IBM announcements - an overview

IBM made quite some interesting announcements this week. I'll just start by giving you a short overview, but I'm surely coming back to some of them in my next posts.

• DB2 10 goes beta as announced in 'IBM DB2 10 for z/OS launches beta program (ZP10-0015)'. As you can see the naming convention stays the same. The beta program starts on March 12, 2010. No idea when it will become generally available. If you want to know more about its functionalities, then you should take a look at Willie Favero's blog. He has a great post on DB2 10.
• It's also the time of year for the preview announcement of the next release of z/OS, so here we go : 'Preview: z/OS V1.12 - Heralding a new generation of smart operating systems (ZP10-0110)'. And apparently a new release of z/OS also means a new and enhanced release of z/OSMF : "z/OSMF V1.12 is planned to be expanded with the addition of z/OS Workload Manager Policy Editor functionality, enhancements to the already valuable Incident Log and Configuration Assistant for the z/OS Communications Server functions, and the ability to add non-z/OSMF application launch points and links".
  
• Here's one I like particulary because it's new for the System z community. IBM took over Diligent in 2008 because of Diligents de-duplication technology. Up to now the Protectier deduplication Gateway was only available for distributed systems. It's a virtual tape solution that emulates LTO tape drives and writes the data to an attached disk array. But, in the meantime it also de-duplicates the data which can lead up to a 'compression' factor of 25 times. Now this technology has also become available for System z. There are two announcements, one for the hardware and one for the software part. I'm definitely going to elaborate on these announcements in one of my future posts.
• A bit out of my league perhaps but we also had the announcement of the Power7 engine and some new models. The press release gives a good overview of this new announcement : 'IBM Unveils New POWER7 Systems To Manage Increasingly Data-Intensive Services'. I just have to mention this one as this will normally become the engine for a future generation DS8000.
• And before I forget it, we already had the announcement of GDPS V3.7 last week : 'IBM GDPS V3.7: Enterprise-wide infrastructure continuous availability and disaster recovery (ZS10-0013)'.
  

IBM GDPS V3.6

It was a bit confusing to me : first there was a preview of GDPS V3.6 and only two weeks later there was already the announcement itself : 'IBM GDPS V3.6: Enterprise-wide infrastructure availability and disaster recovery (ZA09-0005)'. This is a very comprehensive matter, so I'll just recapitulate the enhancements you can also find on the IBM GDPS page.

"Improved coordinated disaster recovery across heterogeneous platforms by supporting:

• Distributed Cluster Management (DCM) support for GDPS/GM
  This is designed to allow the management and coordination of planned and unplanned outages across distributed servers for GDPS/GM together with Veritas Cluster Server managed clusters.
• GDPS/PPRC Multiplatform Resiliency for System z (xDR) support for LSS Sharing between z/VM LPARs
  This is designed to support mulitple z/VM LPARs sharing the same disk Logical Subsystem with GDPS/PPRC. It also allows multiple z/VM LPARs to share the same disk using the z/VM Cross System Extension (CSE) function.

Increased availability with:

• Reduced-impact initial copy and resynchronization of Metro Mirror volumes
  This reduces the exposure window where the environment is without Freeze or HyperSwap protection by reducing the need to throttle down the copy process. There is no need to schedule copies to off-shift times to run.
  
• Remote Pair FlashCopy support
  This is designed to provide the capability to allow a FlashCopy relationship where the FlashCopy target device is a Metro Mirror primary device. The value of FlashCopy can now be combined with the value of HyperSwap.
• GDPS/PPRC Timer Support
  This allows the GDPS controlling system (K-sys) to continue processing, even when the server it is running on loses its time source and becomes unsynchronized. The K-sys is will be able to complete any Freeze or HyperSwap processing it may have started instead of being in a disabled WTOR state.
• Two K-Systems support:
  GDPS/PPRC Multiplatform Resiliency for System z (xDR) allows GDPS to manage Linux guests on z/VM and native Linux on System z LPARs with the same quality of service as for z/OS LPARs. This includes support for HyperSwap for Linux data, and the ability of GDPS to manage Linux and z/VM LPARs, and IPL the Linux and z/VM operating systems. Two controlling system (K-sys) LPARs can now be defined, both with awareness of the non-z/OS environment. The two K-sys support removes the various exposures associated with running with a single K-sys and provides the same level of protection for xDR with two K-sys that z/OS enjoys with two K-sys.

Simplified system management with:

• New GDPS Health Checks
  Three new GDPS Health Check are being delivered with GDPS 3.6 can be used to inform customers when their configuration differs from IBM-provided best practice values.
• Query Services
  GDPS V3.6 will introduce a new interface to allow queries on GDPS monitored resources. This is designed to help simplify problem determination."

Not really connected to the announcement, but you might want to check out the FAQ if you have some practical questions concerning GDPS. And I'd also like to remind you I also discussed Basic HyperSwap in our previous System z Newsletter (only Dutch or French).

Redbook : GDPS Family

I still find this blogging business a bit strange. Last month I was a bit worried because I didn't have much to post. But then again, I didn't want to post just for the sake of it. But see, this week, I made as many posts as last month. Apparently my blogging just reflects the real world. Sometimes just nothing happens and then all of a sudden everything comes together. It's like having nothing to do for ages and then you get three invitations all for the same date. Synchronicity ? Now . . . back to business.

The Redbooks Newsletter is a great source to stay up to date with the redbook publications for your specific interest areas. You can subscribe over here. Though it's not a new one, the redbook that caught my attention this time is 'GDPS Family : An Introduction to Concepts and Capabilities'. This is a 180p. introduction to GDPS in its various forms. I guess the Table of Contents gives you a good idea of what's being discussed. Overall,

• Chapter 1. Introduction to Business Resilience and the role of GDPS
• Chapter 2. Service Availability technologies
  Explanation of the underlying technology such as Parallel Sysplex, copy technologies, synchronous vs asynchronous, Recovery Time and Recovery Point Objectives . . .
  
• Chapter 3. GDPS/PPRC
  Protecting data integrity, Active/Active and Active/Standby discussion, Freeze Policies, Open LUN support, Management with lots of examples
  
• Chapter 4. GDPS/PPRC HyperSwap Manager
• Chapter 5. GDPS/XRC
• Chapter 6. GDPS/Global Mirror
• Chapter 7. Combining Local/Metro HA with out-of-region DR
• Chapter 8. RCMF/PPRC
  This was unknown to me : it offers "a small subset of the function provided by GDPS/PPRC and GDPS/PPRC HM, providing just the ability to manage your remote copy configuration using a set of NetView panels. It is not a disaster recovery offering because it does not include the Freeze function and therefore cannot guarantee consistency across multiple LSSs". It may be a first step towards the implementation of "full GDPS/PPRC or GDPS/PPRC HM in the future".
  
• Chapter 9. RCMF/XRC
• Chapter 10. Sample continuous availability and disaster recovery scenarios
  Some case studies going covering single site solutions and various two data center solutions.
  

IBM GDPS V3.4: Enterprise-wide disaster recovery

Last week (March 13, 2007) IBM announced its Geographically Dispersed Parallel Sysplex V3.4 with GA on March 30, 2007. And as the title suggests, IBM wants to stretch GDPS in order to secure an enterprise-wide disaster recovery. This should also include 'open systems'. The announcement says : "IBM is also announcing Implementation Services for Geographically Dispersed Open Clusters (GDOC) to provide automation, testing, and management capabilities to support world-class high availability and near-transparent application recovery for open systems". This is called Geographically Dispersed Open Clusters (GDOC). As a matter of fact, this looks like a very promising evolution in the GDPS story. You can read more about GDOC in the announcement letter and in another one that was already published last year : "IBM Implementation Services for Geographically Dispersed Open Clusters".
Other highlghts of this new release :

• Management of Linux images and data, when Linux is running native on a System z server, not limited to Linux images running as a z/VM guest.
• Enhanced availability for a three-site GDPS Metro/Global Mirror configuration for various disaster and recovery environments.
• Enhanced GDPS/PPRC system management with a graphical user interface.
• New GDPS health checks for the z/OS HealthChecker to check for GDPS "best practices."
• More flexibility with additional automation for IPL messages.

You can find additional information on the IBM GDPS page and there's a very nice white paper called "GDPS: The e-business Availability Solution". This white paper has been reworked and contains already information on V3.4. But it's really a very readable introduction to GDPS and its functionalities.