaches: When data leaves your own data center, it becomes
vulnerable to unauthorized access.
• Shared Storage Risks: Cloud providers often store data from multiple
clients together. This increases the risk of your data being
compromised, especially if a competitor's information is stored in the
same location.
• Mitigation: Encrypt your data before uploading it to the cloud and
use secure transmission protocols like SSL to protect it during transit.
• By implementing strong encryption and secure practices, you can
reduce the risk of data theft while leveraging cloud storage.
Cloud Storage Providers
• This section highlights some major cloud storage providers. While the
list isn't exhaustive, it includes key players in the industry to help you
evaluate their offerings and find what suits your needs.
• Amazon and Nirvanix lead the market, with other notable names like
Google, which is launching GDrive, EMC preparing a storage solution,
and IBM offering multiple options under Blue Cloud.
• Amazon Simple Storage Service (S3)
• Amazon S3, launched in 2006, is one of the most popular cloud
storage services. It simplifies web-scale computing for developers by
offering a reliable, scalable data storage infrastructure—the same one
Amazon uses for its global websites.
• Key Features:
• Flexible Storage: Store and retrieve objects ranging from 1 byte to 5
gigabytes, with unlimited total storage.
• Unique Keys: Each object is accessed through a unique developer-
assigned key.
• Access Control: Objects can be private or public, with customizable
user permissions.
• Standard Interfaces: Compatible with REST and SOAP, making it easy
to integrate with any Internet development toolkit.
Design Requirements of Amazon S3
• Amazon designed S3 to meet the following key requirements:
• Scalability:
S3 scales seamlessly to support unlimited storage, high request rates, and
numerous users, enabling web-scale applications.
• Reliability:
Data is stored with durability and 99.99% availability. Amazon ensures no
downtime in its operations.
• Speed:
Built for high-performance applications, S3 minimizes server-side latency.
Performance bottlenecks can be resolved by adding nodes to the system.
• Cost-Effectiveness:
S3 uses inexpensive commodity hardware, tolerating frequent node
failures without affecting system performance. Its hardware-agnostic
design helps reduce infrastructure costs.
• Simplicity:
Combining scalability, reliability, speed, and low cost with ease of use
ensures S3 is accessible for any application, anywhere.
 Design Principles
• Design Principles
• Amazon utilized the following distributed system design principles to meet the requirements of Amazon S3:
• Decentralization
• Fully decentralized techniques are employed to eliminate scaling bottlenecks and single points of failure.
• Autonomy
• Individual components are designed to make decisions based on local information without reliance on other components.
• Local Responsibility
• Each component is responsible for maintaining its own consistency, ensuring this burden is not placed on its peers.
• Controlled Concurrency
• Operations are designed to minimize or eliminate the need for concurrency control.
• Failure Toleration
• Component failures are treated as normal, allowing the system to continue operating with minimal or no interruptions.
• Controlled Parallelism
• Granular abstractions enable parallelism, improving performance and robustness during recovery or the addition of new nodes.
• Small, Well-Understood Building Blocks
• Rather than building an all-encompassing service, small, modular components serve as building blocks for other services.
• Symmetry
• All nodes in the system are functionally identical, requiring minimal or no node-specific configuration.
• Simplicity
• The system is designed to be as simple as possible without compromising functionality.
• How Amazon S3 Works
• Amazon S3 is designed for scalability, high availability, and low latency at affordable costs. Here's a simplified
breakdown:
• Data Storage:
• Stores objects up to 5GB in size, each with up to 2KB of metadata.
• Objects are organized in buckets, owned by AWS accounts, and identified using unique, user-assigned keys.
• Access Methods:
• Buckets and objects are managed via REST or SOAP interfaces.
• Objects can also be accessed using HTTP GET or BitTorrent.
• Bucket and object names are formatted for easy HTTP access, e.g.:
• http://s3.amazonaws.com/examplebucket/examplekey
• http://examplebucket.s3.amazonaws.com/examplekey
• Access Control:
• Access is managed through Access Control Lists (ACLs) for each bucket and object.
• Authenticated URLs can be generated, allowing temporary access for a specified time.
• Additional Features:
• S3 can act as a BitTorrent seed, enabling efficient data distribution.
• Buckets can log HTTP access data to another bucket for analysis and data mining.
• Developer Benefits:
• Simplifies storage by eliminating concerns about safety, availability, and server maintenance.
• Allows developers to focus on innovation rather than managing infrastructure.
• Pricing:
• Pay-as-you-go model:
• $0.15 per GB/month for storage.
• $0.20 per GB for data transfer.
• Example: Storing 1TB costs approximately $1800/year, making it a question of total cost to serve, not just storage cost.
• Nirvanix: Cloud Storage Overview
• Overview:
Nirvanix provides cloud storage solutions powered by custom software and file system
technologies. Their services run on Intel-based storage servers across multiple U.S. locations, with
plans to expand globally.
• Key Features of Nirvanix Storage Delivery Network (SDN):
• Storage Delivery Network (SDN):
• A global cluster of storage nodes powered by the Nirvanix Internet Media File System (IMFS).
• Stores, delivers, and processes storage requests intelligently based on the best network location, ensuring
superior user experience.
• Supports multiple file copies across geographically distributed nodes for unmatched data availability.
• Infinite-Capacity NAS:
• Transforms a standard 1U server into a virtually unlimited network-attached storage (NAS) device.
• Seamlessly integrates with existing backup and archive processes.
• CloudNAS for Linux:
• Mounts the SDN as a virtual drive accessible via NFS, CIFS, or FTP.
• Allows storage administrators to manage files using standard permissions and access controls.
• Advanced Functionality for Administrators:
• Automated file replication based on policies.
• A single global namespace scaling to petabytes.
• Secure and encrypted data storage on Nirvanix’s global storage nodes.
• Enterprise Adoption:
• Initially focused on developers, cloud storage is now being recognized
by enterprises as a low-cost storage tier for specific applications like
backup and archiving.
• Nirvanix offers a highly available and cost-effective solution for
businesses requiring scalable, secure, and reliable storage.
• Benefits of CloudNAS
• Significant Cost Savings:
• Reduces costs by 80–90% compared to traditional storage solutions.
• Eliminates large capital expenditures while achieving 100% storage utilization.
• Secure and Reliable Storage:
• Offers encrypted offsite storage, easily integrating with existing archive and
backup processes.
• Provides built-in disaster recovery and automated replication across three
geographically dispersed nodes, ensuring a 100% SLA.
• Quick Data Access:
• Enables immediate data availability in seconds, unlike offline tape solutions
that take hours or days.
• Use Cases:
• Long-term archiving using established backup/archival systems.
• Centralized, shared repositories for departmental data.
• Disk-to-disk-to-cloud setups replacing tape for archival storage.
• Easy backup for all computers within a department.
• Availability and Pricing
• Free for 2TB Contracts:
• Companies with a 2TB or greater contract with Nirvanix can use CloudNAS for
free.
• Optional 24/7 support is available at $200/month per server.
• Risk-Free Trial:
• The "Experience Nirvanix Now" program offers unlimited use of Nirvanix
tools, including CloudNAS, with 50GB storage for 15 days at no cost.
• For more details, visit www.nirvanix.com/ExperienceNirvanix.aspx.
Standards
• Standards are the foundation of the World Wide Web and play a
crucial role in enabling cloud computing. They ensure:
• Connectivity: Standards make it possible to connect to the cloud
seamlessly.
• Content Development and Delivery: They facilitate the creation and
distribution of content across cloud platforms.
Cloud Application:

• A cloud application is software that runs on the cloud, removing the need
for installation on client devices. For these applications to work, there must
be a standard way for the client and cloud to communicate.
• Communication Basics
• Computers need a common language to communicate, similar to humans
needing a shared language. Without it, communication is impossible.
HTTP: The Communication Language
• What is HTTP?
• HTTP (Hypertext Transfer Protocol) is the standard protocol used for
transferring data between the cloud and your computer.
• It is stateless, meaning it doesn’t remember previous interactions. While
this reduces server load, it requires methods like cookies to track user
activity.
• HTTP in Action: A Simple Example
• When your browser requests a web page:
• Browser Request:
• GET/HTTP/1.0 Host: www.velte.com
• GET/HTTP/1.0: The browser asks for the home page using HTTP version 1.0.
• Host: www.velte.com: Specifies the website the browser wants to access.
• Server Response:
• HTTP/1.0 200 OK Content-Type: text/html <head> <title>Thank you
for visiting Velte Publishing.</title> {Web page content here} </body>
• HTTP/1.0 200 OK: Confirms the request was successful.
• Content-Type: text/html: Indicates the data format being sent (HTML in this
case).
• The rest contains the actual web page content.
• XMPP
• The Extensible Messaging and Presence Protocol (XMPP) is being talked about as
the next
• big thing for cloud computing. The problem is that current cloud services—
including SOAP
• and other HTTP-based protocols—are all one-way information exchanges. This
means that
• clouds do not operate in real time and might have difficulties clearing a firewall.
XMPP
• allows for two-way communication and eliminates polling.
• XMPP (also known as Jabber) is a protocol that Google, Apple, AOL, IBM, and
LiveJournal have all signed on with.
• The Problem with Polling
• What is Polling?
Polling is a method where a client repeatedly pings the host at regular intervals to
check for updates. For example:
• Email clients poll servers to check for new messages.
• APIs for many web services rely on polling.
• Challenges with Polling:
• High Load: Polling creates a heavy server load, especially during high-traffic
events. For instance, in 2008, Twitter faced spikes of up to 800 requests per
second, leading to service outages.
• Inefficiency: Polling wastes resources by constantly sending requests, even when
no updates are available.
• Efforts to Solve Polling Issues
• Salesforce.com:
Attempts to avoid polling by sending notifications directly to web services.
• Challenges: Developers face complexity, and firewalls need special configurations to allow
incoming messages.
• XMPP: A Potential Solution
• What is XMPP?
• Extensible Messaging and Presence Protocol (XMPP) was originally
designed for instant messaging but is also suitable for cloud services.
• Advantages of XMPP:
• Two-Way Communication:
• Eliminates the need for polling by enabling real-time updates.
• Extensibility:
• Built on XML, making it flexible and ideal for cloud computing.
• Efficiency and Scalability:
• Can handle millions of concurrent users on a single service.
• Challenges:
• Not HTTP: Since XMPP is not based on HTTP, adoption is slower due to the
reliance on existing web standards.