|
|
|
The VPE installation program installs intercepts at OPEN and CLOSE. During OPEN, VPE modifies the VSAM ACB in order to invoke the following techniques to improve VSAM performance:
During CLOSE, VPE performs statistics acquisition and reporting.
VPE utilizes a proprietary technique for optimizing VSAM I/O buffers. VPE creates large numbers of VSAM I/O buffers in virtual storage above the 16 Mb line (when available) in order to dramatically reduce I/O operations.
The performance of many applications suffers from the result of accessing VSAM work files, which cannot be tuned to avoid CI and CA splits. Deferred Write (DFW) allows splits to occur in virtual storage resulting in decreased CPU and DASD utilization when splits occur.
VPE invokes VSAM's Sequential Insert Strategy (SIS) automatically and when appropriate when records are added to a VSAM data set in order to improve performance. SIS reduces the number of CA and CI splits when adding multiple records at one point because it splits the CI or CA at the insertion point rather than in the middle of the CI or CA. Because half empty CIs and CAs no longer proliferate, FREESPACE can be reduced without increasing splits.
VPE invokes VSAM's Read Look-Aside feature automatically and when applicable. VSAM's Read Look-Aside feature reduces I/Os by looking first in the existing I/O buffers for the CI being read before reading from disk.
| Copyright © 2012 CA. All rights reserved. |
|