This week, I want to continue with my impressions of the Guide Share Europe (#gseconf22 and #gseuk) conference, which took place at Whittlebury Manor on 1, 2, and 3 November. I’ll mainly focus on some more of the sessions that I attended. You can read my previous post here.
Wednesday started with breakfast, which was another opportunity to chat to other attendees. At one stage I was discussing banking with someone from a UK bank and the many ways I felt that banks could improve the service they offer to businesses. For example, because I often get paid in dollars and euros, I would like to have accounts linked to my business account that would simply accept and save those currencies. I could then choose when to convert the payments into sterling. As always, talking to real mainframe users was very illuminating. It also meant that I missed the start of the opening keynote.
I was on time for the “Db2 for z/OS Data Sharing: Configurations and Common Issues”, which was presented by Mark Rader, who is a Db2 for z/OS specialist in the IBM Z Washington Systems Center. Mark started off by explaining that Db2 data sharing and Parallel Sysplex constitute the infrastructure that provides enterprise applications using relational data the highest levels of availability, non-disruptive scalability, and dynamic workload balancing. He looked at the parallel sysplex components and the data sharing concepts. He then drilled down inside the coupling facility to look at lock structure, shared communications areas, and group buffer pools. He then looked at different ways of sharing before moving on to performance issues, availability, and workload balancing. It was a really interesting presentation.
For the next session, I was in the IMS stream watching Tracy Dean, who is an IBM Product Manager responsible for z/VM Tools and IMS Tools, and Ron Bisceglia, who is a Distinguished Engineer and Principal Architect for Rocket Software. Their presentation was called "Customer Experiences Managing the IMS Subsystem". They were both good presenters with a wealth of experience, and it was really interesting to hear about the problems that IMS-using sites have come across (I was going to say, 'got themselves into'!), and how those issues had been resolved using appropriate IMS tools.
After lunch, I was back in the security stream to a presentation about the “IBM Z Cyber Vault” given by Diego Bessone, Global IBM Z Sales Director. I know Dell and Hitachi have similar products, so I was very interested to find out more about IBM’s approach. Basically, what are called ‘safeguarded copies’ of your whole system can be taken every hour (or whatever interval you decide). These copies are stored on air-gapped hardware – an IBM DS8000. At intervals of your choosing, one of those backups is used to IPL a mainframe. This is used to validate the z/OS system, the subsystems and data structures, and the application data. If the backup fails the test, then users can restore from earlier versions. There was some criticism in the room of the approach. For example, the DS8000 isn’t air-gapped if it’s plugged into an infected mainframe. There seemed to be no thought given to onsite procedures like needing someone to be continually booting a mainframe from a backup, or the time it would take to get the right person at a site to give permission to stop all work on the mainframe and start a recovery. There also didn’t seem to be an easy way to find the ‘right’ immutable backup to restore from. I thought it was a great idea, but just needed a bit more work.
I followed up by attending the "What does IBM Z Cyber Vault Mean for an IMS environment?" This was presented by IBM's Tracey Dean. While discussing the Cyber Vault, this was much more practical, for example listing the tools needed to ensure that the IMS data structure was validated, and the tools that were needed for recovery. Tracey also looked at the IMS tools needed for forensic analysis – tools that run in the production environment and produce useful reports about what has happened. Tracey also looked at various recovery scenarios. My one suggestion would be that there is more to recovery than just data recovery. In the event of a breach, bad actors may well have installed back doors and time bombs before encrypting of corrupting data. So, even if all the data – including in-flight tasks – were restored, the hackers could still get in and do the same or worse the next day.
The final afternoon session I attended was "What if You had to Quickly Recover an IMS Database, Could You?", given by Rocket Software's Ron Bisceglia. His was a very practical session, asking questions like whether to recover index data sets or rebuild them. he looked at how to reduce recovery time, MSUs, and human error. He discussed what type of recovery to do, as well as developing a good backup policy. Importantly, he looked at ensuring recovery readiness. he said that recovery needs practice and recovery plans need testing.
I had a meeting, so missed the final keynote. I did go to the drinks session and to the gala dinner. Again, it provided an opportunity to talk to real users, and there were some students on our table, who seemed very enthusiastic about a career in mainframes.
All-in-all, a very interesting day.
The third part of this article will be published next week.
No comments:
Post a Comment