Turning HP and IBM into 'frenemies': Not so clever now, heh?
Cisco's problems run deeper than having $40bn-deep pockets that allowed it to indulge in its expansion into lower-profit consumer businesses that, in theory and very likely in practice, would have helped drive up network bandwidth demands at service providers and inside of corporations.
What seems to have escaped Cisco is that the network traffic was going to go through the roof even if it never expanded into these consumer adjacencies, as Chambers talks about them. And, more importantly, network requirements would have busted through the ceiling even if Cisco did not make a move two years ago into the converged server-networking market with its "California" Unified Computing System. Millions of Apple iPads and a zillion smartphones were going to do the job, along with Netflix, YouTube, Hulu, and other video-streaming services.
All Cisco had to do was wait, and then count the money.
All Cisco seemed to be focused on was growth for the sake of growth and propping up its share price. This is a mistake that young companies make. It nearly killed IBM two decades ago, for instance.
As a server market watcher, I welcomed Cisco's entry in the server space two years ago because of the interesting value proposition that Cisco was offering and because of the engineering work the company had done to make a converged blade server platform. Cisco shook things up, that's for sure. But maybe more things got shaken than Cisco was bargaining for.
Turning Hewlett-Packard and IBM into frenemies instead of allies in the networking arena may not be worth it, over the long haul, if Cisco cannot maintain a sustainable engineering and price/performance advantage with its UCS machines over these two server makers. Particularly with Juniper Networks, Mellanox Technologies, Arista Networks, Extreme Networks, and others that compelling alternatives to Cisco gear.
HP made a decision to get serious about networking ahead of Cisco's jumping into the server racket, so maybe HP started cheating on Cisco first. The poker cards were in the air and the guns were out by the time HP shelled out $2.7bn to buy 3Com in November 2009. IBM ate Blade Network Technologies for $400m last September and is rebadging Juniper and Broadcom gear, a big reversal for Big Blue. The company sold off its networking business to Cisco in 1999 and became one of its biggest champions in the data center.
It is hard to say how much of an impact the lack of pushing by HP and IBM, the two largest server makers in the world, has had on Cisco's top and bottom line. And with upstarts like Arista Networks coming out with faster and denser switches, Cisco has problems here, too.
Cisco's statement today did not address any of this, of course. Wall Street and the Cisco top brass want a scapegoat and it is getting one in Cisco's consumer businesses. We get the feeling that this is far from over, particularly if sales of UCS servers, either directly or through the Acadia Vblock partnership with EMC and VMware, do not pick up.
Cisco has done remarkably well in ramping up its UCS business – as well as Sun Microsystems did with its initial Opteron-based Sun Fire x64 servers back in the middle of the last decade. But a strong start does not make a business sustainable and profitable at a level to which Cisco shareholders have become accustomed. ®