Thursday, March 31, 2011

The Novell Connection Magazine

Some time ago I had a post with a Picture of the Day.Well here's a similar one :

Where did I find it ? Well, it accompanies an article of the Novell Connection Magazine. The article is about server proliferation and all the problems it causes. The solution is of course very simple : "The New Linux Math: zEnterprise System = z196 + IFL + zBX + SUSE Linux Enterprise + zManager". Oh well, I'm not going to take part in the Suse vs. RedHat discussion. This is just another comprehensive article on the use of Linux on System z and how the story will even get more interesting once IBM will provide the x86 based servers for zBX later this year.

I also learn that apparently the IBM Smart Analytics Optimizer is a Suse Linux Enterprise-based appliance. Such an appliance is built by Suse Studio "with an integrated, preconfigured, and fully supported enterprise-class Linux operating system ready for deployment anywhere from desktop to cloud". This technology "to easily build scalable System z server images" will also be made available later in 2011.

Tuesday, March 29, 2011

zPrime : the saga continues

Once in a while the zPrime saga pops up again. Today I read a story about it on the DancingDinosaur blog of Alan Radding : 'Save through zPrime on System z'. It's an interesting post because for the first time (to my knowledge) it tells the story of a client who's actually using zPrime on its production site. It's about a European customer who has installed three zIIPs and is redirecting lots of workload to these zIIPs instead of running it on the traditional engines. A pity it doesn't tell us how many traditional engines they still use nor on what machine they are running (z10 EC or z10 BC). I'm assuming it's a z10 EC.

Alan Radding draws the parallel with the IBM Solution Editions which are the IBM alternative for making the mainframe cost competitive. But, to him, this is just not enough. IBM has to choose between losing mainframe clients (even with the Solution Editions) or cutting down its software costs as losing clients will even cost more to IBM. I must say it remains an interesting issue on how to keep people on the mainframe with or without zPrime.

One side note : I'm intrigued by the savings of "nearly $1 billion Euros each year" they are making with these three zIIPs. As they are redirecting workload from traditional engines to zIIP, I gather they save the MSUs of three engines. On a z10 EC that would be about 450 MSUs. As I reckon this is not a small shop, these are the upper, low(er) cost MSUs. This means they save about 80 million euro on software cost each month ? What kind of software do you have to run for this kind of bill. I don't know. I'd love to see some more information about this.

Still, an interesting story. As I always say, just check it out !

NOTE : As remarked in a comment, the article I refer to no longer seems to be there ?!

Monday, March 28, 2011

Traditional mainframe, meet the x86 "software mainframe"

Here's a nice 2-part article by Beth Pariseau "on the closing gap between distributed systems and mainframe technology". Starting point is a phrase from VMware Inc. CEO Paul Maritz saying "VMware was working to create a 'software mainframe' with its virtualization prowess". A bold statement two years ago, but the author states "the two worlds start to look more alike when you combine server virtualization for x86 platforms; advanced automation and orchestration tools for distributed systems; increasing power and virtualization integration in commodity chips; and the trend of scale-out computing".

On the one hand it looks like such a 'software mainframe' is just re-inventing the wheel : "For example, server virtualization vendors are launching portals that would allow end users of applications to request resources from the centralized computing pool and have them automatically provisioned as well as continually load balanced and monitored". Isn't this a possible definition for cloud ? But then again "In the mainframe world, this is simply how resource provisioning has always been done".

On the other hand this 'software mainframe' keeps growing because users have other "motivations - the foremost being cost and the flexibility that comes with scaling out smaller commodity systems rather than building up one centralized supercomputer".

The second part discusses the evolution of the mainframe in recent years, allthough not every one seems to realize this : "When they hear the word 'mainframe', today’s generation of IT pros may picture green-screen terminals and room-sized computers, but the traditional mainframe hasn’t stood still. With the announcement of its zEnterprise 196 last summer, IBM started to blend the mainframe and distributed systems worlds".

How will these two worlds evolve ? Will there be a "total convergence or life on separate clouds" ? Take a look at the article to compare the author's viewpoint with yours ?

Monday, March 14, 2011

RealDolmen System z e-zine N° 17

The 17th issue of our RealDolmen System z Newsletter has been sent out this afternoon. As we target the Belgian market, those who read Dutch or French can find this and previous newsletters on the RealDolmen Website : choose the Dutch versions or the French Versions. If you don't receive it yet, you can subscribe over here. If you have subscribed, but for some reason no longer receive it, please let me know.

The main topics of this issue are the WebSphere DataPower Integration Appliance XI50z for zEnterprise and the preview announcement of z/OS 1.13 and z/OSMF 1.13.

As usual there are also some hints, tips, links, an overview of recent announcements, EOS dates of software and operating systems and an agenda for mainly Belgian events.

Enjoy the reading !

Hot Topics Newsletter Issue 24

I'd rather've seen some of the articles of the new z/OS Hot Topics Newsletter half a year ago, because great emphasis is on the introduction of the IBM zEnterprise System. On the other hand it has the advantage it can already look back for more than half a year and give some insights from early experiences.

The first 25 pages are almost entirely dedicated to the zEnterprise with a general introduction, a glossary, some info on migration, an interactive guide to working with and setting up an ensemble (even for the less experienced HMC user), an introduction to the Smart Analytics Optimizer and how it's integrated with DB2 for z/OS. An especially interesting article (to me, at least) is 'Policy, Performance and Pizazz' describing how "zManager provides policy-based monitoring and dynamic processor adjustments to ensure that multi-tier applications that span System z hardware and BladeCenter hardware are managed with sufficient processor resources". zManager provides IRD-like management for the zBX as it can adjust processing allocations of virtual servers. Nice !

What else ? A lot of course, so I'll just give a brief selection of some articles. There's e.g. one on DB2 10 and security. There's a FAQ-like article on the VSAM CA Reclaim function that reduces DASD usage and data set reorganization. And as I've indicated already several times that z/OS 1.12 should give you some performance benefits, there's proof of this in the contribution of Steven Jones : 'Your mileage may vary, but on z/OS V1R12 it is definitely going to be better !'.

You can download the PDF version from the z/OS Hot Topics Newsletter page.

And, as I always say, just check it out !

Thursday, March 10, 2011

Preview z/OS 1.13 and z/OSMF 1.13

How interesting are previews ?

I sometimes wonder whether it's a blessing or a punishment when you have to do a presentation on a new z/OS release. It's one of those obligatory things that pop up from time to time at events and you know most of the clients are two or three releases back, they're still implementing version n-2 and don't give a damn about what's new in one of those future releases. Yeah, more functionalities they won't need and surely once again more CPU consumption. Well, we discussed the latter already a couple of times, so I'll skip that one for now. But how about blogging about it, isn't that pretty much the same pain in the neck ? Well I must admit I tend to do some copy/paste work because you at least have to mention it, no ?

Still, I think this time it's a bit different. I wanted to do a short issue on z/OS 1.13 and the according z/OSMF for our next newsletter. But I ended up with a three page article. The (very short) exec summary ? You really should start giving z/OSMF a chance, if you haven't done this so far. It's getting nicer every time. Oh, and as it's an exec summary I should of course add : and it will save you money. For those of you who get the newsletter, it's coming out next week and what I'm going to tell from here on also appears in our newsletter.

So, the preview announcement was already a couple of weeks back : 'Preview: z/OS Version 1 Release 13 and z/OS Management Facility Version 1 Release 13 are planned to offer new availability, batch programming, and usability functions (ZP11-0011)'. Availability will be september 2011. And before I continue : this is a preview, things may change until actual announcement.

z/OS 1.13

Let's first have a look at z/OS 1.13 itself or at least at a couple of the new things. One of them is a new base component, the z/OS Batch Runtime environment. Batch on System z (and particularly with Cobol) remains mission critical to lots of enterprises. On the other hand, a real batch window seems to be more and more a thing of the past. Enters the z/OS Batch Runtime environment. This component provides the framework for Java-to-COBOL interoperability, for transactional updates to DB2, and for sharing database connections between Java and COBOL. In other words, it will "enable COBOL and Java to interoperate for DB2 with transactional integrity so you can enhance and extend existing COBOL batch application programs using Java".

Next to that there are also some JES2 JCL enhancements designed to make programming JCL easier and give you more performance and more control of your batch applications. Two examples :
  • "Support is planned to allow jobs for which journaling is used to be stopped after a currently running step has finished and held for restart in the following step. This is intended to allow less-disruptive system shutdowns.
  • Support is planned for job return codes. This support will be designed to allow you to specify that the job return code be set to the highest return code encountered by any step, the last step, or a specified step in the job. This will help make it simpler to interpret the results of job execution"
There's a lot more of course, but something about ISPF caught my eye because as far as I remember this is something that's always been there in Xedit :
"ISPF is planned to provide support for : Line command level Edit macros, in addition to the existing Edit macro support. This new design is intended to allow you to write macros to be used as line commands, in addition to those you might have already written for use as Edit primary or initial processing commands."
OK, I must admit, when I say 'always', my professional memory does only go back for about a good 20 years.

z/OSMF 1.13 : a small recap

Click on image for larger version in new window

z/OSMF was first introduced together with z/OS 1.11 as a uniform management 'platform' for z/OS system engineers. Usually they have to go through lots of different interfaces for different tasks and this means losing time for experienced system engineers and a high learning curve for new people. z/OSMF has the advantage over former wizzards and stuff that it is really executing on the mainframe. So, people were a bit hesitant, reluctant in the beginning about using it or not and many decided not to use it. Well, I see some change coming and people are reconsidering their choices. And I think they should and z/OSMF 1.13 will only further encourage them to do so. Anyone willing to share their experiences with z/OSMF, I'll be glad to hear from them.

I'll just highlight the three most interesting functionalities that will be added in z/OSMF 1.13 : Software Deployment, Storage DASD Management and Capacity Provisioning. Some are just a start, but they indicate which direction z/OSMF intends to go and this only looks more promising.

z/OSMF 1.13 : Software Deployment

The "software deployment function is designed to provide the functions needed to create and deploy a copy, or clone, of an existing SMP/E-installed software image, including IBM software installed using ServerPac, CBPDO, or fee-based installation offerings, as well as ISV or customer software". Some possibilities :
  • Identify, modify, delete software instances
  • Generate jobs to copy a software instance
  • Verify cross-system and cross-product requisites, verify fixes
  • Copy ALL parts of the software
  • Copy the inventory (SMP/E CSI) along with the software
  • Help identify actions including catalog, configuration, and security updates
It's able to clone software
  • Locally, either on a single system or system-to-system within a sysplex
  • Remotely, system-to-system across a network, even multiple sysplexes
Worth a try ? I do think so !

z/OSMF 1.13 : Storage DASD Management

This looks like a first step for easier DASD Management : it's "a new, simplified process for adding capacity to SMS pool storage groups" via z/OSMF. For this the concept of reserve storage pools is introduced : it refers to containers of pre-defined available volumes. A reserve storage pool refers to a group of volumes which are available for future use. The reserve storage pool resource is designed to replace the need for a storage administrator to manually maintain a list of defined but unused volumes.

With the Storage Management task, users can:
  • Discover reserve storage pools which exist in the current system
  • List volumes in a reserve storage pool
  • Add volumes to a reserve storage pool
  • Maintain metadata for reserve storage pools
Worth a try ? I do think so !

z/OSMF : Capacity Provisioning

I won't go into much detail on On/Off Capacity on Demand (OOCoD), the Capacity Provisioning Manager (CPM) and the stand alone workstation based Capacity Provisioning Control Center (CPCC). There are several possibilities for the CPM that go from warning you about extra resources that are needed up to adding the resources automatically based on predefined schedules or workload requirements. To manage this you need the CPCC tool which runs on windows.

For the moment the "Capacity Provisioning Manager application" on z/OSMF "is designed to support easier monitoring of z/OS Capacity Provisioning Manager (CPM) status. This capability can simplify the work of a z/OS Capacity Provisioning administrator and provides functionality to monitor connections and to view reports for domains, configurations, and policies. Capacity Provisioning Control Center is planned to continue to be available as a separate Microsoft Windows-based stand-alone client". But I have a slight suspicion that we'll see all those components integrated in a future release.

Worth a try ? I do think so !

I can only conclude by saying : the people having to talk about this z/OS 1/13 and z/OSMF 1.13 preview will surely have some interesting material to talk about. So, for our Belgian customers, come and listen to Jan Tits next Wednesday on March 16, 2011 at the GSE z/OS Working Group meeting at the RealDolmen site in Huizingen.

Friday, March 4, 2011

Upcoming Belgian Events - interesting addition

I really want to add this event to the list I posted last week ànd I definitely want to stress the additional z/OS (read DB2) track of it in Brussels. The event takes place on April 6, 2011 in Luxemburg and on April 7, 2011 in Brussels. The content ?
"During this half-day event, we will provide you with a comprehensive approach to information integration and governance – driven by quality, compliance, risk, cost reduction and value creation imperatives.

You will learn more about:

You can find more information over here and the links above give you the entire agenda with keynote and breakout sessions. But there's more, take a look at this additional program that's taking place in Brussels in the morning.
  • 09:30 - Welcome & opening
  • 09:45 - DB2 10 Technical Overview
    by Roger Miller
    , DB2 for z/OS® Technical Evangelist, Designer and Architect, IBM Software
  • 10:30 - DB2 10 for z/OS Migration Planning
    by Roger Miller
  • 11:45 - How to get the DB2 9 and 10 for z/OS Performance benefits and Reduce Costs ?
    by Cristian Molaro
    - Independent Consultant - IBM Information Champion
What do you think ? Is it worth your time ? I think it is !