Page 2

Page 3

Pages 4 and 5

Page 6

Page 7

Page 8

DB2 UDB for IBM iSeries Computers - The Sage application port to iSeries consists of the application executing in a Linux partition accessing data stored in DB2 Universal Database for iSeries.

The support provided by commercial databases today is being driven by the rapidly changing needs of the modern application landscape. eBusiness applications in particular are shaping database characteristics including:

• Support for unpredictable but potentially very large transaction volumes
• Support for data interchange capabilities preferably using XML
• Data analysis in support of Business Intelligence and Customer Relationship Management functions
• Support of industry standards in order to minimise the impact of integrating a heterogeneous environment
• Support for the Linux environment

Additionally, because a wider range of scarce and expensive skills are in demand there is an increasing desire to make database systems as ‘self-managed’ and as available as possible.

IBM DB2 is a family of database products available for hand-held devices, workstations, departmental servers, and host systems. Because of its focus on maximising data access, application support, application extensibility, scalability, reliability, and minimising database management it is called DB2 Universal Database (UDB).

DB2 UDB has three different code bases:

• zSeries (z/OS and MVS)
• iSeries (OS/400)
• Unix (IBM and non-IBM), Windows and Linux

The reason for the different code bases is the architectural differences between the platforms and operating systems. However, major implementation differences between the operating systems do not translate into major differences in supported functions and levels of support. IBM ensures this through a variety of organisations, internal and external, tasked with managing DB2 family compatibility. This also had the advantage of enabling R & D output to be shared among the three development laboratories so that design techniques, algorithms, and program code can be leveraged. There will always be some differences among the code bases, some short term, others long term. Each of the IBM eServer platforms has its own unique characteristics which make it attractive to particular market or application areas. Consequently each platform tends to address different customer requirements with their own priorities. This leads to the different IBM operating systems having their own release schedules to meet those customer requirements and priorities, and of course IBM has no control over the release schedules of non-IBM operating systems. As a result we will have short term differences among the code bases. Long term differences, some of which may never disappear, are due to the different operating system architectures and the architectures of the underlying hardware platforms. However, these tend to be isolated to functions supporting the database configuration, and to some database utilities. For example, the iSeries automates the allocation and management of disk storage space as a result of its Single Level Storage architecture. Consequently, unlike the other members of the family, DB2 UDB for iSeries does not need or have the concept of table spaces. In fact it does not ‘see’ disk drives, only a huge address space which physically comprises main memory and disk storage. This is one of many benefits derived from the database being integrated within the OS/400 operating system.

A document has recently been made available to assist professionals who are developing applications for multiple DB2 platforms or are interested in porting applications among those platforms. This can be found at http://www-919.ibm.com/developer/db2/downloads/cpsqlref.pdf and downloaded. Its title is "IBM DB2 UDB Reference – SQL Reference for Cross-Platform Development".

DB2 UDB is gaining momentum. IBM research invented the relational database and SQL. Its rapid development and increasing customer acceptance is being funded by a continuing solid investment in R & D. The volume of data management patents continues to outstrip the combined efforts of all its competitors.

Today’s ever-increasing reliance on heterogeneous environments for application deployment makes conformance to industry standards an absolute requirement for any database. The new standard for SQL, defined by ANS and ISO, is the 1999 standard, formally referred to as SQL 3. DB2 UDB for iSeries has already delivered nearly 80% of the core standard requirements. It is therefore in a very strong position for modern application development and deployment relative to all other leading relational databases. Most of the other leading industry standards are also followed and a list is shown in the accompanying presentation.

The strategic initiatives for DB2 UDB for iSeries development are:

• Openness and Industry Standards support
• This is to allow ISV’s the maximum opportunities for application development on multiple database technologies and migration of existing applications
• Database technology leadership
• Aim for consistency across the DB2 family and maximum leverage from sharing R & D
• Continue to focus on iSeries fundamental strengths and core values
• Ensure continued industry leading platform availability
• Provide a scalable application deployment model that embraces customer environments from the smallest to the largest (from tens to thousands of end users)
• Ensure that the focus on ease of use is never compromised so that the total cost of ownership value proposition continues to lead the industry

IBM launched eServer around three value propositions:

• Innovative Technology
• Applications Flexibility
• New Tools for managing e-business.

Project eLiza represents the next stage of development under the "New Tools" banner. It is aimed at delivering servers that are self-managing. Self-managed servers will provide the following characteristics:

• Self-Configuring
• Self-Healing
• Self-Protecting
• Self-Optimizing

DB2 UDB for iSeries already provides many of the features embodied by eLiza. For example:

• Data is automatically spread across available disk drives
• The allocation of system resources can be automatically balanced dynamically to meet changing workloads
• New releases of the operating system automatically provide upward and backward compatibility
• When you purchase an iSeries system it includes the operating system and hence the database. Thus every iSeries arrives with the operating system, and hence the database installed and configured for use

• Following a system outage OS/400 can automatically maintain the integrity of database objects such as tables and indexes (access paths)
• The recreation of ‘damaged’ or unsynchronised indexes can be performed automatically by the system in a controlled way that minimises the impact on applications that are running
• Journals (logs) can be used to automatically recover applications and maintain transaction integrity. Management of the journals can be automated by delegation to the system

• iSeries is object-based and much of its implementation is below what is known as the machine interface. That is, all entities such as tables, programs, message queues, etc. are objects and all objects have the same structure which includes an object header. The object header tags the object as being of a particular type. This ensures, for example, that, unlike the PC environment, a data file cannot suddenly be changed into a program and executed. This provides fundamental protection from viruses.
• The object-based approach also provides the foundation for security. Authentication and resource security is implemented at the object level.

• The DB2 UDB for iSeries optimizer is fully integrated within OS/400 and is cost-based
• The database automatically captures and maintains statistics using indexes and a statistics manager. No manual statistics collection process is necessary.
• Database access controlled by the optimizer can be automatically adjusted and optimized based on any changes, manual or automatic, made to the execution environment. For example, changes to main memory, numbers of processors, numbers of disk drives, etc.

Focussing on the traditional strengths of the iSeries high availability has always been a key characteristic and an important ingredient in its low total cost of ownership. Loss of availability results in reduced end user efficiency and leads to higher costs. Independent analyst assessments of mainstream commercial systems have typically placed the iSeries second only to parallel sysplex deployment of the IBM mainframe. The iSeries hardware and the OS/400 software have a reputation with the customer community for high reliability even without the implementation of high availability functions. This is one of the key reasons for iSeries consistently topping the lists in customer satisfaction surveys. For even greater availability there are a number of options available including cluster proven independent auxiliary storage pools, disk mirroring, raid protection, etc. Database availability is optimised through not having to take it off-line for maintenance tasks such as statistics gathering, data balancing, index balancing and backup. For example, if new disks are added to the system OS/400 can be requested to automatically re-spread data using background tasks with minimum impact on system performance. In the rare event of a system failure, for example as a result of a power loss, recovery can be fast and automatic. For example, the database will not need restarting, table and index integrity is ensured on re-booting the system, and if database changes have been journalled (logged) transaction integrity can be maintained via commitment control, the system automatically synchronising the database with transaction boundaries.

Scalability is another key strength of the iSeries. Customer installations around the world vary from the smallest companies with only a handful of end users to large enterprises with several thousand users. DB2 UDB for iSeries underpins the scalability requirements by providing the necessary characteristics such as large table and index sizes, symmetric multiprocessing (SMP) and massively parallel processing (MPP), parallel data access methods, and rapid data save and restore.

Finally, another key ingredient of low total cost of ownership is usability. Focussing solely on DB2 UDB for iSeries it achieves many of its ease of use characteristics through the underlying architecture of the system. Its integration within the OS/400 operating system means that there are no external database interfaces for the customer to worry about. For example, interfaces with the software that handles security, transaction processing, communications, etc. are all part of OS/400. These are managed within the IBM development laboratory so there are no integration considerations even when new software releases become available. The result is that many tasks that other databases require a Database Administrator to perform are handled automatically. Most of those tasks that are not handled automatically by the system can be managed through the iSeries Navigator GUI. This GUI is used to manage most operational tasks on the system. This high degree of integration and automation also extends the ease of use philosophy to the application development environment.

Finally a word on Linux database support on the iSeries. Applications running in a Linux logical partition access data in DB2 UDB for iSeries in an OS/400 logical partition via ODBC. Early in 2003 an additional option should become available. That will be the ability to access data in DB2 UDB for Linux within the Linux logical partition.

View the IBM Presentation on DB2

Previous    Exit    Next