August 13, 2004
IBM has announced z/OS 1.6, the latest iteration of the company’s mainframe operating environment. z/OS 1.6 supports IBM’s zSeries Application Assist Processor (zAAP), a specialized JAVA execution environment, and enhances new application technologies, including z/OS UNIX commands, 64-bit C/C++ application development support, and C/C++ compiler options for z890 and z990 servers. Other improvements include support for up to twenty-four processors in a single z/OS LPAR on a z990 at general availability, followed by 32-way support planned for 2005. In addition, IBM announced z/OS.e 1.6, the latest edition of a specially priced OS for new applications on z890 and z800 servers. z/OS 1.6 is the first version of z/OS to require the z/Architecture, and runs only on zSeries servers (z800, z890, z900, z990). z/OS and z/OS.e 1.6 are planned to be available on September 24, 2004. No pricing information was included. Separately, Endress+Hauser, a leading process control device manufacturer, announced that it has deployed one of Germany’s largest Linux-on-mainframe platforms at the company’s Weil am Rhine facility. During this deployment, the company upgraded its ERP applications from SAP R/2 to R/3, and migrated its existing SAP R/3 systems from UNIX to a Linux on IBM mainframe platform. Overall, E+H moved nineteen production SAP systems onto fourteen LPARs, and shifted related SAP databases on IBM DB2 onto six LPARs. The application servers are using thirty-six IFLs (Integrated Facility for Linux) running on two IBM eServer zSeries z990 systems, and support over 3,500 users worldwide.
As the longest-running franchise in IT history, IBM’s mainframe solutions face the sort of unprecedented challenges and opportunities anyone does if they lead the pack for 4+ decades. The challenges, however, while reasonable enough on the surface, are also entirely predictable. In an IT industry obsessed with the sexy new gizmos and religiously assumed progress, a gracefully aging and eminently vital technology is not just a curiosity but also an openhanded affront. IT pundits and IBM competitors who could not find, let alone create, a similar solution with both hands, are quick to dismiss mainframes as anachronistic relics of IT’s prehistoric age, but such yammering is more reflective of fossilized perceptions than the state of mainframe solutions. Interestingly enough, enterprise customers appear to be pleased as punch with IBM’s eServer zSeries offerings, and have contributed to the 40th birthday celebration by lighting the candles on notable quarter-over-quarter mainframe sales leaps since the company introduced its enhanced z990 and z890 systems last year.
So what’s the deal, anyway? How can a supposedly moribund technology continue to be so darned lively? A couple of reasons. First, while IBM provides continuing support for traditional mainframe markets and workloads, the company also constantly explores new vistas for mainframe (and mainframe customer) opportunities. z/OS 1.6 offers a prime example of this strategic effort, with its zAAP processor environment and C++ options and support. In addition, the company continues to leverage its mainframe investments and experience to drive innovation across the rest of its product lines, such as IBM’s ongoing efforts in virtualization. What does this mean overall? That when customers, such as Endress+Hauser, are looking for solutions that offer up-to-date technologies supported by a foundation of traditional reliability, security, and flexibility, they turn to IBM. Some self-appointed IT paleontologists continue to define the mainframe as a dinosaur, but they are deeply misinformed. Dinosaurs died out, in part, because they were unable to adapt to changing circumstances. The IBM mainframe survives and thrives because it continues to evolve and, with developments like z/OS 1.6, to define the leading edge of enterprise IT.
The People’s Republic of China announced this week that it is forming an open source alliance between a number of Chinese software companies and notable U.S. vendors such as HP, IBM, and Novell. The new alliance includes Chinese Linux vendor Red Flag Software, which has been successful in pushing for open source Linux standards in China and Asia at large. Red Flag has been distributing Asianux, an enterprise version of Linux for the Asian market. Earlier agreements between China, Japan, and Korea formed last year were also planning on developing and distributing Linux alternatives to Windows in those countries. Meanwhile, Microsoft announced that it would begin shipping a stripped-down version of Windows XP to markets through local OEMs in Malaysia, Thailand, and Indonesia, as well as two other countries yet to be identified. The Windows XP Starter Edition will be priced at $36 per copy, and will offer limited functionality, including limited open windows, no PC-to-PC home networking, 800x600 maximum screen resolution, no sharing printers across a network, and the ability to create multiple user accounts on a single machine. Microsoft stated that the Starter Edition is presently a twelve-month pilot program to valuate the market response to Starter Edition and to determine if it will be rolled out in other countries.
Asian countries are continuing to demonstrate their insistence on maintaining autonomy and control of their domestic IT markets by attempting to limit the intrusions of external technology into their markets. By embracing Linux and open source, countries across Asia are seeking to build their own IT offerings for what will eventually will be the largest markets in the world for IT products. Companies such as IBM, Intel, HP, and Novell are hoping that by joining a friendly alliance of Chinese partners, they will gain access to this potentially lucrative market, even if it means unprecedented concessions to these partners to gain access. Even so, a small sliver of the Chinese market could be a very substantial revenue stream for any or all of these companies.
Microsoft continues to take a different path; one that seems to be more reactive to Linux in these markets than anything else. The Thai government has been operating a Linux development project and Microsoft apparently decided that it needed to come up with an alternative that could at least slow the Linux adoption rate in that county. We suspect that the Starter Edition gambit will largely be a non-starter, as Internet access will demonstrate very quickly the shortcomings of this stripped-down version vis-à-vis the full XP package to users in these three countries. Will they pop the extra baht to get the full version? Probably not, and instead turn either to Linux or to bootleg versions of XP readily available for pennies on the dollar. The contrast between Microsoft’s approach and those of the Chinese software alliance could not be more dramatic and instructive. Microsoft continues to try and force entities to accept their proprietary products in an effort to grab dominant market share through vendor lock-in. After seeing increasing numbers of North American enterprises seeking alternatives to such a long-term, sole-source vendor, should it really come as any surprise that sovereign nations might feel the same way as well?
Microsoft announced that it will be releasing Service Pack 2 for Windows XP in the coming weeks. The full service pack is approximately 256MB, but most users running Windows XP Home Edition or Professional Edition will have downloads of between 80 and 100MB, if they have installed the SP1 and regular updates that have been distributed. The initial version of SP2 will be available in English only; versions in some twenty-five other languages are due out within the next two months. Microsoft stated that SP2 is a major upgrade in computer security, wherein a number of default security settings will reflect a newer urgency in this area. While the release of SP2 is imminent, many enterprises indicated that they would not deploy SP2 until it has gone through significant testing to determine what impacts the service pack will have on existing application and infrastructure integration.
While there is no doubt that improved security for the XP operating environment is becoming increasingly critical, we suspect that despite the urgency surrounding the security issue many enterprises are going to avoid being the first on their block to install the service pack without undergoing significant testing. Given that this service pack more resembles an OS upgrade than a collection of patches and bug fixes, we believe that such caution is warranted, especially considering the mischief that new security settings could have on key enterprise applications and data transparency. Closing security holes is generally a good thing to do, but if in doing so one hampers access to business-critical data or applications then the value of that improved security is taken into the net negative.
In our mind the legitimacy of such concerns in unassailable. It makes a great deal of sense for enterprises to seek the least disruptive means by which to deploy a significant OS upgrade. At the same time, however, we believe it is important to note that such an upgrade would have been at least unfeasible five years ago, and in all probability impossible. The evolution of software upgrades from hard copy floppy disks, to CDs, and now automatic downloads would not have occurred were it not for the proliferation of persistent, high-speed connections in both homes and small businesses. These connections allow for the downloading of upgrades that by themselves dwarf the earlier versions of entire operating systems. For most users, the download of SP2 will be a largely invisible activity happening in the electronic background as they pursue their daily tasks at the keyboard. While the specific issues surrounding the content of SP2 may stir reasonable debate, there should be little acrimony over the ongoing value proposition of persistent high-speed connections in the present and well into the future.
The risks of patent assertions against Linux were recently noted in a study conducted by the Free Software Foundation (FSF). The patent attorney that did the Linux review found 283 registered software patents. A third of the patents are owned by Hewlett-Packard, IBM, Novell, Oracle, and other Linux backers. However, Microsoft owns twenty-seven of the patents, many if not all of which could be used as the basis of patent lawsuits against the Linux kernel. The study concluded that the existence of the patents alone would not kill Linux, but it is inevitable that Linux will face a patent challenge in the future. FSF noted that the findings would be similar for any other widely used software program that underwent patent investigation scrutiny; however, open and public domain software are more vulnerable to patent assertions because programs and code are openly available for review against existing patents.
The Free Software Foundation’s study has already had an impact on the plans of the various Linux constituencies and will continue to have in the future. For example, the City of Munich had earlier announced that it planned to replace its 14,000 Windows Desktops with Linux, but in view of the above news, the Bavarian government has put the project on hold. Vendors are aggressively scrambling for position, impact, and profitability in the open source world but their actions may be slowed by the requirement to identify and resolve patent issues. Vendors with a strong Linux commitment and strategy such as IBM and Novell have all but indicated they will use Linux patents only defensively and not offensively. Microsoft’s Steve Ballmer and Bill Gates, in a recent meeting with Wall Street Analysts, indicated that the commercial software industry faced an Open-source risk that threatens the sector’s profit in the next decade. At the same time, they indicated the need for increased focus on protecting intellectual property and predicted a dramatic increase in the number of Microsoft patent filings. There are major PR and other issues for Microsoft, should it decide to defend its business via patent assertion. The above are a sign of things to come in the Linux and open source world.
One must view the FSF study with some skepticism. The Linux kernel is complex and there are a large number of software patents that it must be compared with. It is difficult if not impossible to match all the applicable patents to the Linux kernel. Nevertheless, identifying and resolving patent issues remains a critical component of every major open source project. This is an issue, but the real core problem is deeper and more fundamental. The present U.S. patent system and body of law is flawed. One of the necessary changes in the system is to stop issuing patents for minor process and technique discoveries. For example, the Amazon.com patent for one-click purchasing would not be patentable under a reformed patent code. Granting Amazon-type patents complicates and stifles innovation. In fact the Public Patent Foundation is lobbying for reform of patent laws: they claim that half of the patents issued by the U.S. Patent Office contain no innovation. Another problem with the current patent environment is that companies with major financial resources, like Microsoft, are able to acquire a large number of patent assets. These organizations can then use their patents to combat smaller, more innovative competitors, as well as the Linux and other promising open source efforts. Therefore, it is likely that Linux will face a patent challenge. This may slow Linux adoption; however, we believe that the Penguin will ultimately prevail.
This week Hewlett Packard made a bid for European
services firm Synstar. HP have
offered $300 million in cash for the company, which has over 1,500 customers in
eight countries in
In May this year, HP merged the servers and storage, software, and services groups together to produce the technology solutions group. They also created a customer solutions group. These groups are tasked to create products and solutions that respond to customer needs and combine the appropriate hardware, software, and services, rather than focusing on individual brands of products. This is a different approach that that of IBM, for example, who keep hardware, software, and services in separate divisions. While HP have what is left of the legacy of Digital’s services, that group was mostly focused on infrastructure maintenance, and HP must extend service beyond that if they want to execute their Adaptive Computing vision with any real credibility. Purchasing a company with presence in multiple countries will help them more than attempting to grow organically. So far, HP appear to be setting themselves up to provide customers what they need in the ways in which they need it. The question naturally is one of execution.
HP’s latest results indicate that while services has been successful in the last quarter, and managed services — a specialty of Synstar — was the fastest growing part of services, servers and storage have suffered. In particular, they are seeing much less Alpha and NonStop business, and storage has suffered as well. IT managers are focused on getting the most out of their infrastructure, and managed services — which involves handing day-to-day management of some systems to a service provider — is one way to accomplish that. Both IBM and HP have seen some large deals from outsourcing and managed services, which mean these customers have given over at least partial control of data centre operations to those vendors and trust that these vendors have the capability to run some or all of IT operations. In a market where individual products contribute less and less to vendor margins, it is crucial that HP develop the services business that can help customers find business value in their IT infrastructure.