Worldwide server revenues are up 5.7% from last year, according to a recent report from research firm Gartner. The release said that IBM had “solid increases” in its System p (Unix) and System z (mainframe) lines.
As a result, IBM’s market share when it comes to revenue grew from 29.6 to 31.2, extending its lead over HP in that department to 4.6 points. HP still led in server shipments, however, with IBM trailing a distant third.
There’s a fairly easy explanation for this. IBM just came out with its newest mainframe, the System z10, at the end of February. It stands to reason that the orders would flow in during the second quarter, and that they would be big orders — mainframes cost millions of dollars. It will be interesting to see if IBM can maintain that lead in revenue market share in the coming quarters.
By some estimates, since this past February when the z10 was introduced, Linux has accounted for more than half the total mainframe MIPS IBM has sold since the z10 was introduced last February. I asked a few experts to react to that statement:
Charles King of Pund-IT Research Hayward, Calif. is not surprised. “Linux has been on the mainframe for almost a decade,” King said. “For customers that already have mainframes — leveraging z/vm rather than buying mountains of x86 servers on a year over year basis been getting a lot of play and that’s where a lot of the additional numbers have been coming from.”
“IBM doesn’t make a lot of money selling Linux operating environments, but they do sell a lot of hardware and support services. Linux has been a boon for IBM and for the mainframe. I wonder if the IBM mainframe would continue to have the presence today if Linux wasn’t around.”
David Boyes of Sine Nomine Associates (the company responsible for porting Linux to Big Iron ten years ago) agreed, in email response:
Consider that [IBM] pretty much hands out IFLs like acid drops, and if you plan to do any serious DB/2 or Java work with z/OS, you pretty much have to have a zIIP or a zAAP in order to not drive your CA and IBM software bills through the roof. Since those specialty processors always run at full rated speed (as opposed to an increasing percentage of sub-capacity standard engines due to the loss of the Flex-ES machines for the low-end market and the high cost of z/OS software), the proportion of new IFL and zIIP/zAAP mips being sold dramatically outstrips the percentage of standard CPs being sold.
My take is that the amount of traditional z/OS processing (which is the critical demand driver for traditional mainframe MIPS) continues to decrease, but is being offset by new workload on z/VM and Linux, and additional MIPS on z/OS to handle processor intensive workload like Java, which is exactly where the zIIP and zAAP come into play. Now, the fact that we NEED that many MIPS for Java apps says more about the quality of the software and the tooling than the system it’s running on — there’s a lot of genuinely rotten Java code out there that simply assumes that processor resource is infinite, and you’ll just buy a bigger box if it doesn’t run efficiently — but a $400K price tag vs a $125K price tag for the specialty engines is a compelling argument to absorb new workload.
We contacted IBM for confirmation, but no comment so far.
The cloud is hot. So hot. But not everyone is buying into its hype. One of those people is Bruce Arnold, a Miami-based web designer who recently blogged about “techno-marketeers” throwing out terms like “revolutionary” and “innovative” like they’re bead necklaces at Mardi Gras.
Arnold’s latest peeve is cloud computing and SaaS.
“Both terms are bogus,” he writes. “The only true cloud computing takes place in aircraft. What they’re actually referring to by ‘the cloud’ is a large-scale and often remotely and/or centrally managed hardware platform. We have had those since the dawn of automated IT. IBM calls them ‘mainframe.’”
Ouch! Arnold goes on to list the current hot hype terms and their ancient equivalents:
1. “Mainframe” is now “Cloud” (with concomitant ethereal substance).
2. “Terminal” is now “Web Browser” (with much cooler games, and infinitely more distractions).
3. “Service Bureau” is now “Saas” (but app upgrades are just as painful, and custom mods equally elusive).
4. Most IT buzzwords boil down to techno-hyped BS (just as they always have).
Please Mr. Arnold, tell us how you really feel.
BusinessWeek had a great story yesterday about American Express’ attempt to retain some of the institutional knowledge that was slowly draining from the company, mostly in the form of retirements. And yes, that institutional knowledge includes mainframe know-how.
The key to the change is a phased-retirement program, where the elder statesmen in the company are able to gradually drop more and more of their day-to-day responsibilities to educate others in the company about everything that they know.
In addition, they get more time out of the office doing whatever they want—be it planning for life in retirement or doing charity work. The phased retiree continues to receive a portion of his previous salary, benefits as usual, and the company in turn gets to hold on to some of its most valuable employees a year or more past traditional retirement age.
Right now the phased-retirement program is only available for employees in American Express’ technology and finance departments. According to Jim Rottman, the company’s head of workforce transformation group, it also allows those employees to “avoid some of the emotional and financial hurdles of sudden retirement.”
Another remedy for older mainframers, at least in the emotional sense? Go jogging.
SAN JOSE – One of the more popular sessions I attended this week at the Share user group conference was called “Things a Mainframer Needs to Know, But No One Ever Tells You.” The panel was a group of mainframe professionals who had been in the industry a while. A lot of the audience members, meanwhile, were newbies to the platform.
So here’s a quick rundown of their suggestions:
Documentation. From Mickey Scott, a senior IBM WebSphere software engineer (sarcastically): “So you’re the new guy. Ideally your environment is well documented by everyone there.” The reality is that may not be the case. So you need to start examining where everything is. One key suggestion from Scott was to look at the syslog.
“That’s all you need,” he said. “You should go through the whole log and see what happens when things are normal, so when you are asked to look at a problem, you don’t go chasing some message that is normally there.”
Start building your libraries. This doesn’t just mean manuals, although it does include that. It should also include MVS system codes, printouts of IBM Redbooks, JES2 commands, and JCL scripts. The list goes on and on. And the panelists recommended at least downloading them to a place that you can access even if you’re network connection is down.
Monique Conway, a technical lead for Tivoli software products, added that it’s a good idea to start saving all your JCL scripts. If (and she stressed if ) your company allows it, take those scripts home with you. Why?
“If you’re new to the mainframe, and this is your first job as mainframer and you’re spending all this time building up your JCL, then you go to another job, all that work is gone,” she said. “So when you build it up, take a copy and bring it home.” Again, she said, if your company allows it. If it doesn’t and you do it anyway, you may soon be looking for a new job.
Learn your acronyms. One of the audience members suggested this, saying that you should document all your acronyms and how to pronounce them. You’ll want to know that RACF is pronounced “Rack F” and not “R-A-C-F.” Scott chimed in that IBM tends to have a RAG in their organization — that is, a Random Acronym Generator. Yes, that sounds about right.
Learn how to work with IBM and other vendors. It’s probably best if the first thing you say to a vendor is something that I can’t print here. As one panelist put it: “A large part of what we do is deal with vendor problems, or problem vendors, depending on how you think of it.” Some key points to having a successful session with vendor support:
- Determine the true severity level of your problem, and not just the severity level that you (or your manager) thinks it is. You don’t want vendors to think you’re crying wolf all the time.
- Have your information at the ready: problem type, what happened, when it happened, and anything you’ve already tried to do before contacting them.
- Play nice, but at the time, be persistent if need be. The vendor might tell you that no one else has told them of the problem you’re having. A good answer, according to Steve Conway, the lead sysprog at Freddie Mac? “Somebody’s got to be the first to find the stuff.”
Learn Assembler. Although you don’t absolutely need to learn Assembler to be a mainframe system programmer, it can help when doing problem resolution because you’ll be able to read the system dumps and, as one panelist put it, can be the difference between being a systems administrator (which he likened to a janitor) and a great systems programmer.
Join Toastmasters. Eventually you’re going to have to give a presentation. Eventually you’re going to have to convince a manager to buy new hardware, or upgrade your software release. So it’s probably a good idea if you learned how to speak in public.
And finally: Get a big monitor. In this day and age of multitasking, the bigger the monitor, the better. In fact, get two monitors if you can get away with it.
In an effort to pare down the state budget, California Governor Arnold Schwarzenegger wants to start paying 200,000 state workers the minimum wage, which is $6.55 per hour. The only problem: To make that change, he’s going to need COBOL programmers, many of which he fired last week.
The payroll system is built on COBOL. And as a public-sector consulting firm CEO says in the Sacramento Bee story, COBOL programmers are hard to find. Some of those who came out of retirement to program COBOL part-time for the state were laid off. Others might ditch the state job and go looking for real money. From the Sacramento Bee story:
“COBOL programmers are hard to come by these days,” said Fred Forrer, the Sacramento-based CEO of MGT of America, a public-sector consulting firm. “It’s certainly not a language that is taught. Oftentimes, you have to rely on retired annuitants to come back and help maintain the system until you’re able to find a replacement.”
Retired state employees who have returned to work part-time for the state were among thousands of workers laid off last week.
Forrer said the system has tens of thousands of lines of code, so it is time-consuming to find and replace salaries for each job classification on an individual basis.
If Schwarzenegger wants the changes in payroll to be made, he’s going to have to start rehiring some of the COBOL programmers he fired. And I doubt they’ll be accepting minimum wage. Here are some selected comments from Slashdot:
There are plenty of COBOL Programmers out there, the problem is nobody in IT wants to hire old people.
And a response:
no – the problem is that no one wants to be paid minimum wage to program COBOL
And another response:
no the problem is social security pays more so why go back to 40 hours weeks of coding at that rate!
As I reported on SearchDataCenter.com, a new feature of the next version of z/VM will allow mixing and matching of ICFs, IFLs, zIIPs and zAAPs onto a single LPAR, presumably to help users prevent memory from being wasted and to help in test/dev environments. An important question, however, is where is the new version?
Here’s a quick timeline of the last few releases of z/VM:
- z/VM 4.3: April 30, 2002
- z/VM 4.4: May 13, 2003
- z/VM 5.1: Sept. 24, 2004
- z/VM 5.2: Dec. 16, 2005
- z/VM 5.3 (current version): Announced Feb. 6, 2007; available June 29, 2007
The pattern is that z/VM releases have come every year to year-and-a-half. So I would expect that the next version of z/VM (whether it be called 5.4, 6.1 or something else) will likely come this fall, or possibly even as early as the end of this month. I would expect it sometime in September.
I must admit. As a tech journalist, it’s always fun to watch two vendors go at each other’s throats. That is what has been happening recently between IBM and Sun employees, as they argue over the virtualization benefits of the new System z10.
It all started with a column that Jon Toigo wrote in the Mainframe Executive magazine called “Enterprise Manager: Virtualize This!” Now, let’s be clear about one thing: Toigo is not fond of x86 virtualization using VMware. That was made clear in the article, and is something he’s made clear before. Last fall, he said during a Data Center Decisions conference session (TechTarget runs Data Center Decisions) that VMware is “shoddy” and full of bugs. He recommended running a mainframe instead. Not everyone at the conference agreed with Toigo’s assessment, including other speakers.
Needless to say, Jeff Savit was not impressed. A principal engineer at Sun Microsystems, Savit wasn’t convinced by Toigo’s article. In fact, he said Toigo got a lot wrong:
If you stipulate that another platform can run only 1/4 the work that it can actually run, and omit the very substantial costs on the other platform – z, and believe grossly exaggerated claims about its capabilities, and fail to mention features of other platforms that provide comparable or superior features that z cannot do (VMotion anyone?), well, you’re going to be a few orders of magnitude off.
That set off people in the comments space, one of which was Toigo, who said he would get to the bottom of some of the claims he made in the article, such as being able to virtualize up to 1,500 x86 servers using the mainframe’s LPAR technology. The mainframe can host thousands of virtual servers, but within z/VM, not LPARs, as Mark Post wrote in the comment section of Toigo’s article. Post is an engineer at Novell who focuses on zLinux.
In addition to reading the comments to Savit’s post, you’ll also want to read a post by Tony Pearson, an IBM storage consultant, and the comments that went with that, as well as the most recent addition to the literature, a post by IBM senior IT architect Joe Temple, over at the Typepad Mainframe blog.
And don’t forget the popcorn.
I suspect that readers of this blog will tend to side with the IBMers, but Savit’s post is well thought out, and it’s not like he’s unfamiliar with mainframes, either. Before coming to Sun, he was a VP at Merrill Lynch overseeing their mainframe virtualization.
I think all parties provide a good example of how to properly argue over the Internet. None of them resorted to “your mother” jokes, or asking one another what they’re going to do when the 24″ pythons run wild on you. I kind of wish they did, though. But that’s just me being greedy.
A good indication of how well the quarter went for the mainframe was highlighted by Mark Loughridge, IBM’s chief financial officer, who said during a conference call for investors that “frankly, we were sold out.” He added that IBM also shipped a record number of specialty engines — the Integrated Facility for Linux (IFL), z Application Assist Processor (zAAP), and z Integrated Information Processor (zIIP) — in the past quarter.
All in all, a good quarter for the mainframe, which had the best performance of all of IBM’s server lines. It’s Power Systems brand was just behind it, however, with a 29% increase.
It’s pretty clear what people think about IBM’s purchase of plug-compatible mainframe company Platform Solutions Inc: They bought them to put them out of business.
As we and many others have documented, IBM and PSI were in a nasty legal fight over 1) whether PSI was violating IBM copyright by using z/OS on its Itanium-based servers; and 2) whether IBM was trying to monopolize the mainframe market by binding z/OS only to its own hardware. Now that fight is over, although T3 Technologies, a company that had a licensing and distribution deal with PSI, is continuing the legal battle.
The terms of the deal were undisclosed, although Trevor Eddolls at Mainframe Update threw out the Internet rumor number of $260 million out there. Eddolls also poses the question of what IBM will now do with a server line that runs Intel’s Itanium chip:
The second interesting point is what is IBM to do with the Itanium stuff? Up till now, IBM did not have any Itanium-based servers. It could mean that IBM has to rebadge servers from HP.
Eddoll believes, however, that the main reason IBM snatched up PSI was to kill “off these emulators so it can sell its own – as a way of bringing on-board potential new mainframe customers.”
Gartner, in a report on the IBM purchase of PSI, said the acquisition will end legal action between IBM and PSI, though it won’t resolve the lawsuit with T3. The analyst firm also said that buying PSI may have been cheaper than fighting a lawsuit against it, depending on how the legal proceedings were progressing.
So what happens to the PSI servers? The Gartner report, written by analysts John Phelps and Rakesh Kumar, said that users shouldn’t expect IBM to start selling Itanium-based mainframe servers, but that some of PSI technology may be incorporated into Big Blue:
As IBM looks toward the coming world of virtualization in a heterogeneous environment, it will likely find value in technology that PSI has under development, especially in the area of virtualization of I/O across heterogeneous platforms. Gartner believes that IBM does not intend to promote Itanium mainframe systems, but could use PSI’s engineering knowledge to develop a low-end Power-based mainframe offering.
Most telling, however, is Gartner’s recommendation for current PSI users: “Plan for the discontinuance of the Itanium mainframe system.”