HP ANNOUNCED that it will buy 3Com for $2.7 billion the other day, and with one swipe of pen on checkbook validated Cisco’s strategy. HP also showed that it won’t be able to compete in the converged network arena for years to come.
The 3Com purchase wasn’t one made willingly or strategically by HP. It was simply a matter of its current Procurve and related servers being blown out of the water by Cisco. HP is flailing about tying to come up with a coherent strategy, and likely also with the question of build versus buy.
If you recall, Cisco jumped into the server arena earlier this year, and the server OEMs suddenly got very frightened. The first Cisco servers were a bunch of lackluster blades that promised convergence. We showed you the ‘Ventura’ blade a few weeks ago, and that is one of the three servers that are the backbone, or in this case the bits that dangle off the backbone, for Converged Networking (CN). However, the Ventura blade is a game changer for the server industry.
If you ask Cisco, it will say the backbone is and always will be its switches. Servers at Cisco are compute devices that hang off the fabric and are, well, mostly commodity items. If you think about how many companies you can buy a 1U 48 port 10GigE switch from versus how many companies you can buy a 1U 2S Nehalem server from, it is hard to say Cisco is wrong about this.
Calling the core of their businesses commodity parts is enough to scare IBM, Dell and HP to death, and Cisco did that with Converged Networking. IBM, Dell, and HP were each sent scurrying to find a networking dance partner that had the breadth of products that Cisco does, but, ummm, there aren’t any. Time for Plan B.
So, HP bought 3Com, and will undoubtedly claim that it now has ‘converged’ products just like Cisco does. Converged Networking is not just about plugging blades in, though. It goes a lot deeper than that, especially if you look at the networked convergence board (UCS) and fabric controller (FC). You can’t just buy those off the shelf or code a few lines of scripts to make them happen.
HP’s 3Com purchase is just the first step, but a decent one at that, in getting to CN. It now has many of the building blocks that make up a CN system, but it does not have a CN system yet. That will take years to make and integrate. 3Com is a $2.7 billion baby step, and writing the check is the easy part.
A lot of 3Com’s portfolio overlaps the HP Procurve line, so integrating the networking greybeard will not be an easy task. To top it off, HP will have to integrate the two companies before it can integrate the product lines. Islands of R&D are not the way to convergence. It is a whole, not a list of bits on an order form.
In the end, HP bought 3Com because it was likely the best fit of the non-Cisco networking set. HP had to buy someone, so it did. Cisco is in another league at this point, and is moving from the difficult bits, networking, into the easy part, commodity servers. Now HP has to do it the harder way, but at least it is moving in the right direction.
However, HP is doing that in a very different server marketplace now.S|A
Latest posts by Charlie Demerjian (see all)
- More on Intel’s 10nm process problems - Sep 17, 2018
- Intel puts out another 14nm 2020 server platform - Sep 11, 2018
- Why Can’t Intel Supply Enough 14nm Xeons? - Sep 10, 2018
- Intel can’t supply 14nm Xeons, HPE directly recommends AMD Epyc - Sep 7, 2018
- AMD reintroduces the Athlon name with two CPUs - Sep 6, 2018