It comes from IT Jungle: “IBM v PSI: The Operation Was a Success, But the Patient Died“

This, I think, is a pretty good summation of IBM’s purchase of plug-compatible mainframe company Platform Solutions Inc. (PSI), which was previously suing Big Blue.

Last month I wrote a blog post questioning whether the mainframe specialty processors — and in particular the System z Integrated Information Processor (zIIP) and the z Application Assist Processor (zAAP) — can really save a mainframe shop money. It caused a little stir, from vendors and users alike, who contacted me and talked to each other in defense of the zIIP and the zAAP.

My goal of the post wasn’t to say that they can’t save money, because they can. My point was to explain that oftentimes, the savings comes in an indirect manner. In the case of the zIIPs and zAAPs, I said this:

So you might be able to save in software licensing costs, but the processors themselves cost about $100,000. In talking to users, it seemed to me that the benefit of the zIIP and zAAP was more indirect. By taking workloads to those processors, you can free up room on the central processors. That’s what might matter the most.

So instead of having to buy a new mainframe, you can just buy one or two of these zAAPs and/or zIIPs. The real savings comes not from the reduced software licensing costs as much as it does from being able to buy a six-figure specialty engine instead of a new seven- or eight-figure mainframe.

Some took umbrage with that, saying that savings can be realized immediately with software licensing costs, moreso than I was alluding to. Gregg Willhoit, the chief software architect for mainframe software at DataDirect, added that most customers of theirs don’t consider freeing up space on the central processors to be indirect.

“Most people consider deferred capacity upgrades and the ability to grow existing general purpose processors as close to real money as anything else,” he said. “A lot of people don’t mind considering deferred upgrades as immediate savings.”

DataDirect, which makes service-oriented architecture (SOA) software for the mainframe, has been working to get its applications eligible on the zIIP and zAAP. Willhoit said that for some products, such as Shadow z/Services, up to 85% of the work can be offloaded.

There are other software companies out there that are helping customers offload some work to the zIIP and zAAP, including CA, BMC, and Neon Enterprise Software. Hopefully the list will continue to grow.

Another issue at least one person took was how I said growth of the zAAP and zIIP hasn’t been the same as the Integrated Facility for Linux (IFL), which is another mainframe specialty processor. One person pointed out that IBM says year-over-year growth of specialty engines has been 85%. Impressive? Yes, if you look at the percentage. I would like to see the raw numbers to determine whether the large growth is due to a small base last year. I have yet to been able to get IBM to give me these raw numbers.

I remember a couple years ago when Sun was boasting that year-over-year growth of their x86/x64 servers was 81%. But that’s because they hadn’t previously been selling a lot of x86/x64 servers. If you sell two the first year and four the next, that’s a 100% increase, but it doesn’t necessarily mean there was a huge amount of growth.

In most things, using three systems is more difficult than using two systems. This is true for disaster restart programs as well, so it caught our attention that Hopkinton, Mass.-based EMC Corp. recently released a two-site configuration of its Geographically Dispersed Disaster Restart (GDDR) product. The company already had a three-site version on the market, apparently “cutting their teeth” on the more challenging configuration, and now making the product available to more  shops with smaller, two-site infrastructures.

Like it’s older sibling, the two-site version of GDDR provides a high level of availability in mainframe environments. The product “enables customers to automatically restart host mainframe systems, critical applications and EMC Symmetrix DMX storage systems to minimize the impact of unplanned or planned outages and enhance information availability and protection,” according to EMC’s press release.

This two-site version competes with IBM’s Geographically Dispersed Parallel Sysplex. While IBM sells GDPS as a service engagement, “EMC sells their offering as a product, with services strongly recommended, but not required,” noted Jim Baker, a research manager at IDC Corp. According to Baker, “This means that every change [in GDPS] is cause for a reopener with regard to the professional services price,” while EMC customers know what the costs are up-front with GDDR. And, depending on their sophistication levels (”Very, very sophisticated users,” Baker stressed), customers can implement GDDR without EMC’s assistance.

Another key difference is where the failover control comes from. IBM’s GDPS has K LPAR, meaning that its logical partition is inside the sysplex it controls. Instead, EMC’s GDDR “uses a heartbeat mechanism amongst the entities and can control failover from outside the sysplex,” Baker explained.

While companies that currently run GDPS would have to weigh the pros and cons of switching over to GDDR, Baker believes that the release of a product that competes with GDPS will give customers leverage over IBM.”Competition is a beautiful thing,” said Baker.

GDDR is intended for very large companies that have sysplex in multiple locations and use mainframe-only technology.

“This is a huge step forward for EMC in the mainframe business,” Baker said. “There is still a sharp EMC focus on the mainframe, where most of the world’s transactional data still resides.”

Whenever the California state Legislature needed more money in its coffers, the DMV would pay for it in maintenance costs on its mainframe.

Why? With many applications built on COBOL, changes to the system could lead to time and money spent on altering those legacy applications. One of those changes which came about regularly was whenever the state would hike licensing fees.

But as James Taylor writes, companies can learn to make meals from their mainframe leftovers. What in the world does that mean? It means identifying which parts of your mainframe architecture are static and don’t require a lot of changes — and which do.

In the case of the California DMV, licensing changes were frequent. Other processes, such as the application managing vehicle information, didn’t change much.

And so mainframe migration doesn’t have to mean moving every process off big iron because a few are changing all the time and causing maintenance costs to increase. It means migrating only those oft-changing processes to a distributed platform, where changes might be easier, even if it’s only because of who the company has working for them and what skills they possess. As Taylor puts it:

However, more careful analysis can lead to an interesting insight - much of a typical mainframe system is static, works fine and needs no maintenance. Often only a small portion of the system is responsible for much of the maintenance work.

IBM has been pushing mainframe specialty processors — the IFL, zAAP, and zIIP — as a way for mainframers to save money. But do they really?

Trevor Eddolls at the Mainframe Update blog writes about this very topic, calling it a “bit of a swings and roundabouts situation:”

An organization needs to pay for a specialty processor, and it needs to buy software that can make use of the specialty processor, and then it can start saving money – I hear the sound of Excel Pivot tables being used to present the result of what can be quite a complicated calculation. But is there any software that can make use of this hardware?

Eddolls goes on to mention DB2 as one prime workload able to be moved to the zIIP, and there are some other major software vendors out there — CA, BMC, DataDirect and Neon Enterprise Software — that are looking to offer customers the opportunity of running their software on zIIP or zAAP.

IFL, the Linux processor, has been a success for the most part, mainly because IBM has sold it — and customers have bought it — as a way to consolidate hundreds, if not thousands, of Linux instances onto a single box. That has potential savings implications in floor space, power, systems management, and staff.

Selling the zAAP for Java and zIIP for data applications hasn’t been as easy. Why? The consolidation attraction isn’t there for one, and the benefits are more subtle and indirect. The case for the zAAP and zIIP is that you can save software licensing costs by getting workloads off the central processors to the zIIP and zAAP, where workloads don’t count against the million service units (MSU) rating of the box that is used to calculate software licensing costs.

So you might be able to save in software licensing costs, but the processors themselves cost about $100,000. In talking to users, it seemed to me that the benefit of the zIIP and zAAP was more indirect. By taking workloads to those processors, you can free up room on the central processors. That’s what might matter the most.

So instead of having to buy a new mainframe, you can just buy one or two of these zAAPs and/or zIIPs. The real savings comes not from the reduced software licensing costs as much as it does from being able to buy a six-figure specialty engine instead of a new seven- or eight-figure mainframe.

James Governor, an analyst at RedMonk, has a great post on CICS over at the Mainframe Typepad blog. His basic thesis: CICS is becoming a cash-cow because IBM is invested in SOA-ing it to death instead of in expanding the features of CICS itself:

What happens if you want to change the underlying enterprise data model, for example? You can’t do that without changing the code. You can service-enable all you want, but SOA is as much about component and service isolation, enabling flexibility and portfolio maintainability, than service reuse.

Governor, who lists IBM as one of RedMonk’s clients, adds that down at Impact 2008, an IBM conference on SOA held in Las Vegas earlier this month, it sounded to him like IBM was too interested in, as he wrote it, “extending existing investments” instead of trying to find “net new customers for the box.”

“Leverage existing workloads? I am most interested in net new workloads on Z - and I don’t just mean Linux-based,” he wrote.

It’s an interesting concept, and one that I’m always asking IBM about. New mainframe customers, especially in the United States, can be hard to find. It seems that IBM is investing in existing customers in the U.S. and then grabbing new customers outside the U.S. Whenever I ask about new mainframe customers, IBM always falls back on the BRIC acronym — Brazil, Russia, India, China. Hoplon Infotainment is one of IBM’s most-often touted new customers because a) they’re a new customer; and b) they do online gaming on a mainframe, which is a novelty all its own.

But as much as IBM talks about BRIC when it comes to new mainframers, it seems like they’re throwing bricks in the U.S. There may be some American companies new to the mainframe, but overall for new customers, IBM seems focused overseas.

Novell has announced some discounts for users that want to run its SUSE Linux Enterprise Server (SLES) on a mainframe. Novell already has about 80% of the mainframe Linux market, with Red Hat Enterprise Linux (RHEL) having the other 20%, and so Novell is trying to expand upon its dominance in the area.

With the new Novell pricing, a three-year subscription to SLES on the mainframe will cost what a two-year subscription now costs. And a five-year subscription will cost what a three-year subscription costs now.

There is a catch — isn’t there always? The discounts only apply to users who are coming to Linux on the mainframe through consolidation of distributed servers, or for renewing SUSE Linux on mainframe customers. Although I would guess that probably covers a large chunk of the users who run or would run Linux on the mainframe.

A one-year subscription for patch and upgrade for Novell SUSE on the mainframe is about $12,000. RHEL starts at $15,000 per license per year on the mainframe.

The Novell announcement continues its push on the mainframe. Earlier this year, it came out with a starter kit that lets mainframe users try out SUSE Linux on that platform for free, giving still-skeptical users a chance to try out Linux on big iron.

IBM is expected to introduce its new mainframe next Tuesday. They’re holding an event in New York City, during which they’re expected the roll out the new version of big iron as well as preview z/OS v1.10, which is expected out in September.

I had thought that maybe IBM would make the announcement in Orlando, where the Share user group is holding its conference. But it looks like it will happen in NYC. I’ll be there and snapping pictures when they roll the new machine out. In the meantime, here are some details on the new z10.

For now, it’s just called ?????? ?. But you can still register for the mystery Web event, to take place at 2pm ET on Feb. 26. Hmm, that’s right in the midst of the Share user group conference in Orlando. Wonder what it could be, especially considering a new mainframe system is due out around that time. From the invite:

We have an extraordinary new product that we want to introduce to you. We can’t reveal much about this amazing new product just yet. But we can tell you it can remarkably lower your power and cooling costs. It can help reduce floor space needs, and is designed to deliver significant capacity improvements.

It’s truly an amazing product that your data center has never seen the likes of before. Please be sure to attend the web seminar so you can see for yourself why we are so excited about this announcement.

IBM also wants to invite you to a roadshow called “The Future Runs on ?????? ?,” which will bring the new ?????? ? to a city near you in the month of March. Have you figured out what ?????? ? stands for yet? I’ll give you a hint: The first word probably stands for “System.” You can figure out the rest from there. See you in Orlando.

In a conference call with investors, IBM Chief Financial Officer Mark Loughridge mentioned that Big Blue will be rolling out its “next-generation mainframe” this year, with the announcement and availability coming in late February.

Loughridge said the new version of big iron will have 50% more capacity than the current System z9 and enable “unprecedented levels of workload consolidation,” as well as being more energy efficient. We haven’t heard about any special events that IBM is holding, but Share, the mainframe user’s group, is holding a conference in Orlando in late February, so don’t be surprised if IBM makes the announcement there.

It has been about two years since IBM introduced its latest mainframes, the Enterprise Class and Business Class.