Improve Performance with Parallel Queries

Suppose you have several gigabytes, maybe even terabytes of data. Without warning, senior management needs an immediate answer to a complex question. You want to find the answer as fast as possible. However, you don’t want everyone else on the system to be held hostage while you search and search. Their frustration over delays would only add to your anxiety.

Many Information Systems professionals face this scenario every day. They are backed into corners by conflicting user demands. They receive little in the way of understanding or help to improve the levels of service.

Raising Model 204 to a Higher Power

CCA understands your requirement. Model 204 consistently provides the fastest possible response time for both transaction and query processing.

Now imagine raising the performance of Model 204 by a factor of two, three, four or more. With Parallel Query Option/204 (PQO) you can.

PQO lets you break up those massive queries into smaller, more manageable pieces that are then all searched and sorted—in parallel. PQO finds answers faster because the requests access smaller sets of data. Requesters get their answers sooner and other users’ queries don’t wait as long in the queue. The result is increased management and user satisfaction with the level of service you provide.

Spreading the Power for Model 204

PQO provides transparent access to Model 204 data wherever it resides throughout your organization.

It allows you to share data between production, test, and even historical data regions. Now you can access Model 204 data residing on one region from any number of other Model 204 regions. Further, PQO allows you to distribute Model 204 regions within the same machine, on different machines, or even across different IBM operating systems.

Multiplying the Value

Testing of multi-process performance using PQO shows you can significantly reduce response time over single processor solutions.

When data is split evenly over two processors, query response time is just over half the single processor response time; with three processors it is just over one third. In testing, three separate computers were used so that the processors and the disk I/O subsystems were not shared. In each case the overhead involved in communicating the query to the parallel processors and then receiving the answer was minimal. Additionally, you can combine PQO and MP/204 to utilize parallel CPU processing and realize improvements in system throughput and response time, as well as query completion.

PQO provides obvious benefits when you spread requests across multiple CPUs, however, there is also much to be gained in a single processor environment. Previously, with multiple onlines running on a single processor, or even on tightly coupled processors sharing the same DASD, updates to a database file could only be made from one online. This caused some users to replicate their database overnight by copying it or perhaps re-running the day’s updates against the shadow copy. However, with PQO, customers can directly update a database file from any number of Model 204 onlines. This eliminates DASD space and overnight window batch processing time. It also gives the secondary users real-time access to current information, rather than data that is a day old.

Should You Consider PQO?

Is PQO the answer for you? It is if you:

	Want to avoid file level enqueuing when accessing data owned by another Model 204 online
	Need to reduce the impact of large volume, online reporting applications during peak demand times
	Must access geographically distributed data for query and update processing applications

Features

PQO offers the features you need to boost the performance of your information processing in almost any computing environment. With PQO you can:

	Share Model 204 databases among two or more Model 204 onlines
	Assign ownership for each database to a particular online
	Access an online from many other onlines
	Implement an online as both a client and a server
	Create requests with standard User Language
	Provide location transparent access
	Use APSY support for client and server applications
	Use APSY saved compilation support across nodes to reduce CPU costs
	Support scattered file groups
	Issue transactions that update files at one remote node per commit

System Requirements

	Current levels of OS/390, VM/ESA, VSE/ESA
	VTAM for OS/390 and VSE/ESA
	VTAM and IUCV for VM/ESA
	Horizon/204
	Model 204 Version 4 Release 1.1 or higher

To learn more about PQO, contact your Account Representative, or submit an Information Request Form.