Evangelizing Mainframe
Print Email

tableBASE Provides a Real-Time, In-Memory Table Management Solution

We all know, mainframers use look-up tables in databases like DB2 or Virtual Storage Access Management (VSAM) to retrieve information and validate for business reasons. The data is these tables is usually static or does not change often. In cases like this, doing a Read of these tables impacts the performance of the application. And, in most cases, we read the complete table in program memory and do the processing. Wouldn’t it be better if we had few other options? 
We do, and one such solution is tableBASE, a real-time, memory table management solution on mainframes for faster responses. This DataKinetics product is very relevant to transaction-processing environments and can improve the performance of DB2, VSAM or IMS systems.
We are here talking about a set of tables, or table services (TS), that will be used by mainframe applications to store business data in tables external from program. These TS tables are stored in a library that is a basic direct access method (BDAM) file and can be shared among multiple applications running in batch and online (IMS and CICS) environments.
Users can define their own ad hoc libraries to maintain a private version of tables. This allows them to load specific data into a table without impacting any other users. Ad hoc libraries can be referenced from batch job control language (JCL) and testing tools, such as Xpediter. In most cases, batch jobs should access these TS tables from the library since this is the most efficient and simplest method. On the first read, the entire table is loaded from the library into the batch address space (above-the-line). From then on, all reads to the table will be from the version in memory, thereby providing high-speed access. 
A series of steps usually needs to be followed to set-up TS tables. The steps are:
  1. Define a TS table. A database administrator (DBA) assists the developer in designing the table and then finally "DEFINES" the table on the "test" system.
  2. Load table source data. The data in each table comes from a DB2 table or a queued sequential access method (QSAM) flat file. Users are responsible for designing the source data (DB2 table or file). Developers are also responsible for implementing a process to allow the source data to be updated
  3. Reload the data. Once the source data is updated, the data is loaded into the TS table via the reload process. During a reload, the entire contents of the table are replaced with the data from the source (DB2 table or File).
  4. Distribute. TS tables can be shared by multiple applications across multiple sites. The distribution process can be used to synchronize the contents of a TS table across any of the TS installed sites.
  5. Send. This process should be used to perform the initial release of a table from test to prod. It should also be used to re-release a TS table if the physical attributes (row length, key size, etc.) have changed.
Overall, tableBASE is a good table-manager and can lower costs and mainframe total cost of ownership. It can dramatically reduce batch-processing time and optimize mainframe MIPS usage, offering great performance benefits and more powerful and efficient applications.
I look forward to hearing your thoughts on this so that we can discuss this further. Happy mainframing!
Raghavendra Kulkarni is a mainframe technical specialist at the India-based IT-services company MindTree. He has more than nine years of experience in mainframe-related technologies and products. Currently engaged as a mainframe consultant, he is also a project manager and a pre-sales consultant. He can be reached at Raghavendra_Kulkarni2@mindtree.com.

Posted: 8/2/2011 7:48:14 AM by Raghavendra Kulkarni

Print Email

Join Now!