by Timothy Prickett Morgan

The computer industry was rocked back on its heels last week as Hewlett-Packard, one of the oldest names in the IT industry, acquired rival Compaq, one of the biggest names in the IT industry, for $25 billion in stock. While members of the media obsessed about the ins-and-outs of the megadeal with respect to PCs--something they could get their minds wrapped around--the simple fact is that the PC businesses of the combined HP and Compaq companies only account for about a third of its $87 billion in revenues. The PC business is a commodity business. Both HP and Compaq rely largely on contract manufacturers to build their PCs, and they are built predominantly from third party components that neither HP nor Compaq have much of a say about or much differentiation to offer. That business is never going to be hugely profitable again, thanks to the nature of competitive pressure and volume economics, so Wall Street had just better get used to it. And while Wall Street analysts and the IT press universally bemoaned the fact that even combining HP's and Compaq's respective services businesses resulting in a company that had half the services revenues of IBM, which rakes in over $30 billion a year in services, I don't much care about services in the short run because there is nothing short of acquiring several small, niche services companies that HP can do to try to compete with Big Blue in this area. What is interesting, however, is what happens if a united HP and Compaq can sort out their respective server product lines and maintain their revenue streams. If HP can do that, then it will be the biggest server vendor in the world, and by a wide margin.

I spoke to a number of HP's top brass last week, and they are in no position to produce product roadmaps that show where Compaq's and HP's product lines will be merged and where products will be pruned entirely from the company catalog. The most that any exec would commit to was that HP would have a battle plan together for the combined HP and Compaq products within 100 days, including how the channels, direct sales force and HP Web site would be used to push those products. There's plenty of speculation that HP is going to kill off Tru64 Unix and OpenVMS, Compaq's Unix and proprietary operating systems for its AlphaServer machines. No one is talking at all about the high-end, fault-tolerant Tandem Himalaya line, but this $2 billion hardware, software, and services business is going to figure prominently in HP's long term plans. Similarly, HP has not said what it will do with its own HP-UX Unix platform, and it has left customers in the HP 3000 line hanging, wondering about its plans for the MPE proprietary operating system that runs on the same PA-RISC hardware that the HP 9000 Unix server line uses. Frankly, the HP-Compaq deal didn't cause any of this consternation. Compaq's and HP's customers have been worrying about these issues for years.

What seems clear is that HP is platform agnostic--it will sell Unix, Linux, Windows, and proprietary environments to different server customers for different workloads--and that it, like Compaq, is trying to get to a common hardware platform that can support as many of these environments--either simultaneously or separately--as is technically possible. Only two months ago, as the HP-Compaq merger talks were underway, Compaq killed off the Alpha RISC processor it had been peddling since 1992. Compaq sold the Alpha chip design and all of its intellectual property rights to Intel in exchange for Intel using its chip-making plants to pump out the next few Alpha chips on the AlphaServer roadmap. Compaq has plans to deliver the 1.7 GHz Alpha EV7 21364 processors in the "Marvel" line of 256-way servers sometime in late 2002 or early 2003. Compaq has already committed to porting the Tru64 Unix and OpenVMS operating systems to Intel's "McKinley" or "Madison" generations of 64-bit Itanium processors sometime in 2003, with support for the Tandem NonStop environment on a similar platform by 2004.

With the HP acquisition of Compaq, this server roadmap will definitely change, but not in any substantive way. Instead of porting Tru64 Unix, OpenVMS, and NonStop to a Compaq-derived Itanium platform, how HP and Compaq will likely come up with a unified Itanium platform that makes the best use of all of Compaq's, HP's, and Intel's technologies to create a single environment that will support these platforms. HP may decide to run Tru64 applications in API or ABI environments within HP-UX servers, much as it does for Linux applications today. (Unlike IBM, HP does not directly support Linux on its RISC Unix servers.) Alternatively, HP will support all the different operating systems in the Compaq and HP catalogs natively on a unified Itanium line. What seems clear is that HP cannot support distinct product lines--in terms of the underlying hardware--and that it had to pounce on Compaq before it got too far down the road developing Itanium servers of its own. As far as timing goes, HP has acted intelligently in buying the Compaq server line. Whether or not it can deliver servers that do not alienate OpenVMS and MPE customers, as well as Tru64 and NonStop customers, remains to be seen. HP-UX already runs on Itanium hardware, so at least that much is done.

Assuming that the merging of these two behemoths proceeds according to plan, with closure sometime in the next six to eight months, and HP and Compaq can keep their server channels and customers from freaking out--these are very big assumptions, I realize--then what HP will get is a server business that dwarfs that of IBM, Sun Microsystems, and Dell. HP will knock IBM out of the top spot in the server business, something no other vendor in the history of the server business has been able to do.

Look at server rankings by International Data Corp for calendar 2000. HP-Compaq had just shy of $19 billion in server sales last year out of a total worldwide market of $60 billion in worldwide sales. That's a 31.5 percent share of the market. In 2000, IBM sold $13.6 billion worth of servers and garnered a 22.5 percent share of the market. IBM's market share has had its ups and downs over the past few years, but the general trend has been downward as OS/400 and mainframe sales are flat or down in a growing server market. IBM's aggressive moves in the Unix server market and fair showing in the Intel-based server market have not really reversed that trend. The HP-Compaq share of the server market, by contrast, has been growing by a few points a year over the past year. While Sun owns the Unix server market, beating out IBM and HP two-to-one in aggregate revenues, HP-Compaq had Unix server sales of $9 billion in 2000 compared to Sun's sales of $10 billion. Dell had about $3.4 billion in server revenues in 2000, with all but about $1 billion of that coming from sales of Windows-based servers. HP-Compaq sold twice as many Windows servers, with $5.6 billion in sales in that same year.

by Shannon O'Donnell

There has been a considerable amount of talk in recent months over the fact that OS/400 V5R1 will not support IBM's Office Vision/400 office automation suite of programs. If you're not familiar with OV/400, this was IBM's first attempt at office automation software that we would probably call groupware today. Developed as a collaboration product for the System/34, OV/400 has been a mainstay for document processing, messaging, and calendaring for decades. IBM tried to extend the program to OS/2 and mainframes in the 1980s and early 1990s, but that development never panned out and it failed on those platforms. But OfficeVision worked fine on AS/400s, and it became popular on the OS/400 platform.

When IBM announced a few years ago that it intended to drop support for this product, there was an understandable cry of outrage from the AS/400 community. Thousands of companies worldwide had embedded OV/400 functionality into the very heart of their applications for everything from document printing to database access. If IBM wasn't going to support the product, what were they supposed to do? Many in the AS/400 community just flat out didn't believe that IBM would follow through on this and that there would be a last minute reprieve. IBM, however, has been clear that it wants OfficeVision customers to move to Lotus Notes and Domino.

Despite this, IBM did offer a reprieve of sorts to OfficeVision customers. OV/400 was scheduled to be removed from OS/400 when V4R4 was released. However, for reasons best known to IBM, its shelf life was extend through one more release to V4R5, albeit with the very strong warning from IBM that they really were going to drop it in V5R1. Extending the product this way, after all the press announcing that OV/400 would not last beyond V4R4, may not have been the best move for IBM. Many in the industry figured that if IBM extended OV/400 from V4R4 to V4R5, then they'd probably extend it from V4R5 to V5R1 as well.

It turns out that IBM meant what they said this time. If you upgrade from V4R4 or V4R5 to V5R1, the OV/400 product you had installed will be deleted. If you find that you must absolutely have OV/400 on your V5R1 system, there is a way to go back and install it. To do so, either you'll need to have the OV/400 Licensed Program Product on CD, or you'll need to have saved it before the upgrade to V5R1 using the Save Licensed Program (SAVLICPGM) command. On your V5R1 system, turn off digital signature checking and then you can install the OV/400 program.

This method will allow you to continue to use the OV/400 product at V5R1. However, if you use OfficeVision in this manner, you are doing so at your own risk. In a recent conversation with IBM developers at Rochester, I was told that anyone who installs OV/400 on a V5R1 system might get an unexpected and unpleasant surprise in the not so distant future.

According to my sources, one of the reasons that support for OV/400 was dropped in the first place was because most of the people who had developed and supported that product within Big Blue have either retired, quit, or moved on to other areas of the company. As time went on, no one at IBM really wanted to be associated with the OV/400 product because it was perceived, and rightly so, as old technology. So, it became harder to get people to work in the OV/400 testing and support area. As new versions of OS/400 were cranked out, that meant fewer people were testing the OV/400 product to ensure that it worked with every component of OS/400. Finally, there was no one left who either knew how to or wanted to work with OfficeVision on the development side. IBM obviously does not see OV/400 as a strategic future product worthy of investment, so there has not been a push inside the company to assign people to this task. This lack of testing for OV/400 with the other components of current and future OS/400 releases could cause you lots of headaches and crashes in the future.

OS/400 is based on the concept of shared components. That is, a component written to support database access, for example, may also be used by the Java Toolbox, Operations Navigator, and even OV/400. When developing and testing a new release of OS/400, IBM devotes resources to ensuring that when a change is made to any shared component, that component continues to function as intended with every other product or component that uses it. If IBM does not have the money or people to test OfficeVision with every other component of OS/400, then problems could arise. It is almost inevitable that at some point a developer at Rochester will make a change to a shared component and that shared component will no longer function correctly with OV/400. The only way you'll know that this happens is that OV/400 will no longer work.

If you're an OS/400 shop that is counting on still being able to use the unsupported version of OV/400 at V5R1 and beyond, you may wake up one day to find that your operations staff has installed a new PTF overnight and suddenly the Edit Document (EDTDOC) or Print Document (PRTDOC) commands no longer work. The point is that because there is no one left at IBM who is willing or able to test the integration of an unsupported product with the rest of the shared components in the OS, you can not ever tell when the OV/400 product you're depending on to support your mission-critical applications will stop functioning.

The developer source I spoke to at IBM Rochester had this advice: If you are at V5R1, you can still--for now--install the unsupported OV/400 product, using the method I outlined above. However, you should only do so if installing this unsupported product is part of a larger plan to migrate your OV/400 document processing to another solution. Whether that solution involves Lotus Domino, Microsoft's Word, Inventive Designers' DTM/400 or any one of several document processing solutions available on the market is less important than the need to implement a firm plan of action for moving away from OfficeVision. Your business is simply too critical to rely on a product that could stop working at any time without any warning from IBM.

by Timothy Prickett Morgan

The iSeries product line has won two relatively prestigious awards in recent weeks, and given all the negative news that is going on out there in the world, I thought I would give you a positive story for a change.

The iSeries product line has won the top spot in the VARBusiness Annual Report Card for the midrange category. The VARBusiness awards, which are given for all sorts of categories of products and subcategories based on features and business practices, are generated through surveys of over 5,000 resellers and partners of the major vendors in the IT industry. The xSeries line won the top spot for entry level servers.

A few weeks ago, the iSeries was also awarded best-of-show at the LinuxWorld Conference & Expo in San Francisco. This is something of a coup for the iSeries, given the fact that the open source community, generally speaking, has no idea what the AS/400 was or the iSeries is. Clearly, IBM's support of Linux on the iSeries--and the innovative manner in which it was supported--merit recognition by the Linux community, and it has justly received it. Back in February, IBM's zSeries mainframe line took best-of-show at the LinuxWorld trade show in New York. IBM rolled out native support for Linux on the mainframe six months ahead of native support for Linux on the iSeries.

Save TIME AND MONEY with our AS/400 - iSeries Report & Data Delivery Systems RJS Software Systems, Inc. - http://www.rjssoft.com Delivering AS/400 reports via email, web, Lotus Notes, Domino or CD. Whether it's Native AS/400 or PC-based, we have the solution. * WinSpool/400 report download and conversion * Text converters (ASCII, RTF, PDF, HTML, Spreadsheet) * Form converters (AFP, EZ/Print, Formation,    Formsprint, CreatePrint to PDF) * Split reports with AS/400 Report Splitter * Email Report Server, Web Report Server * Domino Report Server * Batch Report Server delivery to LAN, IFS and FTP * SMTP/400, POP/400 & FTP/400 Email/Transfer APIs * Query and Convert AS/400 Files with DataExport/400 * Web-Based AS/400 Data Access via Active Server Pages For a FREE FULLY FUNCTIONAL DEMO CD, please visit our Web site at http://www.rjssoft.com. Contact us at 888-RJS-SOFT or email us at sales@rjssoft.com

by Timothy Prickett Morgan

If you didn't buy a server between the beginning of May through the end of July, you appear to be in good company. According to market statistics compiled by International Data Corp, server revenues plummeted by 16 percent in the second quarter as shipments declined by 3 percent on a worldwide basis across all server types.

Server revenues declined by 4 percent in the first quarter of 2001 to $13.3 billion according to IDC, and fell 16 percent in the second quarter to $12.1 billion. That also represents a sequential decline of 9 percent, and considering that the first quarter of this year was nothing to brag about, the rather dramatic decline, which was worse than the analysts at IDC had expected, is probably causing a lot of consternation among the server vendors of the world. But just so we keep our perspective, remember that shipments were only off 3 percent, which seems like a miracle.

The analysts at IDC say that they believe what kept shipment numbers from crashing is the rollout of inexpensive Intel-based servers and low-end RISC Unix servers that were typically bought by dot-coms and telecommunications companies over the past couple of years for their Web infrastructure. There were never more than several thousand dot-coms and telecom companies in the world, and there are millions of traditional companies who are now individually, on a much more modest scale, building out their own Web infrastructures. They are buying rack-based, high-density edge servers for these purposes, just as dot-coms did. Companies that depend on midrange and enterprise servers also continue to upgrade their existing machines and swap in whole, new machines to support their growing workloads. Still, shipments of rack-mounted servers were up 46 percent, but revenues were only up 13 percent, showing how hot and cutthroat this market is. Revenues in regular tower servers were off 23 percent, and these machines still make up the bulk of server sales.

Companies are still investing in new applications--particularly supply chain management and customer relationship management suites, which promise increased sales and better profits with a relatively fast return on investment (whether these apps will deliver on that promise is another matter)--and that is driving server sales. Companies are shopping around, driving down server prices as well. All of these factors contribute to the overall revenue decline in the worldwide server market. The question now is whether or not Q2 2001 was the bottom of the dip in the server market, or if Q3 will look even worse. With Japan in recession and Europe and the United States teetering on the edge of recession, no one wants to make any predictions. Except me, of course. I would not be surprised if the economies worsen still and the server market shrinks back to $50 billion or, if it gets really bad, $45 billion in 2001 compared to $60 billion in 2000. And even if the economies of the world improve, I think server spending will lag at least six months behind optimism. Even if Q2 was the bottom and we are in a rebound, because of the price wars in the Windows, Linux, and Unix server markets, vendor factory revenues for servers--which is what IDC is counting--will probably not be able to rise above $55 billion for 2001. That's still a 9 percent decline for the year.

For the second quarter of this year, IDC figures that the server market in the United States declined by 25 percent, Japan by 20 percent, Canada by 13 percent, and Western Europe by 8 percent; the rest of the world declined by 11 percent.

As for revenue growth share, IBM was the only vendor that saw server revenues increase--by five percent compared to last year's second quarter--and Dell's revenues remained flat. In terms of revenue market share, IBM had 26 percent of sales, Sun Microsystems had 16.5 percent, Compaq had 16.1 percent, Hewlett-Packard had 14 percent, and Dell had 7 percent. IDC reckons that 1.04 million servers were shipped in the second quarter, and that Dell increased its shipments by 31 percent and IBM increased its shipments by 6 percent compared to last year's Q2. When all the numbers were in, Compaq got 25 percent of worldwide server shipments in Q2, Dell got 17 percent, IBM got 14 percent, HP got 10 percent, and Sun got 7 percent. Had HP and Compaq already been merged and running at the same sales rates, the combined HP-Compaq would have had 30.1 percent of revenues (about $3.65 billion) and 35 percent of shipments (about 365,000 machines). IBM, by contrast, had about $3.15 billion in server sales in Q2 2001 with 146,000 machines shipped.

Those numbers, more than anything else, explain by HP CEO and chairman Carly Fiorina proposed an acquisition of Compaq and why Compaq CEO and chairman Michael Capellas, who will be demoted to president of HP after the acquisition is completed, accepted the offer. In a server world dominated by Intel chips, the vendor who sells the most Intel chips will have the lowest component costs, the highest margins, and the most wiggle room to engage in a price war. This is why Dell, IBM, and Sun had better pray that HP and Compaq totally bungle this deal and that they leave their installed bases exposed as they complete the merger.

Keep the Enterprise Foundation Strong. * Automated, Real-Time, Change-Based Replication of Data and Objects * Rapid Manual Switchover to Backup to Accommodate Maintenance * Rapid Automated Failover to Backup in Disaster Scenarios * 24 x 7 System Access You absolutely depend on data and applications. They are the foundation of your enterprise. If your systems go down, so does your business. Power failures, lightning strikes, fires, floods and earthquakes are rare, but they do happen. What's not rare is maintenance. You must regularly reorganize your databases and upgrade your core applications, systems software and/or hardware. You need to backup your databases nightly. MIMIX® automatically and transparently maintains real-time replicas of your IBM® AS/400® and iSeries 400 data and objects. It lets you quickly switch to a backup system to accommodate maintenance. Or, it can automatically failover to a backup during an unexpected primary system outage. Click on the hotlink below for an informative MIMIX eBrochure! http://www.lakeviewtech.com/solution/highavailability/index.asp

by Joe Hertvik

Software management vendor SoftLanding is distributing a set of open source green screen tools for managing IBM's WebSphere Application Server; it is doing so through the IGNITe/400 e-business user group. The tools, which are distributed by IGNITe/400 as part of the WebSphere Tools Open Source Project and which do not yet have a catchy name, allow administrators to perform WebSphere commands from a 5250 interface rather than using the graphical WebSphere Administrative Console. In the first release of the tools, which can be downloaded from the IGNITe/400 Web site, SoftLanding and IGNITe/400 are providing the following commands:

* Start Application Server (STRAPPSVR) and End Application Server (ENDAPPSVR) commands, which start and stop an application server within WebSphere (with the application servers corresponding to the jobs you see in the QEJBSBS subsystem)

* Export Application Server Configuration (EXPAPPSVR) command, which exports a WebSphere configuration to an XML document for backup or export to another server

* Work with Web Application (WRKWEBAPP) command, which works with an individual Web application, rather than the WebSphere application server that it runs on

* A WebSphere Tools menu (GO WSTOOLS), that provides an access point for the other commands as well as options to start and stop WebSphere itself

The obvious advantage to using these 5250 command tools is that you can forego the use of the interactive WebSphere Administrative Console for these functions (the Administrative console is usually installed on a Windows NT or Windows 2000 server in an OS/400 shop) and perform the functions directly on a green screen or inside a batch job. The batch job option is particularly useful because it lets you automate things such as backup routines, where you may want to shut down the server, perform the backup, and then restart the server.

SoftLanding and IGNITe/400 caution that they are not trying to build a complete green screen replacement for the WebSphere Administration console. Their tools are best thought of as a supplement to IBM's tools. Nonetheless, SoftLanding and Ignite/400 say they expect to build on the functions they are offering with the first release of the open source WebSphere tools, and add that they will happily withdraw the tools from distribution if IBM provides corresponding new commands for WebSphere. (IBM, this is a hint.)

The tools have been tested with WebSphere V3.5.3, but the companies say that they theoretically should work with any WebSphere V3.5.X release and possibly WebSphere V4.0. However, since WebSphere V4.0 isn't going to be generally available for iSeries until October 19 (see the August 20 issue of The Four Hundred for more on WebSphere V4.0), it's too early to say whether the tools will be compatible with WebSphere V4.0.

To use the tools, you need to have the OS/400 QSHELL interpreter installed along with the following PTFs:

* PTF SF64740, if you're running OS/400 V4R4

* PTF SF64741, if you're running OS/400 V4R5

The WebSphere green screen tools are provided as open source programs under the GNU General Public License (GPL) to encourage the OS/400 community to make contributions. If you download the tools and provide them to another company or individual, you are required to abide by the licensing terms of the GPL. SoftLanding developed the tools, and IGNITe/400 is hosting the open source project and providing the resources necessary to manage the project.

The WebSphere tools run through another WebSphere Tools command, XMLCONFIG, which exposes all the actions you can perform with IBM's XML Configuration Tool (XMLConfig). XMLConfig is a program that can run WebSphere management commands based on the contents of an XML document. SoftLanding's XMLCONFIG (don't you wish they would have named this command something different to avoid confusion with IBM's XMLConfig tool) front-ends XMLConfig by parsing the desired action as command parameters into a QSHELL command string, and then executes the resulting QSHELL script. As such, XMLCONFIG forms the heart of the open source WebSphere Tools, and each tool uses it to build the XMLConfig commands to manage WebSphere. IGNITe/400 tells me that this is the same strategy IBM uses in its other WebSphere-enabled products, such as iSeries Access for Web and WebSphere Host Publisher.