blade servers

HP’s 3Com acquisition: Vendor market activity spurred the move

HP has made a somewhat-surprise move and acquired 3Com, a venerable but now fallen giant in the networking field, and the move is very likely an explicit a changing of the guard with respect to enterprise networking.

We’ve noted that the control of network decisions is passing to data center IT personnel, and thus that network sales engagement means having IT products in the portfolio. HP obviously has them and is just as obviously now planning to expand its network portfolio.

But we think Brocade’s purchase of Foundry and Cisco’s decision to enter the server space is the catalyst for the move. 3Com has a new data center switch product that could significantly augment the HP portfolio, and also a relationship with Huawei and a position in Asia.

Tapping the China market is likely an early goal for the deal. The HP move could put pressure on IBM, which has been relying on partnership relationships in networking rather than fielding its own product. It may also put some of the smaller vendors in the LAN space into play, including Extreme, and possibly Brocade or Juniper.

VMware virtual infrastructure sparks vendor positioning flurry

The next version of VMware’s Virtual Infrastructure, named vSphere 4.0, was announced with a flurry of high-level endorsements from Cisco, Dell, Intel and EMC.

The high-level song and dance can be attributed to the importance of the so-called “data center virtualization” space, which is the host side of cloud computing. IBM and HP, who are both miffed at Cisco’s blade aspirations and determined to carve out their own niche in the space rather than share one, won’t be a part of the big show.

This space is becoming the tail wagging both the IT and network dogs — a relatively small market segment with great strategic importance. We believe that this space was a major reason for the Oracle/Sun deal, for example; Oracle gets a cloud approach and storage, both critical in this new hot area. HP’s new Matrix announcement is aimed in this space, too. IBM will have to respond to the product flood arising from the face-off between HP and Cisco over the critical binding between networking and the data center. That could induce IBM to move closer to Brocade or Juniper, and the question of which way IBM jumps could be critical for both companies.

IBM to acquire Sun? Benefits for the cloud

IBM is said to be in talks with Sun on acquiring the company, which has been troubled since the NASDAQ bubble burst. The move would give IBM a commanding position in the server/hosting space for the Internet and a strong position in telecom. A move would very likely create considerable angst for Cisco, which now faces a broader blade server collision with the industry’s most formidable tech giant.

We believe the move would be good for both companies. We also believe it is an indicator that IBM sees the transformation taking place in what’s called “cloud computing,” the creation of a new relationship between network and IT resources for both network operators and enterprises. Sun’s credentials are strongest in the network space, but it also has considerable cloud talent and its own cloud computing offering.

Cisco’s generalized data center solution — risky business

Cisco made its long-awaited “California” announcement, but the event didn’t fully address the critical questions of just how the new products will be positioned. At one level, Cisco announced itself as a general-purpose blade-center-based IT competitor to IBM, HP and other computer vendors. At another level, Cisco appeared to be putting forward positioning that makes a distinction between a new Unified Computing System and a classic data center.

UCS is heavily linked to network virtualization, strongly linked to storage networking and somewhat linked to cloud computing. Most will recognize much of the UCS positioning from prior Cisco presentations on network support for virtualization, and Cisco trotted out a bunch of partners from Accenture to EMC to try to defuse concern that it might be trying to be a one-stop data center solution provider.

Our view of California from the first has been that anything that presented a generalized data center solution would be incredibly ambitious and risky for Cisco, and we believe that Cisco’s efforts to avoid the risks have not been sufficient, at least in terms of positioning. Counter-pressure from a tightening of the IBM/Juniper relationship seems inevitable if Cisco really goes after the data center market. Juniper’s Project Stratus fabric approach to switching appears more directly responsive to the data center opportunity and less likely to create IT vendor angst.

But will Cisco really do what the product material suggests? The early sales targets and beta tests, according to our sources, are primarily like service delivery platform (SDP) applications, content servers and other special missions.

Cisco’s blade aspirations short on software?

Cisco’s new blade offering, Project California, is likely to be strong on blade and short on software other than UC. This could be a shortcoming that will make Cisco more vulnerable to competition than even some incumbent players might be. Cisco’s project has leaked out for a year or more, giving people plenty of time to position against it.

Dell recently commented extensively on its own view, which might be described as a “blade ecosystem” approach that would involve not only hardware but software and management tools. The implication is that Cisco won’t go there, but we find that very hard to believe given Cisco’s recent software push and the fact that it has management software tools already. In addition, rumors Cisco might even buy VMware have continued to swirl. We think that Cisco’s launch of a blade server would likely come when market conditions appear to be easing, but it may make some M&A moves much earlier.