January 2006

Why Data Serving on a Mainframe

Why Data Serving on a Mainframe
Page 2

Contents

Introduction ..................................................................................................................................... 3
A common view of data ...................................................................................................................... 4
Making the most of your data and protecting it ....................................................................................... 5
Addressing the objectives for Corporate Data Serving ............................................................................. 6
IBM System z and IBM DB2for z/OS – synergy and strength ....................................................................... 9
DB2 for z/OS – foundation for enterprise data serving ............................................................................. 10
Summary ....................................................................................................................................... 11
Why Data Serving on a Mainframe
Page 3

Introduction
One of the few constants in the world of Information Technology is change, and that the rate and pace
of change increases over time. As clients evolve to become on demand businesses, they need to closely
integrate their people, processes and information across the enterprise and across its ecosystem of
partners, to speed up core business processes, reduce costs and enable more rapid "sense and response"
reflexes to changing market and business conditions, opportunities and threats.

A key element of this approach is IBM’s vision of Information on Demand. This focuses on
enhancing business value and lowering risk through the capturing, creating, integrating, analyzing
and optimizing of all types and sources of information throughout their lifecycles. This approach is
designed to help businesses become more reactive based on the information they obtain. The benefits of
this approach are numerous and include the ability to make decisions based upon accurate information
and real time monitoring and control, the ability to better manage risk and compliance, collaborative
benefits across the ecosystem through sharing of information with partners, and reducing overall IT
costs through the simplification of the infrastructure.

One key factor that underpins information on demand is that businesses may have to re-evaluate
their approaches to data serving. The purpose of this paper is to review the impact that these changes
have on data serving and the requirements that this generates as you consider the way you deploy data
serving technologies. It then reviews how the IBM mainframe can help you address these requirements
and can play a significant role as the data serving hub of the enterprise.

Ever since computing started to be used commercially, data serving has been a key element in
business use of information technology. Ensuring that applications and their users have access to the
data they need is critical to supporting business processes. Historically, the mainframe has performed
the role of a central data server for many large enterprises, taking advantage of its typically high data
throughput, scalability and strong security and resiliency capabilities. This has made it an excellent
choice as a data server to help support mission critical core business applications. However over time,
as businesses and their infrastructures have evolved with multi-tier solutions and technology spread
across the company, so corporate data has become more fragmented across the infrastructure.
Why Data Serving on a Mainframe
Page 4

Today, as businesses continue to evolve, and the reach of their technology expands within the
business and out to customers, partners and suppliers, there is an increasing need to re-assess their
approach to corporate data serving, being driven by two of the key requirements. Firstly, there is the
need to present a common view of data to many users, using many different applications both within
and outside of the business to help control costs and improve responsiveness to customers and suppliers.
Secondly, and somewhat conversely, there is the need to protect the data from unauthorized access,
driven by regulations, customer service and to maintain competitive advantage.

A common view of data

Probably the best way to demonstrate the driving forces and the complexity behind this issue is to look
at a specific example. Let’s take something as simple as an individual’s bank account. It is not that many
years ago that all transactions related to the account were processed within the bank. If you withdrew
cash from the account you had to go to a branch to do so. If you bought something from a retailer, you
paid by check, and those checks were sent to the bank to be cleared. Access to the data was only required
by a small number of applications and a few users.

Today’s picture is totally different. It is much more complex and involves many more parties and
places new strains on your approach to data serving. If you need cash, you can still go to your branch, but
typically you would go to an ATM. You are no longer limited to normal working hours, or your bank’s
location; you can access funds 24/7 around the world. The ATM you are using need not be part of your
bank’s network; it may be part of a competitors network. If you want to work with your account, you have
access to the information by phone, or increasingly over the Internet, again at any time, from anywhere.
When you make purchases, you use debit or credit cards, with retailers gaining authorization for
payment electronically day or night. Finally, of course, the bank may use the data to customize offerings
to meet your needs.

This is, of course, just one of many examples that span almost all industries, and as businesses
become more ‘on demand’ we expect that this issue will become even more prevalent. We see many other
examples today. Consider the impact that the Internet has had for data access for online retailers, or the
requirement for access to patient information, whether you are the local doctor or in a Trauma Center
across the country, or the impact on access to reservation systems for online bookings. The list is large
and keeps on growing!
Why Data Serving on a Mainframe
Page 5

So what does this dramatic change in data accessibility require? The data needs to be accessed
in many ways, by multiple applications and available to a large number of authorized users and
applications around the clock and around the globe. It must be updated in real time, constantly
reflecting the most recent transactions. Managing this across multiple fragmented databases, or
keeping multiple copies of the data aligned is extremely complicated and can be very costly to manage.
Having fewer copies of the data that all the applications and users access could be a logical way to help
simplify the problem and increase efficiency. However, to achieve that, it would likely require a server
that is highly scalable, very reliable and security-rich, and designed to support very high I/O rates with
very high bandwidth for connectivity and for tens of thousands of users.

Making the most of your data and protecting it.

There is no doubt that the vast amount of corporate data that many companies store can be a significant
asset to the business when fully exploited. It may be used to help drive down costs through exploiting new
channels such as the Internet, allowing customers direct access to the data they need so as to help reduce the
staffing costs needed to provide the service to your customers. It can be used to develop better information,
to understand more about individual customers so that you can customize services to their needs, or to
identify and understand trends in the market providing the potential for competitive advantage.

It is clear that making better use of your data can help provide significant benefits. However,
increased dependency on data access for business processes along with opening up data access to a
much larger number of users brings with it a critical need for data availability and protection. In today’s
world, the Internet has meant that users expect to be able to access information anytime of the day, from
anywhere in the world. This means that your systems need to be available 24/7, 365 days of the year. If a
customer cannot access the information they need, not only may you have a customer service issue, but
there is a real concern that a competitor’s Web site may be only a couple of ‘clicks’ away.

It is of course impossible to generalize about the cost of ‘downtime’ of applications as there are
so many factors involved. Surveys have shown how cost of downtime varies by industry, from tens
of thousands of dollars to millions of dollars per hour, and it would probably vary from business to
business depending on size, business process design and many other variables. There are many other
factors to take into account. How critical is the application to the business? Is the application internal to
the business or open to customers and partners? How many users are affected? Are the transactions lost
completely or deferred to a later time? If the transaction is lost, is it a customer transaction that could be
gained by a competitor? What impact does that have on future potential transactions from that user?
Why Data Serving on a Mainframe
Page 6

Probably the only generalized comment that can be made regarding the cost of downtime is that
there are costs to the business, sometimes substantial costs. What is also probably the case is that as
businesses become ever more dependent on IT within their business and in interfacing with customers,
partners and suppliers, the cost of downtime is likely to rise.

Along with the need for high availability, there is a very critical requirement for security designed
to protect the data and maintain privacy. The requirements span many aspects of IT security. These
aspects include helping to secure data through encryption, whether that data is being transmitted across
a network or ‘at rest’ being stored on media such as tape; providing user identification and checking
that users only have access to data that you want them to be able to access; and helping to secure a key
corporate asset, the data, and the applications and the infrastructure that support the data.

There is no doubt that the public at large has a growing concern regarding protection of data.
This is driven by an understanding of the amount of data that is stored by businesses about themselves
and by the increasing need to provide sensitive information, such as credit card details, electronically
across networks, where the ramifications for any mis-use of that data is considerable. This issue has
been well understood by the legislature and there have been a number of regulatory requirements
put in place designed to address data protection with varying degrees of severity by industries and
countries. In some cases, the costs for non-compliance can be significant. However, nothing illustrates
the importance of data protection and security features more than the highly public, headline news
that has resulted from security breaches in recent years. These occurrences can have an impact on
business reputation and customer trust and may have the potential to affect the both the bottom line
and shareholder value.

Addressing the objectives for corporate data serving

We have seen so far that there are three critical objectives that have to be addressed for today’s corporate
data serving environment. Firstly, the objective to support fewer copies of data that is accessible by
multiple applications across the enterprise can require a highly scalable processing environment.
Secondly, the objective for access at anytime, from anywhere, can require a system designed for
extremely high availability. Thirdly, the objective for data protection and privacy features can require
strong security capabilities.
Why Data Serving on a Mainframe
Page 7

The latest mainframe system from IBM, the IBM System z9™, continues to evolve the technology
capabilities for which the mainframe is renowned. Its strengths in areas such as scalability, availability
and security are just some of the reasons why the IBM mainframe has for many years been a leader in
data and transaction processing, supporting many mission critical applications. Indeed, its presence in
the finance industry is a testament to these capabilities. Let’s now look at each of these areas and see
how the IBM System z9 performs.

Let’s look first at scalability. Mainframes have for decades been viewed as ‘large systems’ and the
latest System z9 is no exception. The combination of increased processing power as well as the ability to
have 54 processors in a single server produces a top of the range model, the S54, which is almost double
the capacity of the top of the previous IBM eServer™ zSeries ® 990 (z990) range. Then, if you require
even more capacity, in theory, up to 32 System z9 servers could be ‘coupled’ together utilizing Parallel
Sysplex ® technology, with a single application being run across multiple servers.

There is of course much more to scalability than just raw processing capacity; you need to be able
to keep the processors busy. IBM Mainframes have for many years been focused on what we refer to as
a balanced system design. As the name implies, this design approach is about balancing components
within the server, so that in addition to scalability in processing power, there is also scalability in
memory and in I/O capabilities so that the large quantities of data can be made available to support
a high volume of transactions. To enable a high volume of transactions, IBM mainframes utilize a
combination of virtualization and workload management capabilities designed to enable resources to
be shared across applications and reallocated dynamically between them based on transaction volume
and on preset prioritization according to your business needs. This combination of a balanced system
design, virtualization and workload management capabilities help IBM mainframes to be capable of
running at near 100% for long periods of time.

One other key element of scalability is the ability to apply additional capacity dynamically.
IBM mainframes provide a broad set of functions that we refer to as Capacity on Demand. One of the
capabilities that this provides, with the right pre-planning and configurations, is additional capacity
that can be added either permanently or temporarily without the need to stop the system, so that you
may be able to change the capacity of the system over time to match your business requirements without
having to interrupt the service you provide to users.
Why Data Serving on a Mainframe
Page 8

Let’s now turn to availability. In today’s IT market, almost every platform from every vendor
has very strong availability claims and it is easy to think that all platforms are pretty much the same.
However, it is important to realize that it is possible to measure availability in many different ways.
For example you could measure at a highly granular level such as processor availability, or you could
measure at the server level, or you could measure the entire system. Each of these approaches, and many
more, are totally reasonable, but the ramifications of the different approaches can vary substantially.

If we take the personal computer that almost all of us have on our desks, I expect the majority of
us have experienced some kind of failure, the system gets ‘locked’, an un-recoverable error, the ‘blue
screen’, and so on. But what caused the failure? Was it a hardware problem, or the operating system?
Was it a problem in the middleware or in the tools and utilities that help support the system? From your
perspective as a user, the origin of the failure is not the issue and wherever the failure takes place, the
fact that the rest of the components are 100% available does not help you. Whatever the cause, the
effect is the same: you cannot access your applications and proceed with the activities you planned. The
ramifications are probably at least frustration, and may lead to lost productivity and missed deadlines,
which may also have financial implications.

Of course in this environment, there is normally only a single user affected, but with a large
corporate data server there may be tens of thousands of users affected, both inside and outside of the
enterprise and the ramifications are probably much more severe. IBM mainframes’ design point for
availability is application availability, with availability capabilities and considerations taken across
all the system elements. There is, of course, a strong focus on hardware availability, with redundancy
built into a number of components from processors to power supply. But there are also availability
characteristics built into the operating systems, the middleware, the tools and utilities, and the storage
and networking technologies designed to enable you to access to your applications and data.

In the scalability discussion, we looked at the mainframe’s Parallel Sysplex capabilities that enable
a single application to span multiple servers. This capability is primarily about availability, designed so
that in the very rare event of an individual server failure, the application can continue to run across the
other servers within the sysplex. Not only is this approach designed to provide continued application
availability in the event of such a failure, the workload management capabilities built into this
environment are designed so that the reduced resource can be prioritized to your most critical workloads.
IBM System z™ technology is designed for up to 99.999% availability with Parallel Sysplex clustering.
Why Data Serving on a Mainframe
Page 9

Finally, let’s look at security. IBM mainframes have for many years provided a security-rich
environment and have been used by enterprises around the world to support their mission critical
applications. This includes the financial industry, where security requirements and applicable
regulations can present special challenges. Indeed, IBM mainframes have received some of the highest
levels of industry security certification. The mainframe’s strengths in security stem in part from its
history of supporting sensitive data for large enterprises, resulting in security features being built into
its design for many decades, as well as a systems wide approach with security capabilities built into
hardware, operating systems, key middleware and so on.

The mainframe’s security capabilities fall largely into three categories. First, there is user
identification and authorization, designed so that only eligible users can access their applications and
data. As the mainframe has for many years been designed to run multiple applications simultaneously
on the same server, these capabilities are well evolved. When deployed, Multilevel security (MLS) is
designed to be an efficient way to help reduce the number of copies of data required while increasing the
effectiveness of access management. Second, there are intrusion detection services, designed to help
detect and prevent unauthorized intrusion to applications or data. Third, there is encryption, so that
data that is being transmitted over the network or stored on tape for archival or distribution purposes
can be protected, should that data get into the wrong hands.

IBM System z and IBM DB2 ® for z/OS ® – synergy and strength
The foundation for these traditional strengths for data serving on the mainframe lies with the tight
integration that System z and z/OS share with the database – DB2 for z/OS. DB2 for z/OS is written
to exploit the System z platform and as a result can offer advanced features and function. IBM DB2 for
z/OS delivers rich function for highly scalable, industry-leading high availability IT infrastructure for
your enterprise data and on demand business applications. The combined power and capacity of IBM
System z with the high performance and availability of the z/OS operating system, and the strength
of the DB2 for z/OS data server can expand and extend your IT infrastructure and the business value
of your data. The combination of DB2 and System z provides a unique competitive advantage for on
demand environments by providing a flexible, cost effective and optimized foundation for Information
on Demand. This foundation allows you to better manage risk, supports your efforts to demonstrate
compliance with policies and standards, and helps to simplify management of your information
infrastructure. These capabilities are important to enable customers to use their core business data to
drive insight and gain competitive advantage.
Why Data Serving on a Mainframe
Page 10

DB2 for z/OS – foundation for enterprise data serving

DB2 for z/OS Version 8 is the most robust release to date with strong customer acceptance. Key
improvements include enhanced scalability, easier application development and porting, even stronger
security, and significant enhancements to help move closer to the goal of continuous availability.
Management of very large databases and 64-bit virtual storage support make systems management
simpler, while continuing to lead the industry in scalability and availability. This new version broke
through many old limitations in the definition of DB2 objects, including SQL improvements, on-
line schema evolution, longer names for tables and columns, longer SQL statements, enhanced Java
and Unicode support, enhanced utilities, support for more log data sets, support for more partitions,
and more. Extensive SQL enhancements provide for even tighter DB2 family compatibility, which
in turn helps ISV and application developers deliver and port applications quickly and efficiently.
DB2 continues to provide industry-leading availability by minimizing planned downtime. Database
changes, such as adding a partition and upgrading to a new version can now be done without an outage
– a breakthrough for availability. Multilevel security at a row level, improved flexibility for on demand
business and new encryption options provide a breakthrough for security. Version 8 and its enhanced
suite of tools is a significant milestone, addressing significant customer requirements and allowing for
ongoing incremental database and application growth.

DB2 has used the function of the System z and z/OS platform extensively for many years and plans
to continue to do so with future releases. The tight integration between hardware, software, and z/OS
drives unmatched capabilities that are core to delivering a flexible, cost effective IT infrastructure
that can deliver capabilities needed to help you to better manage risk, to support efforts to demonstrate
compliance with policies and standards, and to achieve lower cost thru an optimized IT infrastructure.
These capabilities are important to support key information-intensive applications that provide
business advantage. Our plans are for future releases of DB2 for z/OS to build on the foundation of
DB2 for z/OS Version 8, providing increased integration with the DB2 family and with the System z9
platform itself. New trusted security context and database features are expected to expand on the
concept of granular security first offered in DB2 for z/OS V8. Fast table replacement (a.k.a. cloning) is
targeted to help improve the availability of many Web-based applications. Improvements are planned to
SQL, optimizers and algorithms to help improve performance.
Why Data Serving on a Mainframe
Page 11

We have seen how DB2 and the mainframe have strong capabilities to help address today’s data
serving requirements. However, the mainframe’s evbolution in recent years has involved much more
than just these traditional capabilities. The introduction of a broad set of open and industry standards
such as Linux ® and Java™ along with the introduction of new middleware such as WebSphere ® , has
provided a much more flexible environment that can allow greater interoperability with other platforms
and may help improve access to the wealth of data on DB2 and to the mission critical applications that
are key elements of the mainframe environment.

Summary
In today’s on demand world, providing the support for internal and external clients requires an
increasing dependence on IT, with more users and more applications requiring access to a common set
of data in real time that is accessible anytime, from anywhere across the globe. So how do you meet these
needs? The answer will, of course, be dependent on your individual situation and requirements; there
is no ‘one size fits all’ answer for data serving. However, if you feel that meeting your individual situation
requires consolidating to fewer copies of data, running on highly scalable servers, while also providing
a high level of security and availability to provide the service your users demand, then the new IBM
System z9, which builds upon the inherent strengths of the IBM mainframe to deliver industry-leading
data and transaction serving capabilities, may well be the right choice for your data serving needs.
© Copyright IBM Corporation 2006

IBM Corporation
New Orchard Rd.
Armonk, NY 10504
U.S.A.

Produced in the United States of America

01-06
All Rights Reserved

IBM, IBM eServer, IBM logo, DB2, Parallel Sysplex,

System z, System z9, WebSphere, z/OS, and zSeries are
trademarks or registered trademarks of IBM in the United
States, other countries, or both.

The following are trademarks or registered trademarks of

other companies
Intel is a trademark of Intel Corporation in the United
States, other countries or both.
Java and all Java-based trademarks and logos are
trademarks of Sun Microsystems, Inc. in the United States,
other countries or both.
Linux is a registered trademark of Linus Torvaldsin the
United States and other countries.
UNIX is a registered trademark of The Open Group in the
United States and other countries.
Other company, product and service names may be
trademarks or service marks of others.

Information concerning non-IBM products was obtained

from the suppliers of their products or their published
announcements. Questions on the capabilities of the non-
IBM products should be addressed with the suppliers.
IBM hardware products are manufactured from new parts,
or new and serviceable used parts. Regardless, our
warranty terms apply.
IBM may not offer the products, services or features
discussed in this document in other countries, and the
information may be subject to change without notice.
Consult your local IBM business contact for information on
the product or services available in your area.
All statements regarding IBM’s future direction and intent
are subject to change or withdrawal without notice, and
represent goals and objectives only.
Performance is in Internal Throughput Rate (ITR) ratio
based on measurements and projections using standard
IBM benchmarks in a controlled environment. The
actual throughput that any user will experience will vary
depending upon considerations such as the amount
of multiprogramming in the user’s job stream, the I/O
configuration, the storage configuration, and the workload
processed. Therefore, no assurance can be given that
an individual user will achieve throughput improvements
equivalent to the performance ratios stated here.
1
Financial Impact of Downtime Per Hour (by various
Industries) Source: ©Eagle Rock Alliance, LTD. All Rights
Reserved 2003
2
Logical Partitioning Certifi cation – On March 14, 2003,
IBM eServer zSeries 900 (z900) was the fi rst server to be
awarded EAL5 security certifi cation. The IBM eServer
zSeries 800 (z800) achieved EAL5 on June 6, 2003 and as
of May 13, 2004 the IBM eServer zSeries 990 (z990) joined
this elite group.

ZSW01811-USEN-00