Softpanorama
(slightly skeptical) Open Source Software Educational Society

May the source be with you, but remember the KISS principle ;-)

Google   

Nikolai Bezroukov. Portraits of Open Source Pioneers

For readers with high sensitivity to grammar errors access to this page is explicitly prohibited :-)

IBM Marries Linux to Outsourcing

I believe IBM is effectively taking over Linux, not through ownership but by influencing the influencers: manipulating the people and press involved in guiding its use, evolution and acceptance. Witness, for example, IBM's success in manipulating the press and lots of serious Linux players with regard first to the SCO lawsuit and, more recently, in the fawning attention paid its opening of 500 mostly expiring, and mostly irrelevant to open source, patents.

Paul Murphy Open Solaris and strategic
consequences Perspectives CNET/News.com

Introduction

Those who are interested in computer history knows that Microsoft's march toward monopoly was aided by anti-IBM backlash in the tech community. Microsoft was a clear alternative for IBM and for many people that was very important, if not decisive, consideration: the level of hate of IBM in late 70th to early 90th was  comparable or even stronger then anti-Microsoft feelings today.  For example the term "FUD" was first defined by famous hardware designer Gene Amdahl after he left IBM to found his own company, Amdahl Corp.: "FUD is the fear, uncertainty, and doubt that IBM sales people instill in the minds of potential customers who might be considering Amdahl products." All-in-all in 70th to early 90th IBM behaved like a ruthless predator that can go to almost anything to protect its System 360/370 monopoly and destroy competition.

Therefore it's rather funny to see IBM "on the other side of the barricades" and attacking a new monopolist  -- Microsoft.  For a while IBM was considered again a nice company and even enjoyed some positive feelings from mainstream IT.  In a dialectical way that is changing again as IBM became a flagman of  IT outsourcing.

It's no secret that IBM has become interested in Linux partially due to anti-Microsoft  feelings or hedging as well as due to its value in marketing of IBM hardware brands, especially mainframes. IBM love story with Linux started in 1998 with its investment in Red Hat. Of course the desire of revenge for OS/2 was somewhere in the background, but  in no way it was the main reason;  respectable business cannot afford such a luxury as revenge, especially if you sell the servers with Windows preinstalled for many billions of dollars ;-).

Linux marketing and riding Linux PR wave was probably more important asset that as IBM correctly calculated can positively influence the outcome of their fight with another Unix heavyweight, Sun Microsystems.  Also in the enterprise environment the word "Linux" implicitly means "outsourcing" and here IBM wanted to have a first mover advantage.

It would be interesting to know who was behind such a shrewd, bold and rather dangerous move. I suspect that the chief architect of Linux moves was Samuel J. Palmisano and if this is really the case I cannot but admire the level of executive talent he has. You really need to be extraordinary person to be able to make such a bold move in conservative IBM environment.   On high management level Linux was also considered to be dangerous for its price fixing stamp that can really depress earning from other software products for any company that adopts Linux or other major GPLed product. This "GPL infection" when customers start to demand to open other products or worse stop paying for them was an important consideration on high level.  But Perlesano (or somebody else) managed to understand that open source can serve ass a Trojan horse that gives its consulting business access to lucrative accounts around the world and can serve as the basis of outsourcing using offshore labour. And if  the consulting fees charged for the switch to open source are comparable to the license fees that would have been paid to Microsoft that gives IBM the possibility to beat Microsoft in its own game and simultaniously with good revenues enjoy the status of freedom fighter.

For example, an instant question about IBM adoption of Linux from major customers was: when they will open Websphere and Tivoli ( to name just two important software products in their product line.) The second and more dangerous question was: if they use free labor is not this only fair, but necessary to cut the prices.  I was once present at the meeting where IBM lost its bid exactly for this reason, although it has a reasonably solid offering: the company went to the competitor, that (at this time) was not associated with OSS.

So one important thing for IBM was to "compartmentalize" OSS: open source is good only as long as it fit into following three niches:

In short the strategy was to develop closed-source highly profitable commercial applications that can work well with already existing open source infrastructure and kernels, dialectically repeating the move that Microsoft so brilliantly played with DOS.

Such "selective Linux adoption" permitted IBM getting probably 90% of PR effect, while sustaining less then 10% of financial damage associated with "price fixing/zero price" stamp of any company associated with major OSS products. The latter had sunk half a dozen companies that tried to ride Linux bandwagon with Corel as the most notable example. 

At the same time both IBM brass and top specialists were far from becoming "GPL enthusiasts" and understood perfectly well that technically Linux is nothing new or interesting. Most of them (correctly) thought about Linux as a funny fringe OS, overhyped Unix clone without significant technical merits, as one of my IBM's friends aptly put it, more of  a young religion then a technical phenomenon. That's why in many IBM technical conferences (Tivoli is one example) talk about open source is a taboo. Also to use GCC after IBM compilers is not very appealing idea for any professional who have grown up on IBM products (BTW IBM was, and probably still is, one of the leading compiler vendors with such masterpieces under its belt as Fortran H, PL/1 optimizing compiler and PL/1 debugging complier).  Even in the programmer editors area good old Xedit  and its derivatives (eastern orthodox editors) can give Emacs a run for its money. Moreover even with important possibilities for savings due to outsourcing developers that Linux opens, the black flag of anarchy (GPL license) remains incompatible with the blue IBM's corporate logo and, especially, with IBM's practice in the area of intellectual property.

Ironically, IBM pioneered the strategy of  "intellectual property profiteering" (see IBM and Software Patents) which later was picked up by companies like Acadia Technologies.  Acacia is the only publicly traded company in the world that does nothing but acquire and license patents, and sue other companies over the patents it acquires and licenses; the company can even be slightly respected because among others they target the adult entertainment industry with several streaming media patents (implicitly trying to correct a very questionable US Supreme Court decision about pornography as a free speech) :

...the company has also redoubled its search for broad new patents. Basically, the plan is simple: find other patents held by inventors or smaller companies running out of money (or just patience); buy the patents; license them out.

"Companies in effect outsource their intellectual property departments to us," says chairman and CEO Paul R. Ryan. "An IBM or GE can do it themselves, but for small or even midsized companies, it isn't practical to assemble a team of attorneys and engineers to research, write, and file patents; find potential infringers; and ask them for licenses." Indeed, large companies have known for years how lucrative patents are. The most successful is IBM, whose licensing department made over a billion dollars last year. Though that's just a fraction of the company's $89 billion revenue, licensing is a low-expense, high-margin operation. (IBM will license almost any of its technologies and rarely has to go to court.)

Mired in a slump in the early '90s, IBM marshaled all its patents and sent out letters demanding licensing dollars from alleged infringers. It used high-pressure tactics to attempt to obtain licensing fees. It was actually IBM which created this "new business model" and small firms now have an increasingly difficult time breaking through patent "barb wire" amassed by large firms. IBM is the world's patent leader (it got 22,357 patents from 1993 to 2002) and earned roughly $10 billion in licensing fees from them.

"When you have so many competing property rights, the cost of clearing permissions is very large, and it becomes a greater and greater tax on what people can do," said Tim O'Reilly (founder of  O'Reilly & Associates Inc. which in the last century was the major OSS software books publishing house; later priorities somewhat changed). As Jonathan Krim wrote in Washington Post on Dec 11, 2003 in his paper Patenting Air or Protecting Property (TechNews.com):

But to critics, the system is used primarily by big companies to shake down the little guy, and they are even suspicious of some of those now calling for reform.

"Those of us who work with emerging technologies have been living with this," said Silicon Valley attorney Gary L. Reback, who often represents start-ups in battles against large, entrenched competitors. "But it has gotten so badly out of control that companies the size of Intel are beginning to be disadvantaged."

Reback often tells the story of how a team of IBM patent lawyers went to Sun Microsystems Inc. in the 1980s and claimed that the then start-up was infringing on seven of its patents.

After Sun engineers explained why they were not infringing, the IBM lawyers responded that with 10,000 patents, they would be sure to find some infringement somewhere, Reback says.

Instead, Reback said in a 2002 Forbes magazine article, IBM said Sun could "make this easy and pay us $20 million." After some negotiation on the amount, Sun cut a check.

Jerry Rosenthal, IBM's vice president of intellectual property, denied that the incident occurred the way Reback described. He said IBM supports several improvements in the patent process, including additional challenge procedures.

"But we need to get off this issue of too many patents," he said. "It's the quality of patent." That view is generally shared by the powerful community of patent lawyers. The American Intellectual Property Law Association also supports increasing patent application fees to provide more resources to the patent office.

Here is how Gary Reback recollects this incident in his 1992 Forbes paper Patently Absurd:

My own introduction to the realities of the patent system came in the 1980s, when my client, Sun Microsystems--then a small company--was accused by IBM of patent infringement. Threatening a massive lawsuit, IBM demanded a meeting to present its claims. Fourteen IBM lawyers and their assistants, all clad in the requisite dark blue suits, crowded into the largest conference room Sun had.

The chief blue suit orchestrated the presentation of the seven patents IBM claimed were infringed, the most prominent of which was IBM's notorious "fat lines" patent: To turn a thin line on a computer screen into a broad line, you go up and down an equal distance from the ends of the thin line and then connect the four points. You probably learned this technique for turning a line into a rectangle in seventh-grade geometry, and, doubtless, you believe it was devised by Euclid or some such 3,000-year-old thinker. Not according to the examiners of the USPTO, who awarded IBM a patent on the process.

After IBM's presentation, our turn came. As the Big Blue crew looked on (without a flicker of emotion), my colleagues--all of whom had both engineering and law degrees--took to the whiteboard with markers, methodically illustrating, dissecting, and demolishing IBM's claims. We used phrases like: "You must be kidding," and "You ought to be ashamed." But the IBM team showed no emotion, save outright indifference. Confidently, we proclaimed our conclusion: Only one of the seven IBM patents would be deemed valid by a court, and no rational court would find that Sun's technology infringed even that one.

An awkward silence ensued. The blue suits did not even confer among themselves. They just sat there, stonelike. Finally, the chief suit responded. "OK," he said, "maybe you don't infringe these seven patents. But we have 10,000 U.S. patents. Do you really want us to go back to Armonk [IBM headquarters in New York] and find seven patents you do infringe? Or do you want to make this easy and just pay us $20 million?"

After a modest bit of negotiation, Sun cut IBM a check, and the blue suits went to the next company on their hit list.

To reiterate the fact it is important to understand that Big Blue now reaps almost a billion a year in licensing revenue. The trend accelerated as other entrepreneurs - savvy or amoral depending on your perspective - slipped their patents into their holsters and went looking for a fight. And to what extent such a business mode is compatible with OSS/Linux marketing is a very interesting question. As we will see IBM managed to succeed in maintaining a very intricate balance of using Linux for PR marketing and "business as usual".  That success is a testimony of a very talented executives and PR people working at IBM.  IBM is a major player in IT not only due to its size and patent/class licensing portfolio but also due to the caliber of people employed. 

The major reason of commitment of IBM to Linux was that IBM brass saw Linux as the way to solve some important internal problems (that we will discuss below) not as a poor an anti-Microsoft play (or, more correctly, desire to keep Microsoft in check). The latter was definitely a side benefit, but not the main consideration. I would like to stress it again that in no way top executives can afford such a luxury as revenge to a business partner.  But IBM has a long memory and after losing PC and OS/2 battles a lot of IBM executives wanted to get even with Microsoft. 

IBM "100 Featherless Penguins
 in the Steel VM Cage" Road Show 

Although Linux put some pressure on AIX, IBM was able much better capitalize on it than Sun (or HP) and not only sustained much less damage, but even managed to get some ground from both Sun and HP.

The main difference was that Samuel Palmisano saw the possibility of using Linux as a powerful unifier of desperately different and sometimes even competing with each other  IBM's hardware product lines as well as for saving VM/CMS. The latter was really a brilliant idea. VM-based "software partitioning" technology that for many years were developed in VM/CMS was really unique technology that IBM never was able to market properly.  That might help to explains why IBM brass closed eyes on potential problems with GPL license and "crazy guy from Boston".  As one Slashdot reader aptly noted: 

Re:IBM and OSS (Score:1)
by Loge (http:/itcouncil.com/) on Wednesday June 27, @10:52AM (#124770)
(User #83167 Info | http://itcouncil.com/) It is certainly true that IBM is fully behind Linux, but that has little to do with any commitment to Open Source!

When is IBM going to release the code to its most strategic software products, such as WebSphere, DB2, or zOS? When hell freezes over, that's when!

According to Software Magazine, IBM makes more than $40 billion from its closed-source software right now. Do you think it is going to throw all of that licensing revenue away just to eke out a few more dollars from service or hardware opportunities? No way!

IBM's software packages are highly leveraged cash cows, and all of the "street cred" with the hacker community in the world wouldn't justify sacrificing them.

Everyone talks about the $1B that IBM is spending on Linux, but you don't hear as much about the $1B that it is spending on WebSphere, which it develops and releases the old-fashioned way -- with proprietary code and very high-margin licensing fees.

Make no mistake, IBM believes in the traditional software business just as much as Microsoft or Oracle. Its embrace of Linux as an operating system is simply a tactical maneuver to gain leverage in adjacent software, service, and hardware markets.

I would call this move a Napoleon-style move,  because everybody else including executives from Sun and Oracle tried to play Linux card mainly against Microsoft.  Here the Linux wave was played as an internal restructuring and marketing move as well as a sharp attacking tool against traditional competitors and first of all Sun.

The key idea was not new: "a good will created by supporting an open source product puts competitors on the defensive".  Noble motives is an important currency in any wars and why the war for the market domination should be any different. In a sense Samuel Palmisano used Linux as a Trojan horse both to attack competitors, especially Sun, as well as a driving force for the reorganization of  IBM's own corporate camp, a force to fight entrenched corporate bureaucracy within IBM itself. 

From the internal perspective this move created the possibility to unify three completely different IBM product lines (a huge disadvantage that helped Sun to became a number one Unix vendor and even eat some part of IBM's market share in mainframe market despite the fact that as a hardware vendor IBM was probably always ahead of Sun). It also created an extremely valuable opportunity to use Linux marketing hype to IBM's direct commercial advantage and introduce outsourcing under politically correct slogans.

I am not sure to what extent Lou Gestner participated in this transformation but playing Linux card has some Machiavellian component in it. Basically, this reinvention of IBM made employees expendable. Old IBM had been one of the best firms to work with, but to survive, it crushed its employees. Of course, it is easy to cut costs this way.  Cutting salaries and benefits is always easy and very profitable game if there is no resistance and the danger of mass strike.

But despite potential huge human costs the cornerstone of this strategy was a  brilliant idea (both technically and from marketing standpoint) to replace CMS in VM/CMS tandem with Linux and market this revitalized mainframe VM/Linux configuration as the best thing since sliced bread. Later this idea was logically extended into AIX which now also can serve as host for VM-style software-based partitioning and is capable of running Linux in one of the partitions. The ability to run Linux on top of AIX creates a lot of advantages and permits introducing those new capabilities in a very attractive to customers and "politically correct" way.

For mainframe users the move was of  tremendous importance due to instant creation of qualitatively better environment while preserving all major advantages of mainframes: suddenly they can run most Linux applications, overcoming mainframe isolationism. It also provided mainframes programmers with a free and more or less modern development environment. Here is the essence of  Sam Palmisano's strategy in his own words:

eWEEK: Spending $1 billion on Linux—it was great as a bumper sticker. But is Linux going to be a significant investment year after year?

Palmisano: The billion-dollar figure included marketing expenses and support centers and was spread over time. But, yes, Linux and open industry standards will be ongoing investments. People want to spin it as an operating systems debate or a Windows substitute. It has nothing to do with that. It enables applications to be developed against an open infrastructure. It gives people speed and flexibility, and they don't get themselves tied down with the infrastructure.

We also looked at the adoption cycle, including kids being trained in schools. We talked to partners developing on the platform; we looked at the work in the scientific community. And we looked at the Internet. By any measure, it was the fastest-growing server environment on the Internet. So we said this has got the potential to change the way applications are developed and deployed. That's where we came from and where we still are.

eWEEK: Were you the driving force?

Palmisano: It was the week before Christmas 1999. We sat around the table at my house: myself, Bill Zeitler, Irving Wladawsky-Berger, John M. Thompson and Nick Donofrio. We decided to launch a new business organization in early January. I told them I wanted to announce at LinuxWorld a year from then a bunch of enterprises using Linux. I didn't want universities and supercomputers. I told them I wanted German and Japanese banks—the most conservative places in the world.

It's still the fastest-growing [is] entry-level server platform by far. And 15 to 20 percent of our mainframe growth was Linux-based. But this thing is really about open industry standards. People want to polarize the debate. It's not about operating systems. To implement a true computing grid, whether it's a shared grid or your own internal grid, you need to work in this heterogeneous world. No one company is going to be able to dictate that the grid will be .Net [or] the grid will be Solaris.

It's not about what you like. It's about what solves the problem. The only way this problem will be solved is open industry standards.

It was December 2000, when outgoing IBM's chief Gerstner announced this astronomical billion dollars investment into the reorganization of IBM marketing, support and development efforts on a new open standards platform under the PR slogan "investment in Linux".

Skeptics instantly suspected that this is actually a billion dollars for IBM outsourcing efforts and restructuring toward outsourcing. Some might think that whose money still might be better spend on improving Lotus Notes interface (the most ugly interface I encountered in mainstream product: out of ten user only two dislike it, other eight hate it, hate, hate it ;-). Or on debugging Websphere and really horrible Commerce Suit (popular year 2000 joke stated that IBM's codenafts -- two men wearing spacesuits that were used in IBM's advertising campaign in 2000-2001 need to wear a spacesuit because there was no other way to survive IBM's buggy e-commerce environment ). 

But this time skeptics were wrong. That move that has little to do with direct outsourcing efforts and nothing to do with feeding GPL enthusiasts and parasites although IBM did hired several key Linux developers (for example, Ted Tso; see his interview   to Linux journal). A main target was reorganization of  IBM consulting organization as well as restructuring marketing and service organizations along the open standards line, so that several competing lines of IBM products from a liability became almost a competitive advantage: you can try the same application of three different platforms from the same company and chose the best.  Of course the largest part was spend on developing and positioning in the market mainframe version of Linux and promoting VM/Linux on mainframes as well as software partitioning on AIX.

In fact, IBM's push of Linux reminds me Sun's famous push for Java: Like Linux is a flawed OS, Java is flawed programming language without (almost) any interesting ideas. In the same wane Linux is just a reimplementation of an old operating system that became popular mainly due to associated software cult (or "young religion"), not due to technical merits.  While Sun promised  "making Windows irrelevant" with Java, IBM promised "make OS irrelevant" with Linux and VM technology, since businesses may not care which OS kernel they run under the VM as long as it is IBM's VM.

As for GPL, the last thing IBM cares is the suit from FSF about any real or imaginable GPL violations. It's probably clear even to Stallman that in this particular case FSF will seize to exist pretty soon and without much noise (or with undesirable for RMS media noise about his personal shortcomings).

I view this rebranding of the IBM server line, coupled with an endorsement of Linux and a renewed middleware emphasis as a honest attempt by IBM managers to bring the company incompatible server lines closer together and unleash some level of synergy that never existed before: 

IBM's efforts at driving Linux across all its platforms, including the mainframe, were reasonably successful in taking away some of the attention from Sun Solaris, especially with developers and independent software vendors. Previously, if you buy a server from Sun you know that you can scale it up to the highest end using just one OS -- Solaris. And that was a great competitive advantage of Sun against IBM that helped them to dominate the Unix server market since early 1990th.  It's mainly because of fragmentation and multiple, incompatible hardware offerings that IBM despite it impressive midrange hardware (PowerPC) found itself so much behind Sun in Unix server business and AIX as semi-forgotten (and under appreciated) flavor of commercial Unix. But with adoption of Linux as the open development platform running under VM on almost any IBM hardware lines as well as better Linux applications compatibility for AIX the game changed in IBM favor. With new faster Power5 CPUs available IBM got an opportunity to counterattack Sun in all sectors claiming that Linux unification is the best way to go (and mentioning  the advantage of "open standards" in the process).

Please note that Linux on mainframe make sense for a very limited spectrum of applications (essentially for applications that that are not very resource intensive)  as mainframe CPUs cannot compete with the latest line of Intel and RISK CPUs neither in price or performance. Major acceptance of Linux and thus the compatibility of Linux applications still are biased toward Intel's 32 bit platform.  Still VM/Linux created a threat for Sun higher level boxes. Sun very quickly noticed this danger and provided a more or less objective analysis of possible problems of potential pitfalls in execution of this idea:

Linux on the mainframe just doesn't compute. Here's why:

Linux on the mainframe is actually hosted by another proprietary operating system, z/VM. The optimized operating system for IBM mainframes is z/OS, formerly called MVS(2). Compared to z/OS, z/VM is a niche operating system with virtual machine (VM) support for new hardware features added late or often not at all(3). And Linux isn't designed to run in a virtual machine; implementation decisions that make sense on PC hardware don't fit well in a virtual machine(4). This is Linux. It's designed for Intel. It's not tuned for the mainframe hardware in which it's running.

Linux on the mainframe is complicated; this isn't Linux running on a two-way Intel server. Despite IBM's claims of easy management(5), customers still need a special machine room and specially trained staff for both z/VM and Linux. Finding mainframe staffing is an obstacle in many organizations(6); combining mainframe and Linux staffing further complicates the matter. Running multiple Linux images still requires administration that needs to grow with the number of images being run.

Linux on the mainframe can't respond to the workload demands of Web serving with high utilization--something IBM touted at the time of its z800 announcement. Horizontally scaled Linux farms are designed to handle unpredictable demand with above average peak loads. As demand rises, a load balancer distributes the traffic evenly across servers, which increases utilization. Because design capacity needs to handle peak demand, server farms often have a low utilization.

Given the relatively low cost of hardware, some organizations find this trade-off acceptable to ensure appropriate service levels. Contrary to what many believe, consolidating a Linux farm into multiple images on a mainframe would not change the demand pattern. Although z/VM can start and stop Linux images, it cannot dynamically add resources to match demand. As a result, a mainframe would need to size for peak demand just as the Linux farm would; high utilization is a myth.

It's neither fish nor fowl. Linux on the "mainframe" is not an open system, and there is little incremental RAS benefit. Although IBM claims "zSeries servers inherit the legendary IBM S/390 strengths in the areas of fault avoidance and tolerance, recovery from failures, and concurrent maintenance and repair for "always-on" availability"(7). We don't believe this to be true for zSeries servers running Linux. The "legendary" IBM S/390 strengths IBM references are the result of decades of development work on IBM's flagship mainframe operating system, known today as z/OS. The fault recovery features of z/OS are not found in Linux. z/VM does have some fault recovery features, but it is not nearly as resilient as z/OS. For example, z/VM cannot take advantage of Parallel Sysplex clustering, and VM hypervisor is an added single point of failure(8).

Applications that run on Linux for Intel need to be recompiled and recertified for each new platform; thus the application portfolio to run Linux on a mainframe is small(9). Consolidation without application availability just can't happen--and if the applications don't run on your platform, or if there are costly ports and changes to be made, cost savings can't be realized. Often the difference in Intel versus mainframe applications makes porting difficult(10). Additionally, different applications are ported to different distributions of Linux (for example, Red Hat, SuSE, and Turbolinux). Getting applications to run on the mainframe might require supporting multiple distributions of the 'same' Linux operating system.

The economics just don't work. IBM claims it is financially justifiable to consolidate as few as 20 Linux servers on a z800(11). With an estimated starting price of $400,000 for a z800(12) with a single CPU engine enabled, that claim seems exaggerated compared to Linux servers that hover in the $1,000 to $2,000 range. Sun's rack-optimized 1U form factor servers start at list prices under $1,000. When customers realize Linux on mainframe utilization will be low, and administration costs have still not been factored in, you can begin to see how the costs will add up. And let's not forget the support costs that will need to be purchased, either from the distributor or IBM Global Services. One example of a distributor's cost on an IBM Multiprise ran in the tens of thousands for the initial services and thousands a month for ongoing service(13).

Most points in this critique are right, but despite this in 1999 and 2000 VM/Linux became the flagman of the high-end enterprise level open source adoption. Actually this environment is (and for some time will be ) probably the only enterprise ready "flavor" of Linux. Still this solution is costly and attractive only to customers that already had mainframes. That means that commercially this solution is a questionable in view of existence Unix-based mainframes from Sun, HP and, yes, from IBM (Regatta p690 UNIX server).  The latter uses POWER4 chip (which IBM deems a "server on a chip") and is able to run 16 separate partitions on the same machine, and you can repartition while the server is in operation. IBM boasts the ability of the server to continue to operate even under the pressure of major system errors, something it claims no other machine can do. Like in Sun high-end offerings hot-swapping of components is also provided, which means the server keeps running while the malfunctioning part is replaced. This machine is half the price of the new SunFire 15K, boasting more processing power. That means that it competes in the same segment as VM/Linux.  Later IBM some up with PowerPC5 series that bring this VM idea even lower down the line of their servers.

The META Group has released a report that essentially says there is no future for Linux on the mainframe.  If META is right, a whole bunch of people should get ready to post their resumes on Monster.com, but I think that the rumors about mainframe demise are exaggerated . The META report says that mainframe configurations of Linux will fall out of favor as soon as 2005, by which time Unix and Intel-based Windows and Linux solutions will have enough mainframe-like features without the mainframe-like costs to make those options the better choice. It is true that  current "mainframe Linux" incarnations are relatively immature, as evidenced by the interminable list of errors/patches on Linux providers' Web sites. They have, therefore, been naturally limited to simple, less-critical applications such as LL Bean's e-mail notification system.

But aside of hardware, the main problem here, as was noted by Sun, is that Linux was not designed to run in VM environment. Still for VM environment penguin (a small bird that cannot fly) became especially appropriate logo. Actually IBM can remove several Linux kernel subsystems (starting with problematic SMP implementation) that are duplicated (on much higher technological level) by VM environment, saving Linux users from troubles with immature SMP as well as related troubles with virtual memory system and may be even scheduler. And please remember that before adoption of Linux, VM (that has a tremendous technological advantages but was almost destroyed by IBM's brass neglect) was just one semi forgotten dusty component among several IBM product lines. A bleak, non-sexy product used by a few universities (I know about Princeton),  and a few survived mainframe software developers slowly but surely moving toward retirement.

For existing mainframe customers this new duo of VM and Linux instantly combined amazing technological achievements of VM architecture with sexual appeal of Linux Penguin. But jokes about featherless penguin in VM steel cage aside,  IBM instantly got a unique possibility painlessly move all (or almost all) powerful tools and applications that exist on Linux (including Sun's Open Office) to mainframe environment: an extremely attractive move for existing IBM customers and greatly increases the value proposition of mainframe even if we believe that Sun's analysis is right (I do). It's really difficult to find a better move for preserving shrinking mainframe market share for IBM and I can understand the level of enthusiasm that prompted IBM brass to approve "Peace, Love & Linux" graffiti campaign in San Francisco and other major cities streets (see below).

Anyway, for IBM Linux reopened market for very interesting and formerly under appreciated commercial product (VM). Generally IBM's efforts of using Linux as a glue that cement all its previously isolated platforms, including the mainframe is a unique example of large corporate open source adoption that really makes sense. I would say that helped to sustain the market for mainframes as widely reported in the media purchase of VM/Linux by Norwegian telecom provider Telia (which probably was a mainframe user before) suggests. It's definitly easier to maintain, say, 100 Linux virtual machines then 100 1U servers with Linux. Here is an interesting opinions of mainframe programmers published in a LinuxToday forum.

anonymous - Subject: S/390 and Linux Big Deal ( Mar 21, 2001, 19:41:53 )
I don't like Microsoft anymore than the next person but Linux advocates spend as much if not more time spreading FUD, half-truths, distortions, and outright lies. Much is made about S/390 systems running several hundred or thousand copies of Linux simultaneously as though Linux were responsible. Sorry, but this is more, if not entirely, a testament to the S/390s virtual machine subsystem. Linux is just along for the ride. As to Crays, all I can is say BFD. Unix was doing that a *long* time ago so why should we be impressed that Linux does it?
 
IdoRexxAllDayLong - Subject: I am a VM/ESA 2.4 programmer. Linux gives to VM what VM has been missing. ( Mar 22, 2001, 04:04:36 )
I program using Rexx and DB2 Server for VM/ESA all day long. I also program Java Servlets on Linux but that is another story. The IBM VM/ESA platform is a good and robust platform. But many companies running mainframes consider VM/ESA to be a toy and useless compared to MVS/TSO (now called OS/390 and z/OS). Those of you around during the late 80's and early 90's will remember that VM/ESA never really got the respect from the MVS crowd. I remember one manager at an oil company telling me in 1991 that VM would be gone in 6 months. It is now an entire decade later and that oil company still has its VM system. In many companies VM/ESA was purely a PROFS system. Most companies have dumped PROFS for other various Email packages. So, when you say "Sorry, but this is more, if not entirely, a testament to the S/390s virtual machine subsystem. Linux is just along for the ride." you are not considering that Linux is as good for VM/ESA as VM/ESA is for Linux. Linux is breathing new life into VM/ESA. It gives a reason to not ditch VM/ESA. One also has to think......why did IBM not consider running 41,000 AIX images on VM/ESA? My personal opinion is the OS wars that occur at IBM. In my opinion Linux greatest strength is its open source and adaptability. No other operating system that I can think of has had an impact that Linux is having. Linux, Apache, and open source will prove to be the greatest software development of our time.

As a subtle, but logical sign of the success of this VM revitalization strategy in late March 2001, a six-stores billboard in New York's Times Square was emblazoned an image of the Linux penguin alongside peace and love symbols that give the multi-million dollar campaign a distinctly 60s feel (for those who do not remember, mainframes became operational in early sixties, I think the first one was produced in 1962).  This was start of famous IBM's "Peace, Love & Linux" advertising complain. As CNet commented:

The global marketing effort has a serious business aim for IBM. The core of the company's $1 billion commitment to Linux this year is an effort to make the system more attractive to corporate clients and sell more computers and software at the same time. The Linux message will appear alongside the company's eServer" brand, linking Linux to IBM computers running corporate networks.
... ... ...
Torvalds personally approved the ad campaign, Guarino said.

IBM told its business partners at a conference in Atlanta this week that "Peace, Love & Linux" is "a rallying cry" for open computing standards. Open Web standards and open software, like Linux, will make it easier for disparate computer systems and software to communicate across networks such as the Internet, IBM argues.

Open standards also undermine such proprietary computer operating systems as Microsoft's Windows and Sun Microsystems's Solaris. Both companies are IBM rivals, and IBM Chairman Louis Gerstner has said both companies' proprietary models are doomed. IBM is far along in a drive to make all its software and hardware Linux-compatible.

And this was money well spend. From IBM's Peace, Love, and Linux ad in Times Square:

In Q4 2000 IBM shipped 100% more OS/390 MIPS than the previous year. And AIX sales were up 49%. Both platforms can run Linux applications. Again, the Linux movement is all upside.

Not everybody was bought:

Re:This is how it would work... (Score:5, Insightful)
by 0x0d0a (568518) on Tuesday July 20, @12:32AM (#9745330)
(Last Journal: Thursday June 03, @01:50AM)
"Peace, Love, Linux" -- IBM

As regards IBM's "we're hip and open source" campaign:

I mean, yes, it's all marketing in the end, and a bunch of people sitting in cubicles trying to figure out how to manipulate me, an open source coder, and people like me. But it's kind of like being manipulated by an attractive woman -- yes, you're being manipulated, but it's *enjoyable*. You don't *mind* it. Having enormous, cold companies pretend to be warm and fuzzy and rub up against you while purring makes you feel *happy*. You just can't help it.

Also too much zeal in defacing public property enraged some people who misunderstood the historical importance of this moment. Here is a typical Slashdot posting from Martin Beauchamp:

IBM's Corporate Graffiti (Score:1)
by fuct (spush@spamalamadingdong.core.binghamton.edu) on Monday July 09, @10:20AM (#124766)
(User #76297 Info | http://www.core.binghamton.edu/~spush)
Scott,

I live in New York City, and I've noticed IBM's latest ad campaign for Linux (The penguin, the peace sign, and the IBM logo). So far so good. What disturbs me is the fact that it has been spray-painted, presumably by some paid agency of IBM, onto the sidewalks of New York.

My question is, why is IBM illegally defacing public property? Those streets are not for rent for any corporate agenda, they were paid for and maintained by the taxpayers of NYC and New York State. In fact, they are one of the few places to look if you don't want to be inundated with ads. Will your company be spray-painting the insides of eyeballs next? Forgive me if I sound a bit disgusted.

Martin Beauchamp

Being a good corporate citizen IBM acknowledged it had authorized an advertising campaign that involved the defacing of public sidewalks with graffiti and royally paid for damages to municipal officials (I think $100K for a city).  With such generous damages payments city authorities instantly realized that it does not make sense to abuse the IBM's generosity by hiring  "the graffiti extermination specialists" and this subtle understanding ensured that graffiti survived at least for a the period of time IBM expected, or may be even more  ;-):

When will you remove the graffiti? (Score:1)
by stefanlasiewski (flarg@[removeme]flarg.org) on Wednesday June 27, @07:30AM (#124762)
(User #63134 Info | http://www.flarg.org)
Several months ago, IBM sprayed 'Peace Love Linux' on a zillion sidewalks in San Francisco, with what you call 'biodegradable chalk' on a zillion sidewalks in San Francisco.

Today, much of that graffiti is still there, even after several large rainstorms. When are you guys going to clean up your graffiti?
 

IBM also engaged itself into "opening proprietary source" politics to help Linux. That includes for example Logical Volume Management System (LVMS) that Solaris 9 has and Linux didn't [LinuxMall 7/00]:

The next piece of code that IBM announced its plans to open source was its Logical Volume Management System (LVMS) technology, which allows different physical drives to be treated as a single volume by the OS, providing important flexibility for high-end systems that need to use a very large amount of disk space: "The LVMS release may represent a milestone accomplishment for the Community and for the extent of IBM's commitment to it. 'IBM is releasing one of its most advanced architectures for a Logical Volume Management System. This architecture is quite interesting as it completely integrates all disk and volume management into a single, highly extensible, easy to use entity.

For IBM with its army of research centers that is not that difficult, so AIX positioning has a lot of similarities to Solaris positioning: it's just a "big bird", the big brother of small penguins that cannot fly. It's interesting that on Itanium-based Intel servers AIX competes simultaneously with Linux and Windows XP. And that requires really talented marketing to explain customers that they will be better off of using AIX without alienating Microsoft and Linux crowd. This delicate situation was not unnoticed by some members of open source community. Here is an interesting comment from Linux Today - Yahoo-CNET IBM says Microsoft arrogant on open source

Xavalon - Subject: IBM and Open Source - very funny ( Jul 6, 2001, 08:55:20 )

Whoaw,

And you all believe IBM is trowing 1 billion $$ into Linux development. Haha.. Not.. Anybody seen the plans how they are about to spend the 1 billion. Lets see: a huge part goes into porting existing IBM software to Linux (e.g. DB2, Websphere, Tivoli etc). However, all closed source so the benefit will only be choice by customers. The financial benefit goes to IBM.

Another part goes to IBM Global Services in training engineers how to configure Linux on OS/390 and Linux on Intel Servers and how to support those ported product. Go and hire those just from school and having a basic Linux training IBM engineers for $200/hour or more. Very impressive.

So there you have the 1 billion that IBM spends on Linux.

But wait, IBM and Open Source is so funny. They pretend to support Open Source. But what have they give you so far:

euuuuhmmm an Open Source XML parser... yep that's about it...
Oh wait I forget the journaling file system..... whoaw I am very impressed!

Now they have 0.001% of their applications in Open Source.

Xavalon

It's clear that due to the brilliance of Samuel Palmisano's vision and the real necessity to find an unifying solution for three major hardware lines IBM played Linux card in a much more profitable way than Sun. IBM also benefited from its larger presence in Intel product space (so far Sun has only a Cobalt line of low end Webservers) and stronger service division. Unlike Sun an important part of IBM's business is making money from Intel hardware. And in the case of IBM's mid-range Intel servers, it's fairly high-margin hardware. IBM also understands that its strong support services and the economy of scale can work against Microsoft in large corporate environments. If some large company will not pay Microsoft, say,  $15-20 millions dollars to upgrade midrange and low departmental servers to W2K and decided just upgrade hardware preserving current OS (or even replacing it with IBM-based OS) they can go out and hire 10 IBM consultants  ($150K/year) to improve their mission critical server-based applications. Casting the bread upon the waters, as they say, the company can expect that in 4 years their proprietary server-based applications are better, while in case of W2k servers your money disappeared almost without trace in Microsoft coffers and the company needs to upgrade all your midrange and low servers again. Of course, not all IBM's consultants are good and maybe savings will be just $5 million dollars (or even less if the consultants will be able to persuade poor ignorant managers that a Quad-Intel 1GHz w/1G of RAM is not a good enough  departmental server and that one need R6000 to serve a couple hundred of pretty useless and mostly static pages on the internal corporate WEB site, but maybe those consultants will manage to find advantage from similar contributions made by others and savings will be more than that. Anyway, that possibility can be readily discussed with high level management.

But IBM strategy has once interesting side effect. With Linux encompassing all major lines of products, IBM essentially positioned itself as the center of open source universe. And as such it is often fighting against its own expensive proprietary products including Websphere, DB2 and Tivoli.  That contributes to schizophrenic personality of IBM. In other works, one problem was resolved another was created.

As we all know Red Hat is essentially an IBM-Intel joint venture and survived and made a successful IPO only due to IBM investments. And as you can understand other Linux-space companies do not enjoy the same level of IBM love, especially if they try to produce a hardware products. IBM also does not want any competition from Linux service organizations even for Linux of IBM mainframes:

Linux executives, speaking on a condition of anonymity, say the current way of doing business with IBM is not as lucrative as they thought it would be. For instance, Red Hat, SuSE and TurboLinux believe they are not being well-compensated for agreeing to support IBM's complete eServer line. At least one executive at SuSE, which supplied the mainframe Linux for a major ISP installation at Telia, says the company will be unable to recoup the costs of creating and supporting S/390 Linux under the existing deal.

SuSE's not alone. Some TurboLinux executives are having second thoughts about agreeing to support the complete IBM eServer line. In TurboLinux's case, company executives say it has its hands full with its own clustering work and attempting to be first to market with Linux for the Itanium.

... ... ...

The Linux companies' complaints aren't limited to not enough payback for supporting Linux on eServer. One top Linux company executive says Linux vendors do not get acknowledged publicly. "[While there is] a lot of value to the IBM name and that's helped with Linux acceptance, IBM also wants IBM to be the story, not Linux, and certainly not the Linux vendors," says the executive.

The last accusation is a little bit funny. It's like Pot calling Kettle black. There might be some install routines added by distributors to the S390 distribution and a price sticker on the box but not much else. Porting the Linux to the S390 were accomplished by IBM programmers.  In Jan 2001 IBM announced that it  will spend $300 million to create it's own consulting practices to advise corporations on Linux deploments. Like IBM installation of Linux on its low end servers eliminated "Linux hardware" companies including Intel-supported VA Linux, IBM's move to Linux consulting eliminated the possibility to extract profits from Linux services. Or at least closed the door to large corporate clients. 

And it was not accidental that in late 1991 VA Linux hardware division was closed and VA Linux repositioned itself as a software company. Linux companies that attempted to pump money into enlarging their service departments in hopes to get the Linux services revenues were hurt too. Linuxcare Inc. soon went bankrupt. I would like to remind that IBM service division is a one of the largest services group in the IT industry, and in early 2001 Dell inked a pact with IBM to offer Dell customers access to IBM's army of services personnel. The pact with Dell sends a very clear message that anybody else looking to make money on Linux-related hardware, software and services:  IBM is the  default Linux vendor in large corporate environment.  Please note how Robert LeBlanc VP of Software Strategy at IBM Software Solutions Division (who was part of the original team that evaluated Linux) commented on IBM's announcement a billion dollar Linux fund:

"The shift started about two years ago. An internal effort was started and I was also part of this team of 6-7 people. We wanted to know what this thing was and whether we needed yet another OS. We also looked at open source. Open source had far more benefits in terms of process. There was a lot of vigor in the process, lot of value in the process. It was a self-policing environment. If you did good work, you got noticed and got to do more good work. If you didn't pull your weight then you were shunned by the community. We also liked Linux. It was built right from the ground up. A small kernel which was very important as was the ability to add modules. Linux could scale upwards and downwards. The open source model was very good. We saw that Linux as an OS had very strong technical underpinnings. Another point to note was that customers were wanting a more open environment. We had just been through the Internet evolution and we saw the power of the Internet a lot before our competitors did. We were also members of the Apache project. The world did not need another web server, but the world needed a very good web server that could grow and expand. The world did not need another OS, but what the world did need was a common OS that was open source and scalable. We made the decision 18 months ago. We're way ahead of HP and Sun. So it's customer demand and we're giving them what they want."

Actually Linux distributors IPO gold rush (Red Hat, Caldera) that made Linus Torvalds a multimillionaire was by and large financed by Intel and IBM.  IBM is also the major producer of Transmeta chips.

That means that like Sun, IBM understands perfectly well that for departmental level severs and intranet web servers Linux is a reasonable choice and you better support it to get money from hardware than ignore it and give this business to a competitor (Microsoft). But as you scale up you will always be able to offer a much better proprietary solution to the customers. With 2.4 kernel Linux seems to be reliable enough both for the front-end and departmental servers. Support and maintenance are relatively cheap, and are improving.  And it will be some time before journaling filesystem, 32bit uids, RBAC, fully threaded networking code might eventually put Linux in higher end systems. Anyway it looks that IBM like Sun seems to be quite determined to preserve a respectable distance from Linux with its own AIX. IBM brass understands that for true enterprise level Linux up still needs a lot (NUMA, better fiber channel support, and better HA) and they are working in this direction...

On midrange level Linux still cannot compete with AIX: when Linux finally makes some milestone, AIX will be already at the next level. Actually keeping Linux as "me too" OS and substituting it for AIX solution for midrange server has its benefits that customers can try Linux and in case of disillusionment can always move AIX. The latter is definitely more attractive/reliable OS for running the enterprise critical applications. As the second after Sun holder of Unix 98 certification IBM is in good position to preserve and even enlarge technical gap with Linux anytime it wishes. As one IBM engineer joked on last Linux World: in certain key areas Linus will understand issues that IBM already resolved in AIX only in the next "Linux five year plan." An anti-Sun angle also was an important part of the game. For low end server business IBM hopes that Linux will harm Sun more than IBM as IBM sells Intel entry-level and midrange servers and Sun until recently does not (now after Cobalt acquisition Sun got its own line of entry-level servers). So each Linux server installation is a direct loss for Sun, but it might help to maintain a revenue stream for IBM's entry-level Intel servers. 

As I will discuss in the next section IBM's "Sun containment with Linux" strategy manifested itself in IBM-Intel partnership in the creation of a laboratory for developing and testing Linux on advanced hardware, the effort that has a distinctive anti-Sun flavor (as Bill Gates proclaimed it's much better to lose business to themselves than to the outsiders.)  But like Sun IBM position is ambivalent. Competition from open source products hurts. Moreover the Windows-baser server business at IBM is obviously much larger than Linux business. That means that Windows NT/2000/XP is and will remain a major part of IBM's portfolio. Linux is OK for making mainframes much more sexy, adding some revenue stream on departmental level servers and stealing a market share from Sun, but the IBM strategy is "about providing choice to customers"  as long as customer continue to use proprietary IBM products and not to replace IBM's product with open source alternatives. As Daniel Lyons noted in his paper Forbes.com Free Software Not For You, Baby:

"Given the amount of advertising that IBM does around Linux, I always smirk and roll my eyes when I'm in a competitive situation with IBM and I'm pushing Linux and they're pushing against Linux," says Martin Fink, vice president of Linux at Hewlett-Packard (nyse: HPQ - news - people ). "This happens frequently. Once or twice a month I'm seeing this. In an IBM account with lots of AIX, they do everything they can to protect their proprietary stuff."

HP delights in pointing this out, claiming it is more "open" than IBM. Indeed, though no company has done more than IBM to encourage big companies to consider using open-source software, now some rivals and open-source advocates are questioning IBM's sincerity. IBM declines to comment.

Originally developed by a loose-knit community of volunteer hackers and hobbyists, open-source programs like Linux typically offer most of the features that proprietary programs do, but at a fraction of the price. Plus, customers can view and even modify the source code of an open-source program, tailoring it to their needs; with proprietary programs, that code is kept secret.

IBM has touted these virtues. Yet right now, Fink says, IBM is fighting to keep one of the 50 largest companies in the world from switching to Linux, squaring off against HP, which is developing prototypes to prove Linux can handle that company's workloads.

"IBM is fighting tooth and nail to keep that company on AIX," Fink says. "We're talking to them about moving to Linux, and IBM is going out of its way to demonstrate why Linux is not ready for their environment. It's really twisted. IBM is doing all this marketing around Linux, and here I am competing against IBM using Linux, and IBM is using AIX."

If IBM is a little schizophrenic about Linux, it is downright hostile to other open-source programs, such as JBoss, an open-source Web server program that competes with IBM's proprietary Websphere program, and MySQL, an open-source database that competes with IBM's proprietary DB2 program.

"The dirty little secret is that IBM likes open-source up to the level of the operating system, but above that IBM does not like it," Fink says.

 

IBM boldly moves AIX into VM space

 

 "We are happy and comfortable with the idea
that Linux can become the successor not just for AIX
but for all Unix operating systems"

Steve Mills,
senior vice president and group executive
of the IBM Software Group
 (as sited in AIS, see also LTD)

AIX 4.x was a most proprietary of proprietary Unixes. Pretty strange and under-appreciated flavor of Unix with some interesting features in kernel space (Mach-based)  that was used primarily for government and (some segments) of banking markets (Solaris dominated in banking).  The last pre-Linux version of AIX, version 4.3, was one of the first first 64-bit Unix to achieve TCSEC (Trusted Computer System Evaluation Criteria) C2 certification from the National Security Agency in the United States and the analogous E3F-C2 certification from Germany's Federal Department for Security in Information Technology.

In 1998 IBM, SCO, and Intel launched a joint "Project Monterey," a cooperative effort to create an open, flexible enterprise UNIX platform for both  the IBM POWER processors and the upcoming Intel Itanium processors. That was called "Project Monterey" was essentially an effort of the Santa Cruz Operation (SCO) server software division (now part of Caldera) and Intel to develop an enterprise Unix to run on Intel's perpetually forthcoming Itanium 64-bit processors. Like all the other RISC processor-based Unix vendors, IBM spent a lot of research dollars on porting its AIX operating system to the Itanium architecture through Project Monterey, only to shelve the results after early performance statistics for Itanium reduced expectations and delayed customer demand. The years it took to see Itanium go from design to reality also coincided with the rise of Linux, which is much better suited to Intel architecture, and represents a more attractive marketing proposition to end users. As the Gartner Group noted[GartnerGroup 2/00]:

 "IBM claims that, with its help, Linux will improve and mature for broad enterprise use within five years. To build its credibility, IBM will form a new unit to identify appropriate technologies from AIX/Monterey and make them available for Linux and the open-source community. IBM will continue to differentiate Monterey and Linux through proprietary intellectual property that will remain unique to Monterey, but IBM predicts some kind of convergence (at the application-programming-interface level) of these two operating environments, although details are vague. ... At this point, IBM has not specified which technologies will be given to the open-source community and which will remain proprietary. IBM, however, is now presenting itself as integrator of Unix and Linux with open-system technologies... So far, no other system vendor (perhaps with the exception of SGI, on a smaller scale) has defined or assumed such a role."

IBM later shelved all Itanium efforts and changed the course to market the new OS as AIX 5L. AIX 5L combines the heritage of previous AIX versions with elements of DYNIX/ptx and UnixWare, as well as a important for marketing "affinity" to  Linux. In IBM opinion Linux on Intel is much more interesting to customers than proprietary Unix on Intel, but on RISK the situation naturally reverses. At the beginning AIX 5L was a pretty foreign for IBM OS that even did not enjoy much support within IBM (as of 1991 IBM still prefers shipping 4.x line). As the name imply if was a rather radical departure from the current AIX 4.x. And not exactly in the Linux direction: although the L stands for Linux it is mainly for marketing purposes. Some superficial things were done to improve interoperability, for example like Solaris AIX 5L also has AIX Toolbox for Linux Applications.  But the principal things of 5L looks more like an attempt to imitate key Solaris features: partitioning,  multipath I/O, NUMA (nonuniform memory access) and capacity upgrade on demand.  It also have a distinct IBM mainframe smell that always made AIX the most obscure Unix flavor in existence. 

In the race among Sun, the No. 1 Unix server seller and No. 2 HP,  IBM as No. 3, desperately needed the ability to split the server into several virtual machine "partitions": domains that each runs its own operating system and has an access to a share of common resources such as memory. The new AIX 5L  (starting with 5.2) fulfills this goal and was able to work with newer pSeries line of IBM's Unix servers, such as the p690 "Regatta", its sibling the p670, and the lower-end p630. Partitioning is the idea very similar to the idea of VM: it lets a computer perform several OSes at once, a useful feature when replacing a host of smaller servers with one big one. This "server consolidation" push is one project companies are undertaking, despite the down economy, and that brings AIX closer to the VM/Linux (see below): AIX dynamic LPAR technology is managed entirely in software, to the extent that multiple virtual instances of the AIX operating environment are abstracted from the underlying series hardware that powers them—an architecture that does sound a lot like VM on the mainframe. IBM borrowed extensively from its mainframe expertise when it designed the new AIX and probably will allow running Linux in a partition as well in the not so distant future.

There are many ways to partition systems. Sun, with its several years' head start over IBM and HP with Unix server partitioning, prefers physical divisions between partitions that ensure even major hardware failure in one partition won't affect the other. IBM, which is "Johnny come lately" in Unix server partitioning but was the pioneer of the VM approach with its classic VM/CMS system, prefers a software approach. Its "logical" partitions use an extra layer of software between the operating system and the hardware, a layer of abstraction that makes it easier to change the resources allocated to a partition. IBM's approach will get another improvement with AIX 5L version 5.3 coming in 2003. Currently, partitions can reside on a single processor, but changes in 5.3 are designed to allow up to 10 partitions per processor. This is very close to VM/CMS capabilities.

Although Sun positions itself as the Java leader, IBM with WebSphere is the major Java software vendor and it also expects to beat Sun to market. Java 2 Version 1.3 support is integrated in AIX 5L and in its plans to offer an improved Workload Manager utility that supports disk I/O management in addition to familiar CPU and memory management features. Solaris currently lacks the ability to manage disk I/O, which can be an important way for Web-critical applications and allow more effective exploitation of unused server capacity.

In September 2001, IBM forged an agreement with Linux vendor Red Hat to push Linux across its pServer product line. IBM plans to introduce Linux offerings for pSeries by the end of 2002 thus introducing the capability to manage mixed Linux and AIX environments and to run Linux on virtual machine provided by pSeries systems (IBM is extending VM capability to all pSeries Unix servers)

The new version of AIX allows administrators to create virtual servers on dynamic partitions that are as small as a single processor and have 250MB of memory. The smaller the partition, the more efficient the resource allocation, and potentially, the lower the total cost of ownership. The benefits of this are improvement to system utilization, a lowering of administration costs, the ability to add resources on the fly and the capability to manage servers better that competing vendors. But don't expect an Intel version. There will not be an Intel version of AIX 5L. While IBM produced a version of AIX that ran well on the first-generation [Intel] Itanium platform, we no longer build an AIX version for Itanium. They have withdrawn from that program and have no plans to rejoin.

IBM will also announce software package for managing mixed clusters of Unix and Linux servers. Clustered Systems Management (CSM) Version 1.3 provides a single point of control for installing, configuring, maintaining and updating IBM eServer xSeries servers running Linux and IBM eServer pSeries servers, or their logical partitions, running AIX. The introduction of these servers will make IBM the only major Unix vendor to offer a complete product line capable of running Linux in VM environment.

But all those Linux moves are not without danger. GPL is incompatible with most IBM licenses and that can harm IBM one way or another but as long as IBM is investing money in Linux the danger is mainly theoretical.  Robert X. Cringely aptly made this point commenting about the other major open source project that IBM supports -- Apache: 

Here is the core argument: There are a thousand Open Source projects that get started out of need or fun, are maintained for awhile for fame, then get abandoned because there is no reason to go on. Eventually, the programmers come to understand that "users" are people who yell at you to fix stuff. So Open Source is inherently flawed. It only works because otherwise unknown programmers can get 15 minutes of fame using the Internet as low-barrier entry into introducing their skill to the world. Since they are introverted nobodies, getting a few emails from unknown users that say "good job!" feels great. But in time, most Open Source projects grind to a halt. The ones that survive are projects like Linux and Apache that have substantial involvement by PAID engineers. One could argue, in fact, that the idea of Open Source software being created by volunteers is a misnomer. Even Linus Torvalds is paid by Transmeta to be the God of Linux.

Open Source has value or people wouldn't still be doing it after 10 plus years. At the same time, complexity breeds inefficiency. Whatever approach you take to the organization of product development some form of the 80/20 rule applies -- 80 percent of the available material is useless to you. We can easily just dismiss the creaking parts of Open Source by bunching them in the 80 percent we ignore.

But ignoring them does not make those parts go away, and here is where we’ll find Open Source’s vulnerability. There is this idea (I’ve written it myself) that Microsoft, for example, can’t compete with Open Source because you can’t compete with a product that has no profit Motive, and can’t out-market a product that has no marketing budget or plan. But Microsoft could still beat Open Source simply by subverting it.

It is possible to hijack an Open Source project since any Open Source team will automatically bend itself around the party doing the most work. What I find most interesting, however, is applying varying motives to the hijacking. What if Microsoft, for example, suddenly started devoting a lot of resources to Open Source development? They could throw a team at all the key projects. But why would they do that? Well, IBM is already doing it. IBM has hired most of the Apache team. IBM has some major pull on what work gets done and does not get done. In some cases, it is frustrating, and other cases not. However, everybody just accepts it because IBM is paying the bills and people can do what they love. Is there an official IBM party line at Apache? Absolutely not! It is just that none of the Apache developers will talk negatively about IBM, even those that do not work at IBM. So in this sense, it already appears that Apache has been hijacked.

Destroying old IBM

But not everything good about this IBM's top management sponsored move to Linux. It started a trend towards offshore outsourcing programming jobs to other countries, China, India and Brazil. As the flavor of Unix able to run on the most cheap hardware Linux is ideally suited for outsourcing programming jobs. There were rumors that IBM Global Services plans to outsource to India more than half the information technology positions it has in the USA. IBM would neither confirm the existence of the plans nor comment on their reported contents.

Though many economists say this cost-cutting will help businesses be more productive, leading eventually to more U.S. jobs, it's unwelcome news for thousands of US programmers, who blame offshore outsourcing for a nagging lack of IT jobs and stalling wage growth. Still IBM leadership wants to "leverage the Asian economies lower cost structure," where a company can get "two heads for the price of one.": IBM pays U.S. programmers with three to five years experience about $56 an hour, including benefits. Chinese programmers with similar experience make about $12.50 an hour, including benefits.

A typical outsourcing deal usually starts with moving existing employees are becoming IBM employees, "...and then let go later like all the other companies that did this before". And IBM scored big wins at companies such as Morgan Stanley, Royal Dutch/Shell Group, etc. Here is one example from the paper by Mary Hayes, InformationWeek

The initial stages of the seven-year, $5 billion outsourcing relationship between J.P. Morgan Chase & Co. and IBM Global Services has included a fair share of misunderstandings, differences of opinions, and complexities. Still, John Schmidlin, managing director and chief technology officer at J.P. Morgan Chase, said today that the company's outsourcing of its IT infrastructure to IBM was a wise decision that he wished he had made years earlier.

Schmidlin, speaking at Gartner's Outsourcing Summit 2003 conference in Los Angeles, said the contract with IBM went live April 1 and included moving 4,000 employees to IBM Global Services. The financial-services firm began its search for an IT services provider two years ago as part of its strategy to cut costs and eliminate distractions not related to its core businesses. "We are not a technology firm," Schmidlin said. The company views technology as an "adjunct to the services we provide to clients."

When evaluating IBM and two other finalists, J.P. Morgan Chase considered factors such as a services provider's ability to understand J.P. Morgan Chase's transformation of its IT infrastructure--including server consolidation and on-demand computing--and offer a detailed a plan for achieving certain milestones toward that effort.

J.P. Morgan Chase also interviewed business-technology executives at other companies who had outsourcing contracts in excess of $1 billion, all of whom warned of bumps related to the transformation of people and processes. Schmidlin extended similar advice. "You'll frequently have misunderstandings, and there'll be differences along the way," he said. "Make sure you have a process in place to deal with the conflicts."

Still, said Schmidlin of the contract with IBM, "If I had to do it all over again, I would have done it five years ago."

Here is the questionnaire from IBM that outlines their strategy:

How do you know if outsourcing is right for your business?

If you can identify with the any of the following situations then it is likely your business is ready to outsource some business functions:

As IBM wrote in  IBM outsourcing A primer for healthcare managers

An emerging technology trend has been the practice of outsourcing IT activities to service providers. Examples of information technology services that are now being outsourced include the management of networks, data center services, call center services, Web hosting and application services.

This important report describes the factors driving this trend and provides a five-phase process model that healthcare executives can use in making the decision to outsource IT activities and then implementing that decision if the outsourcing option is selected. Phases in the process include determining a sourcing strategy, analysis of sourcing needs, vendor selection and contract negotiation, transition to the service provider, and managing the performance of the service provider.

Staring with 2001 low morale in IT shops started to spread and in 2004 is reaching a critical point. Here is a typical letter:

I am a former programmer now entering the MBA world (one of these schools: harvard/wharton/stanford).

[I'm] pretty sick of the devastated software economy. All the people I respect have moved on to something else. The only ones left are those that cannot leave because they are too old or too clueless.

5 years ago, hi-tech attracted some of the smartest and most talented people available. now they attract the socially inept, or those with no ambition beyond writing 200 lines of code per day in a cubicle. I am excluding from this statement those that do research for a PhD in engineering - they are in a different world.

As we all know, this whole "do more with less" approach to IT (indeed across all businesses) has put a lot of strain on work and home relationships. Offshoring and outsourcing, the uncertainty of the business environment, and heavy workloads are all contributing to morale issues. Of the 650 companies that META surveyed for the study, 72 percent said that low IT employee morale is a serious issue.  That also influenced the attitude to the IBM's embracing of Linux and in general attitude toward IBM's technology:

As a technology consultant, I have influenced many corporate customers to purchase IBM products and services for more almost two decades. I have usually been very pleased with the response, value, and quality received. (As an independent eProductivity consultant, I do not profit on the recommendation of one vendor's products over another. I am simply paid for my expertise to design what I believe to be the best solution for my client's stated objectives.) Outside of this, the only tangible benefit to me personally is another happy client -- which is worth more to me than any consulting fee I may receive.

Unfortunately, this is changing, and I believe that the IBM's decision to outsource and their method of doing so is at least partly to blame...

... ... ...

Disclaimer: This is not an essay about the evils of outsourcing offshore to some group that does not speak English fluently. Nor is this a rant about the potential costs in terms of U.S. jobs lost as a result. This not a rant against any particular individual at IBM or any of its outsourced divisions. This is a rant against the experience of being a technology consultant who has to deal with an outsourced organization that apparently does not understand the products, information, and process of serving its customers. The outsourced "IBM representatives" that I dealt with were --  to their credit -- all English speaking and all very polite; and, they usually responded to my email or calls within minutes. I have come to realize that these people were doing the best that they could do with what little information (or perhaps product training?) that they had access to.

The problem appears to be a complete lack of information and knowledge to complete the sale or satisfy the customer, and the unwillingness of IBM management to get involved. The walls between the outsourced organization and IBM are apparently quite high. This reminds me of why I do not shop at a certain chain of electronics store -- well-dressed people, some excellent products, but little or no information or experience on the part of the people I have to deal with to complete the sale.

In a move apparently designed to quell fears about widespread outsourcing IBM has established a 25 million worker retraining fund. Samuel Palmisano announced the fund at the company's annual IBM Partners World event in Las Vegas. Interestingly enough, Palmisano defended outsourcing, citing the need to improve quality of life and standards of living in countries with still-developing economies. The fund, dubbed the Human Capital Alliance, will be made available to IBM workers whose current positions or skill-sets are being outsourced to such places as India, Pakistan and China. IBM said that for the next two years, it will focus on training workers in areas where it sees the most demand, but its unclear where the demand might be. While more and more software corporations say they must take advantage of outsourcing opportunities if they are to compete and stay profitable, U.S. programmers worry that their the jobs being exported and that the trend is for more and more better-paying jobs moving oversees.

Copyright © 1996-2008 by Dr. Nikolai Bezroukov. www.softpanorama.org was created as a service to the UN Sustainable Development Networking Programme (SDNP) in the author free time. Submit comments This document is an industrial compilation designed and created exclusively for educational use and is placed under the copyright of the Open Content License(OPL). Original materials copyright belong to respective owners. Quotes are made for educational purposes only in compliance with the fair use doctrine.

Standard disclaimer: The statements, views and opinions presented on this web page are those of the author and are not endorsed by, nor do they necessarily reflect, the opinions of the author present and former employers, SDNP or any other organization the author may be associated with. We do not warrant the correctness of the information provided or its fitness for any purpose.

Last modified: August 27, 2008