Market Roundup

June 6, 2003




IBM Announces Storage Virtualization Solutions

HP To Try the Thin Client

HP/Intel and IBM Announce Telecom Server Solutions

Palms Joined for the Future Seeks to Create .NET Vision of Business Services


IBM Announces Storage Virtualization Solutions

By Charles King

IBM has announced general availability of the TotalStorage SAN Volume Controller and SAN Integration Server, which the company described as its first storage virtualization products. According to IBM, the new products were designed to provide a common platform and centralized point of control for volume management and storage solutions, as well as to enhance storage utilization and administrator productivity. The SAN Volume Controller, with a base price of $60,000, integrates virtualization software with customized IBM eServer xSeries servers running an operating environment based on the Linux 2.4 kernel, and provides near-linear performance scalability for up to two petabytes of pooled storage. The SAN Integration Server, with a base price of $140,000, is a pre-configured complete SAN solution that leverages the SAN Volume Controller, Fibre Channel switches, and IBM FAStT Storage. The products initially support IBM storage servers and are planned to expand to non-IBM storage systems later in 2003. Customers who purchase the initial systems will be entitled to simple, non-disruptive upgrades as they become available. IBM stated that the Virtualization Family of products would be initiated into the company’s TotalStorage Proven Program, which has produced more than 300 storage solutions pre-tested for interoperability. Additionally, the company indicated that nineteen more ISVs, including SAP and Informatica, had joined the program. The SAN Volume Controller and SAN Integration Server will be available on July 25, 2003.

Storage virtualization is a subject IBM and other storage vendors have been talking up for some time, so it is worth looking beyond the technical details of this announcement to see what lies beneath. In plain English, IBM’s SAN Volume Controller is an in-band device that provides a centralized means of managing and utilizing block volumes of pooled storage data. The company’s SAN Integration Server leverages that same technology with miscellaneous Fibre Channel switches and IBM FAStT storage arrays into an off the shelf SAN-in-a-box. Good enough, but what does it mean? First, it allows IBM bragging rights by finally delivering on some of the storage virtualization vision it has been discussing for years and provides specific points of comparison/competition for similar SAN-based products from vendors including EMC, HP, Sun, and others. There are some fundamental in-band/out of band philosophical technical issues involved, but IBM believes that pairing xSeries servers in the SAN Volume Controller should satisfy most customers’ reliability/back-up concerns, and that basing scalability on adding I/O groups should allow the new products to outrun many individual storage systems. While some might quibble about the solutions limited (i.e., IBM-only) immediate use, we would point out that interoperability may be a nice concept but is one that takes a good deal of sweat to make real. IBM’s initiation of the new products in its TotalStorage Proven Program should aid the eventual delivery of future versions that are dependably interoperable with other platforms.

All things considered, how will IBM’s new products impact the marketplace? Certainly, the SAN Volume Controller should resonate among existing IBM storage customers, and the SAN Integration Server (which uses IBM’s affordable FAStT disks) appears to provide attractive price/performance for an off-the-shelf solution. However, the products’ longer term success will depend on how effectively the company delivers on its multi-platform support promises. While IBM’s Storage Tank initiative was curiously absent from this announcement, these first two products stand as directional indicators for the company’s larger storage virtualization strategy.

HP To Try the Thin Client

By Jim Balderston

HP announced this week that it plans to introduce a product line that would use the company’s blade servers as the processing power behind a line of thin-client desktops. Although no formal announcement was made, the company said that the user terminals would have a small amount of local memory and would access their applications and CPU from a central location. The user would access personal data – and the processing power to manipulate it – through a password login. Under this scenario, a user would get access to an individual blade, although not the same one every time. Computing resources would be managed by HP OpenView software. HP said this arrangement could decrease the cost of enterprise ownership of PCs by as much as 45%.

The idea that HP wants to attack spiraling enterprise IT costs comes as little surprise. Gone are the days to demand for the bleeding edge technologies that will transform the enterprise; instead more and more customers are looking for something that is reliable, manageable, and cost effective. This trend is the impetus for the managed computing environments that various vendors call autonomic or utility computing. HP apparently is making a gambit to tackle the complexity/cost issues with a centralized, thin client offering that to date has not found much traction in the market. Just this week, Oracle shut down its Network Computer division, despite much ballyhoo and backing from Oracle CEO Larry Ellison.

We are not ready to say that the failure of the NC seals HP’s fate for their thin client offering. In fact, the very circumstances and market conditions that HP cites as reasons for this decision may well be enough to finally give the thin client and blade a real shot at traction. But before we go too far down that road, we have to ask: what are the price points for these units and the blades that make them go? Cost savings up front, we believe, will have to be part of the sales package. A promise of cost reduction over the long haul may be too much for gun-shy enterprise IT customers who have been burned so many times in the past. That said, we also have to wonder how this move will play within HP’s utility computing framework. In one way, the company could address the issues of management of complex computing environments by making at least one part of the environment – the desktop – much simpler. Yet for many enterprises, the solution to their complex computing management needs will not include a thin client solution, or even a sort of thin client (remember the NetPC?). While it remains too early – with too few details -- to make a full-blown prognostication on the viability of HP’s offering, we believe that pricing and immediate cost savings will be key to any momentum that this latest incarnation of the thin client will generate.

HP/Intel and IBM Announce Telecom Server Solutions

By Charles King

Intel and HP have announced that they are developing standards-based solutions for telecommunications applications featuring Intel Itanium processors, embedded Intel Architecture processors and other silicon. Intel also named HP as the first premiere member of the Intel Communications Alliance (ICA), a group aimed at developing standards-based modular platform telecom solutions. In a separate announcement, HP announced availability of the HP cc3310, a NEBS Level 3-certified Intel Xeon-based telecommunications solution built on Intel’s TIGPR2U carrier grade server platform. The new server runs Red Hat Enterprise Linux AS 2.1, features single or dual 2.4GHz Xeon processors, 12GB memory and six PCI slots, and supports up to 292GB of disk storage. In an unrelated event, IBM announced new hardware, software, and service offerings for the telecom industry. The company plans to deliver a new BladeCenter for Telecommunications that utilizes the company’s existing BladeCenter technologies along with Intel-based processors and hardened operating systems such as carrier grade Linux for telecom applications. IBM is also developing pre-integrated Linux-based solutions for core network applications, as well as WebSphere Business Integration for Telecommunications, which provides software solutions designed to accelerate new service introduction. Finally, IBM announced new telecom-specific initiatives and solutions including Telecommunications Data Warehouse (TDW), a business intelligence solution; and IBM Content Management, which leverages DB2 Content Manager and WebSphere Portal to increase call center productivity and efficiency.

First off, these and other Intel-based telecom solutions have a singular target in Sun Microsystems, which has largely ruled this sector for years. But while HP and IBM may both be leveraging Linux to unseat Sun (while throwing the occasional elbow at each other), they are approaching this shared goal from markedly different directions. HP is depending on its increasingly critical partner Intel to help deliver the goods. We say “increasingly critical” not because the two companies’ professional relationship has shifted but because as HP moves further along in becoming an Intel-centric vendor, the two companies’ synergies and dependencies are becoming markedly clearer. Since HP is stepping away from silicon design and development, new initiatives by Intel constitute new opportunities for HP. At the same time, HP and Intel each stumble whenever the other trips. Itanium’s rocky development road has likely impacted the sagging fortunes of HP’s enterprise server group. This is doubly bad news for Intel, since HP constitutes Itanium’s largest and most influential backer. The two companies’ vigorous push into the telecom space by dint of Intel’s telecom-related efforts makes perfect sense. After years of bad times, carriers and service providers are looking for new ways to maximize performance and minimize costs, and shifting away from Sun’s proprietary platform to an Intel/Linux-based solution may fill the bill. Intel’s pre-configured telecom platform should aid HP’s development efforts and lower costs enough to make everyone (except Sun) happy.

So what about IBM? First off, the company offerings are notably different from HP’s in two ways. First, IBM’s leverage of its BladeCenter architecture (which packs nearly twice as many blades into a rack as HP’s blade solution) provides it a significant leg up in server density. Additionally, the myriad software and service offerings (largely based on IBM’s middleware solutions) accompanying the BladeCenter for Telecommunications offer a strategic vision starkly different than HP, which has surrendered the middleware battleground to its partners. A question worth asking here is just how rosy the future of enterprise hardware can be unless it is buttressed with solid middleware development. The answer for IBM obviously comes from within. It remains to be seen if HP can do without.

Palms Joined for the Future

By Jim Balderston

Palm Computing announced this week it is acquiring Handspring, a competitor in the handheld space that was started by two of the original founders of Palm. Palm will pay approximately $190 million to purchase its rival, which licensed the Palm operating system for its line of handheld devices. Handspring has recently focused on wireless devices that serve as both organizer and cell phone. As part of the announcement, Palm said it would be spinning off PalmSource, its software division that licenses the Palm OS to various manufacturers, including Sony.

While many of the PDA-specific sites were all a titter about this announcement, we can’t help but to see this for what it simply is: industry consolidation. And as history has shown, most such consolidations happen after markets constrict and competition becomes too fierce to sustain all of the players in the market ecosystem. So is the case here, with sales of handheld devices peaking in 2001 and sliding thereafter.

Will this reunion of the Palm founders make a difference in the market? We are not ready to go that far. One can’t help but wonder to what degree the handheld market has been an inside baseball kind of play, with everyone in Silicon Valley, Austin, and the other outposts of tech mania all hauling around their various personal organizers and communications devices like Sherpas assaulting Everest. In short, these devices were status symbols and totems of the technology revolution that everyone was thrilled to be a part of. Palm Pilots and their market counterparts had to be the next big thing, since “everyone” was carrying one. Now that the universe of “everyone” has been drastically reduced by the information technology slump, we suspect that sales of these devices will not be taking off any time soon. Just as new reports find that people are overwhelmed with email, IM, and the overall instant nature of communications, we can’t help but wonder if non-alpha geeks are actually enjoying not having every waking minute of their lives entered into a device that persistently reminds them that they have to be at a meeting, a dentist appointment, or a soccer practice. While we believe that there is a steady market for devices that help people organize their time and stay in contact with the world, we believe the rate of growth for such a market will be far from the heady days of the end of the 20th century, and that more modest and less dramatic market penetration will occur over time. Will that be enough for the new Palm to reverse its fortunes? Stay tuned. Seeks to Create .NET Vision of Business Services

By Myles Suer announced the launch of a new Web-delivered service aimed at application developers and their customers. The new service, called Sforce, makes publicly available the client/service application development utility. The company announced strategic alliances with BEA Systems, Microsoft, and Sun; and indicated that is plans to work with Borland to enable enterprise developers to use their current development tools, combined with Web services to rapidly build and deliver business applications using a software-as-service model. The Sforce concept seeks to facilitate to development and delivery of Web Service by ISVs through, effectively renting the environment from which Web Services or Objects are developed and then deployed to customers. For this, charges $50/end user/month for deployment, plus $1/MB/month for the first three users with a 10MB free incentive for the first year. The Developer Edition is available online for free and deployment fees are waived for Enterprise Edition customers.

As we look at this announcement, the first question that comes to mind is: What does this do for the end user? To a certain extent, this realizes much of Microsoft’s .NET vision of Web-delivered services. We believe this has interesting implications for users of core business as we see it potentially creating a role for systems integrators in Web-based or ASP delivered software. In this role, integrators could tailor’s core product to users’ needs while enabling to add greater value where the customer has many elements in place; for example, their own accounting system. Over time, this approach proffers the ability to pull information out of a greater variety of services. However, in the near term we believe the ability to get inside and make use of the capability of’s existing software will be the driving consideration. For the overall market, this announcement potentially gives a means by which to go after vendors such as NetLedger who has been building out all of its product suites. But will this signal the beginning of mix mode software, where some resides on the local server and other is leased? Only time will tell. Nevertheless, for business process outsourcers, this announcement may prove very interesting.


The Sageza Group, Inc.

836 W El Camino Real

Mountain View, CA 94040-2512

650·390·0700     fax 650·649·2302

London +44 (0) 20·7900·2819

Munich +49 (0) 89·4201·7144


Copyright © 2003 The Sageza Group, Inc.

May not be duplicated or retransmitted without written permission