Sign in
 
HomeLibraryWikiLearnGalleryDownloadsSupportForumsBlogs
Resources for IT Professionals >
> 
> 
> 
Integration HIS/CICS performance issue

Answered Integration HIS/CICS performance issue

  • Monday, April 02, 2012 5:06 PM
     
     
    Sign In to Vote
    0

    Hi to all,

    I have a performance problem in integration HIS/CICS. I collect some network traces using WireShark and I see there are a strange pattern...

    Strange for me... Probably because I don't know SNA protocol!

    When the sender host send to receiver host the "HPR Status Message" the receiver "wait" 0.1/0.3 seconds every time... Anyone know why? I'd like to avoid this delay :-P

    Thanks :-)

    Ciao

    Emiliano

     

All Replies

  • Wednesday, April 04, 2012 5:54 AM
     
     
    Sign In to Vote
    0

    Hi Emiliano,

    First of all what is HPR :

    The Enterprise Extender architecture carries the SNA high performance routing (HPR) traffic of any logical unit type over an IP infrastructure without requiring changes to that infrastructure. It treats the IP network as a particular type of SNA logical connection. In this manner, the SNA protocols act as transport protocols on top of IP, as does any other transport protocol such as Transmission Control Protocol.

    Some tips to increase performance using SNA :

    • connect directly to the AOR (Application Area) and avoid connecting to the TOR : this will reduce your time of MRO and InterRegion LU62 (change the VTAM entry)
    • use a compression model like RLE (needs to be defined in your Logmode)
    • avoid the use of secondary sessions (the data will not be compressed) and use only primary sessions to connect to the Enterprise Extender (on you logmode you should specify the following :

          => Contention winner limit : 0

          => Partner Contention winner limit : 0

          => Automatic Activation limit : 0

          => Now you should first initiate the connection from the AOR using CEMT I CON(XXXX) ACQ

    Best Regards,

    Steve Melan - BCEE My Blog : http://stevemelan.wordpress.com

     
  • Wednesday, April 04, 2012 6:38 AM
     
     
    Sign In to Vote
    0
    Thanks :-)
     
  • Wednesday, April 04, 2012 3:27 PM
     
     Answered
    Sign In to Vote
    0
    Emiliano

    > Strange for me... Probably because I don't know SNA protocol!

    Did you ask if it was strange for the SNA specialist in charge of your partner Enterprise Extender SNA node?

    > Probably because I don't know SNA protocol!

    Note that not knowing SNA *at all* and being supposed to be supporting the APPN node in the IP-DLC link Service in HIS connected to the APPN VTAM node using Enterprise Extender is infeasible! You need minimally to have had APPN - which will include HPR - education. I trust this applies to your local SNA specialist.

    Anyhow, you can pass the following questions and comments on to your SNA-trained colleague.

    Incidentally, are you experiencing performance problems with CICS which lead you directly to this WireShark trace or is it just a guess that this delay may be relevant?

    -

    > When the sender host send to receiver host the "HPR Status Message" the receiver "wait" 0.1/0.3 seconds every time ...

    Under what circumstances do you observe the slightly delayed response to what WireShark is describing as the "HPR Status Message"? Specifically, is "business traffic" flowing or is the Rapid Transport Protocol (RTP) connection otherwise idle.

    If the former, there may be a problem worth examining.

    If the latter, you should hope for longer delays in order not to take up valuable bandwidth with "are you still alive?" exchanges. Recall that the Internet is being shared by other users - as if it needed saying! - or, if the IP network in question is an intranet, I guess there are still other simultaneous users within your organisation.

    > I'd like to avoid this delay

    So you see that it may actually be more sensible to have a more extended delay!

    -

    > Strange for me... Probably because I don't know SNA protocol!

    There's no substitute for understanding what you are doing!

    I see you got a reference from Steve Melan.[1]

    If you want to get to the same level in SNA/APPN as your SNA colleague, a quick-off-the-ground starting point might be the following redbook:

    Inside APPN and HPR - The Essential Guide to New SNA

    http://www.redbooks.ibm.com/abstracts/sg243669.html

    That would be followed, specifically for HPR practical topics by the following redbook with its extensive description of the adaptive rate-based (ARB) congestion and flow control algorithm:

    Subarea to APPN Migration: HPR and DLUR Implementation

    http://www.redbooks.ibm.com/abstracts/sg245204.html

    And then for Enterprise Extender, the following redbook:

    Enterprise Extender Implementation Guide

    http://www.redbooks.ibm.com/abstracts/sg247359.html

    If any of these are too "heavy", you can research the prerequisites which are typically provided.

    -

    Checking on these references, I noted that there is an option HPREELIV=NO, an operand of the VBUILD TYPE=XCA PORT statement. This will prevent the use of the "are you still alive?" mechanism at the level of the RTP logical connection whenever, as may well be the case with your configuration, the IP-DLC/Enterprise Extender logical connection constitutes the whole of the RTP logical connection. This is because the IP-DLC/Enterprise Extender logical connection operates its own "are you still alive?" mechanism - not for nothing is SNA connection-*oriented*! - and so the "are you still alive?" mechanism operated by RTP is not needed.

    This obviously applies to the VTAM RTP end-point. I don't know of any such restraint available in HIS.

    -

    If you need further help with your trace, in your next post please - you or your SNA specialist - include just sufficient of it for the purposes of checking the context.

    -

    [1] IBM has come up with a very frustrating web page subsite in the so-called "Education Assistant". The best reference I can give to allow you actually to download a PDF file is as follows:

    http://publib.boulder.ibm.com/infocenter/ieduasst/stgv1r0/index.jsp?topic=/com.ibm.iea.commserv_v1/commserv/1.9z/ee/

    Select:
     z/OS Communications Server
     V1R9
     Enterprise Extender and SNA

    You can then download the presentation "Enterprise Extender and HPR enhancements" - although what use an *enhancements* presentation is to someone who knows nothing about SNA, let alone Enterprise Extender, totally defeats my struggles to comprehend!!!

    Furthermore, I would expect your local CICS specialist to be well aware of how to organise and optimise the local MRO environment.

    Note that RLE (Run-Length Encoding) is the most basic way to reduce the volume of *text* data which needs to be sent. It can be used when it is the volume of such data which needs to be sent that is the performance "bottleneck". If the "bottleneck" is the processor power supporting the VTAM the API of which is used by your CICS system, use of RLE will be counterproductive! According to my presentation notes on this topic, there is *no* limitation applied by VTAM to the polarity of the session partners so the limitation mentioned by Steve Melan must be some weird HIS limitation.

    -

    Chris Mason
     
  • Saturday, April 07, 2012 6:13 PM
     
     
    Sign In to Vote
    0
    Thanks Chris, I'll forward your post to the CICS specialist :-)