Browse Prior Art Database

Grouping (Batching) Database Update Requests

IP.com Disclosure Number: IPCOM000100349D
Original Publication Date: 1990-Apr-01
Included in the Prior Art Database: 2005-Mar-15
Document File: 1 page(s) / 34K

Publishing Venue

IBM

Related People

Clark, CM: AUTHOR [+4]

Abstract

The synchronizing instruction in DB2* is the COMMIT request, which is issued by the user to instruct DB2 to process a number of accumulated requests, to verify a correct result from all processing, or to roll back the database to the previous COMMIT point if any request has failed.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 100% of the total text.

Grouping (Batching) Database Update Requests

       The synchronizing instruction in DB2* is the COMMIT
request, which is issued by the user to instruct DB2 to process a
number of accumulated requests, to verify a correct result from all
processing, or to roll back the database to the previous COMMIT point
if any request has failed.

      During an Object Storage Hierarchy Manager (OSHM) process, a
synchronous update includes two requests:  a delete object request,
and a directory update request, followed immediately by a COMMIT.  An
asynchronous update includes only the directory update request and is
queued by DB2 until an eventual COMMIT request.

      When updates are processed synchronously, the data page must be
written to DASD as part of COMMIT processing for each update request.
 When requests are processed asynchronously, multiple updates may be
made to the same page before it is written to DASD.

      The COMMIT request is issued when any of the following
conditions occurs:
1.   A synchronous update is processed.
2.   The work queue is empty.
3.   A queue limit (currently 40 requests) has been reached. Forty
directory entries fit into a 4K data page.
4.   The process routine has requested it.

      When a portion of the database is being processed sequentially,
this results in a significant increase in throughput, due to the
reduction in I/O requests to the database.
*  Trademark of IBM Corp.