Using the MVS high performance option (HPO) with VTAM

The MVS™ high performance option (HPO) can be used for processing VTAM® requests. The purpose of HPO is to reduce the transaction pathlength through VTAM.

Effects

HPO bypasses some of the validating functions performed by MVS on I/O operations, and implements service request block (SRB) scheduling. This shortens the instruction pathlength and allows some concurrent processing on MVS images for the VTAM operations because of the SRB scheduling. This makes it useful in a multi processor environment, but not in a single processor environment.

Limitations

HPO requires CICS® to be authorized, and some risks with MVS integrity are involved because a user-written module could be made to replace one of the CICS system initialization routines and run in authorized mode. This risk can be reduced by RACF® protecting the CICS SDFHAUTH data set.

Use of HPO saves processor time, and does not increase real or virtual storage requirements or I/O contention. The only expense of HPO is the potential security exposure that arises because of a deficiency in validation.

Recommendations

The general recommendation is that all production systems with vetted applications can use HPO. It is totally application-transparent and introduces no function restrictions while providing a reduced pathlength through VTAM. In the case of VTAM, the reduced validation does not induce any integrity loss for the messages.

How implemented

The SVCs and use of HPO are specified in the system initialization table (SIT) and, if the default SVC numbers are acceptable, no tailoring of the system is required.

How monitored

There is no direct measurement of HPO. One way to tell if it is working is to take detailed measurements of processor usage with HPO turned on (SIT option) and with it turned off. Depending on the workload, you may not see much difference. Another way to check whether it is working is that you may see a small increase in the SRB scheduling time with HPO turned on.

RMF™ can give general information on processor usage. An SVC trace can show how HPO was used.

Note that you should be take care when using HPO in a system that is being used for early testing of a new application or CICS code (a new release or PUT). Much of the pathlength reduction is achieved by bypassing control block verification code in VTAM. Untested code might possibly corrupt the control blocks that CICS passes to VTAM, and unvalidated applications can lead to security exposure.

Related tasks
Networking and VTAM: improving performance
Setting the size of the terminal input/output area (TYPETERM IOAREALEN or TCT TIOAL)
Setting the size of receive-any input areas (RAMAX)
Setting the size of the receive-any pool (RAPOOL)
Adjusting the number of transmissions in SNA transaction flows (MSGINTEG, and ONEWTE)
Using SNA chaining to segment large messages (TYPETERM RECEIVESIZE, BUILDCHAIN, and SENDSIZE)
Limiting the number of concurrent logon/logoff requests (OPNDLIM)
Adjusting the terminal scan delay (ICVTSD)
Compressing output terminal data streams
Tuning automatic installation of terminals
[[ Contents Previous Page | Next Page Index ]]