Minutes of the GSE UK IMS Working Group Meeting
IBM South Bank, 16^th March 2000

Chairman: Frank Fleming ( frank.fleming@barclays.co.uk) - 01565-613581

Secretary: Peter Armstrong ( peter_armstrong@bmc.com) - 01784-478835

Administration

• Introduction
  • Welcome from Frank, with thanks for good attendance and introduction of committee members. Andrew Jack is now fit again, but sends apologies for this meeting as he is on holiday! Please identify yourself when raising points / contacts etc. so we know who you are. (Note from Peter :- nobody did so if I got your name / company wrong, I'm sorry).

  • Frank handed out an agenda, a list of attendees, problems / contacts form and a feedback form. There were 38 attendees from 23 companies. Please fill in feedback forms as they help us a lot.

  • Frank gave an overview of the agenda, which was as distributed.

  • The registration form was passed round with a request to update details, especially email ids. Now that we are doing things by email, you can have as many people as you like on the distribution list, which saves you having to pass on the minutes / agenda etc.

• Minutes
  • The minutes of the last meeting were agreed and no matters were arising. Many thanks to Richard Penfold for taking / typing last time.

• Correspondence
  1. UK Storage Management Working Group notice of meeting - Wednesday 22^nd March in St John's Swallow Hotel, Solihull.

  2. Minutes from the March 2000 Large Systems Working Group Meeting

• General
  • Future meetings - 2 day joint meeting with CICS 25th-26th May at Four Seasons Swallow hotel next to Manchester airport. Aiming for balance of DB and DC. Likely cost will be £200 per attendee (including accommodation) with substantial discount for speakers - so volunteer now! See feedback form for question on who will be coming so that we can plan numbers.

  • Next one after that will be 7 September at IBM Warwick - DB only

  • Final meeting 7^th December, probably in IBM South Bank - DB and TM (and a quiz?)

IBM What's New - Alan Cooper, Pete Sadler, IBM

• Details of IMS education. Expecting a new DBRC class with practical sessions in the future. September may be last time that V6 education is run.

• Alan asked if there was interest in an ETO or APPC course. Let Alan or Pete know if you need anything, as the courses are there ready-to-go and up to Version 6. Do you want any DBCTL education? The migration course is effectively dead. Comments to Pete please.

• New course just been created, which includes ITOC etc. 2 days, heavily practical. Oriented particularly towards ITOC plus WebStudio, but also useful for general ITOC. Lab based, so difficult to do as a one-company course. Pete is looking at running it 3^rd or 4^th quarter this year. Frank - want OTMA as well. Pete - yes, would need to include OTMA, MQ etc. Dougie - end of support for ITOC announced on web. Alan - yes, because it becomes IMS Connect in V7.

• IMS Technical Conference is next week in Barcelona. Only 8 customers from whole of UK are booked on it. Why aren't people going? Try harder to go on these conferences - there are lots of excellent education sessions and an opportunity to talk direct to people from the labs. Correct dates are 20^th to 24^th March.

• Pete than took over and talked about IMS V7 Jump-Start Services Offering. Some customers want early access to V7 code, so IBM are offering this jump-start option. Expect to get hold of the code in ca. May/June. Includes 50 hours of IMS advocate support and 3 seats on a V7 class. Pete ran through details, prereqs etc. If you are interested, then please contact Pete or Ron Blake. Can you go from V5 to V7 - yes.

• Back to Alan - latest IMS Newsletter distributed. Two new redbooks on IMS tools, and one on VTAM Generic Resources. The IMS Primer is coming out again - an excellent place to start learning (or relearning) about IMS - SG24-5352-00. Does the V5 Performance Guide contain details on OSAM? Yes. Alan recommends using OSAM for DBs (and obviously VSAM for indexes). Pete quoted a recent example where the book helped a customer reduce CPU from 2 secs to 2/10 sec. Lots of redbooks on Web-enablement.

• Niel Kenyon (from Manchester) has kindly volunteered to take over the role of the IBM rep at our meeting - welcome. Alan then ran us through the IBM people present - Alan, Pete, Dougie, Karen M., Ron Blake, Niel Kenyon.

• Handouts were provided (Note from Peter: - if you want copies of the handouts, then please come to the meeting or talk to a colleague who did - we don't send these out electronically).

IMS Program Restart - Alan Cooper, IBM

• Alan then talked about a new product from Icing Software. Concept is not new (restarting batch and BMP from last checkpoint), but as he reminded us "IBM is in the business of providing IMS tools."

• Frank - there are products which functionally replace DFSRRC00 - does this handle that sort of thing? Alan - I think so.

• Works by writing checkpoints to two dynamically allocated checkpoint-tracking datasets (2 per jobstep) in flip-flop. These are then used at restart time to find the latest checkpoint. Being honest, Alan also pointed out one technical issue, which comes about if IMS abends whilst the last checkpoint is being written and hence the checkpoint is "in-doubt", and showed how to solve it.

• Handouts were provided.

IMS eBusiness and the World Outside - Pete Sadler, IBM

• Things keep changing - fast, so Pete gave us an update on how IMS can be accessed nowadays. He took us through a series of options including eNetwork Host On-Demand, IMS Connect (V7, ITOC for V5/V6), MQSeries Internet Gateway, IMS Object Connector, IMS Component Broker Access, VisualAge Java Servlet, IMS TOC Connector for Java etc etc.

• Alan - one of the major problems is that many of these need knowledge of C++ or Object Oriented Programming and often these people don't talk to the IMS people. Pete - true, and this is where I typically act as the glue in a customer, and this is why the redbooks tend to concentrate on the Web side rather than the IMS part, as they expect you to know the IMS part.

• Ian - is there a Web / Java GSE group? Frank will follow up.

• Alan - you need to get into this area, as it is the future of IMS. Dougie - problem is choosing which one to use?

• Should you design an application for the Web from ground up or just convert your existing one with a different front-end? What does the end-user want? Do the screens make sense? For a prettier, more functional front-end, you may need multiple IMS back-end transactions. Frank - example is Barclays online banking. First screen gives all accounts with balances, which needs several transactions and we couldn't afford to rewrite all of them.

• Handouts were provided.

Hints and Tips, Problems and Contacts

• New hints and tips? Dougie is expecting one from a guy on the listserver. Send any input to Dougie (at work) please. Dougie_Lawson@uk.ibm.com (not his home e-mail - bandwidth is limited at home - he can take stuff home on diskette.)

• Name:Neil Price - TNT

  Problem: Storage Management people moved OSAM data to new devices, with UCBs above the line, which unfortunately is not supported by IMS V5.

  Contacts/Comments:Lots of people need to be involved in a decision like this and they weren't. Will also apply to logs as IMS uses OSAM for logs.

• Name:Ian Tyson - IBM / Royal Sun Alliance

  Problem:What level of response are you getting on PMR?

  Contacts/Comments:Depends on whether defect or non-defect PMRs. Enterprise Assist has contractual obligations. Frank getting very good response currently, even on Sev 3. Howard - 6 months for a Sev 3. Howard - problems can arise when it falls between e.g. IBM and hardware vendor.

• Name:Sue Lakes - AIB

  Problem:Running 3 IMS systems on one MVS and having CSA problems.

  Contacts/Comments:Neil running 4 x DBCTL with 7M private - no Fast Path. Howard - running 4 with Fast Path, and fifth one tends to kill MVS. Modules into LPA a la Performance Book - didn't bring a lot except having to IPL when you want to make a change!

Mainview - Bryan Child, BMC Software

• Bryan reminded us of the history of IMS (and Boole and Babbage), his favourite football team and pop music. IMS has changed over the years, with more bits and (now / in the future) more systems to look after. This is why Mainview is architected to give a single system image. He then gave us an overview of the Mainview for IMS product, with examples of how it can monitor all the new possibilities like Shared Message Queues, n-way sharing, Sysplex etc. as well as everything you ever wanted or have heard about in old IMS systems.

• Mainview also allows you to monitor transactions, which are spread over various bits like CICS talking to DBCTL and DB2, and does cross-referencing between DBD, PSB, ACB etc.

• Now that BMC and Boole and Babbage have joined forces, Bryan told us that the future will see a combination of their product offerings. This means that the monitoring function will determine what is wrong (e.g. a database performing badly), recommend an action (e.g. change the buffering or reorganise the database) and then trigger the action immediately or in a scheduled window depending on your choice. In other words the software will detect when work needs to be done and schedule it for you or tap you on the shoulder to tell you what needs to be done - e.g. only reorganise what really needs it, when it needs it. This is already available for DB2 and will be appearing shortly for IMS.

• Ian - can you see OTMA in DC Monitor? Yes

• Handouts were provided.

Continuous Ops - Features and Requirements

• The world is changing - if you open up your back-end systems to the World Wide Web, then you immediately drive up the availability requirements. So what areas are still outstanding in the IMS world, which stop you running continuous operations?

• Online Change - still need to find a quiet time to do the library switch. When is the quiet time? Frank - fortunately most of our DC stuff is not MFS-driven. ACB changes - need input from application groups. Need to know which DBD uses which PSB which uses which ACB which uses.. There are cross-referencing products available (including Mainview above).

• Frank - use DELTA, but still tend to /STO TRAN beforehand - depends on what you are changing. Mike - problems with conversational transactions. Neil - problem also with BMPs. Howard - looking at possibility of simply starting up a new IMS system with new blocks. What you have to stop depends on the change. Change DB structure - have to stop transactions etc. but if you are just adding a new SENSEG or something, you can keep things running.

• Alan - IBM strategy is to co-ordinate Online Change across a sysplex, not to change the way Online Change works.

• Would be nice to have a deferred /DBR - run what you have in-flight and start up new stuff with new control blocks / new programs.

• Neil - with our web transactions, if it doesn't work, send back a message to say try again later. Try it again and it may well work.

• SAS have been running for 3 years - they use BMC's DELTA for the above.

• Dynamic Allocation - not a problem for HALDB or DEDB as they get this info from RECON (and you can upgrade RECONs from V6 to V7 whilst IMS is running).

• System maintenance - problems with PMO, Linklist, LLA (Freeze). Frank - use PMO Quickfetch and have everything dynamic. Dougie - there are operator commands you can use with LLA, which you could automate. Frank - have two sets of RESLIBs and use ALIAS to point at one or the other; quiesce, switch ALIAS and start up again. How good are PMO etc. Frank - if it is not working, then IMS is not working.

• DPROPNR - had problems changing release. ITSA - did it over the weekend.

• Testing changes - how do you get to an acceptable level of confidence. If you test on real data, then you have to copy beforehand and possibly restore. Users have to sign up that you will keep the systems at a "recent" level of maintenance and this will involve them in user acceptance. Problem is if upgrade costs money and is not perceived as bringing any benefits / useful features. Do you do preventive or corrective maintenance? Counter-argument is how many bugs are you introducing to fix the one problem you have?

• Dougie - SUF is now a reality and you can use it in different ways.

• Do you have to have staff in every Sunday? Howard - trying to get away from Sunday outages.

• There are different levels of continuous availability - you need to define and measure application availability.

• Pete - need to review how you do copies - move to an online / near-online option. Dougie - RECON PRILOG will get you eventually? Peter - no, that was fixed in V4; as long as you take regular Image Copies and /CHE SNAPQ, it will keep running.

AOB

• Need more user presentations. Ideas - shared message queues, changing from VSAM to OSAM and the differences between them.

• Pete - DBCTL education needs - please don't forget.

• Neil - Y2K problem with GENJCL.CA - PQ34596 (HIPER YEAR2000) PER VOL CHECKPOINT ID NOT ENCODED FOR Y2K ON PRILOG RECORD.

• Dougie - COBOL programming - see the following URL: http://www.ibm.com/software/ad/cobol/cobpf120.pdf

• Thanks for coming - see you all next time in Manchester

Feedback

It looks as though most people enjoyed some of the topics yesterday.

The scores on the board are :

1.	IBM Whats new	54444453243444523335445443552354324	=3.7 = 75.4%
2.	IMS Program restart	53444354134424524434444443545344435	=3.7 = 74.8%
3.	IMS eBusiness www	54533444145455524444545454551452435	=4.0 = 80%
4.	Mainview	44423252233333523433433423343332243	=3.12 = 62.3%
5.	Continuous ops	34454343244543524444454454433455334	=3.86 = 77.14%