The professional journalists will hash out the details on the Cisco’s long-awaited announcement today. Some striking aspects:
- Memory – as in DRAM – a key point. Chambers and Intel remarked on this several times. The next gen Xeon’s will address *lots more* DRAM – hopefully more efficiently than the wretched FB-DIMM – to enable many more VM’s per server. “More memory than any other blade” until other vendors announce blades using the same Intel processors.
- Cloud-grade. As in “better than enterprise-grade.” Commodity-based cloud kit setting the standard for name-brand enterprise kit? That goes right to the heart of this announcement’s business problem.
- Cisco’s dream: leading company in IT. Chambers said it and he meant revenue, not technology. Key: make the IT market pie bigger.
- Make the pie bigger. 2 revenue sources: consumer and travel. “Consumers won’t know where their movie is stored and they won’t care” – uh-huh – the audio-only webcast kept cutting out on my “broadband” connection. Travel reduction through video conferencing – sure, that makes sense. But how much compute does that require?
MIA
The announcement was more interesting for who wasn’t there: IBM, HP, Sun, Dell, Google, Amazon and Yahoo. The former 4 because they represent traditional competition and the latter 3 the cloud.
If IBM, HP and Sun aren’t meeting today to plot a radical, Cisco margin destroying open-source router & low-cost switch counterattack – like Seagate, HP and IBM performed on Quantum’s DLT – they’re idiots.
Vision
“Vision” means “ain’t happening now, baby.” So Muglia’s comment about Microsoft’s vision of hundreds of virtual servers working together means that isn’t happening now – mostly for storage reasons. VMware’s storage infrastructure is pathetic; Hyper-V’s is worse; and the Unified Computing System won’t unify much if the storage issues aren’t addressed.
The StorageMojo take
Cisco’s partner-friendly strategy enabled it to build a strong position in data center networking and routing. But now it has declared war on the server businesses of HP and IBM, so “coopetition” is the order of the day: all smiles on even-numbered days and brass-knuckles on the odd.
A gangster once told fictional private eye Philip Marlowe “I could make your business my business” and Marlowe replied “You wouldn’t want to – the pay’s too small.” That’s the server business in a nutshell.
The real money in servers is storage. While Cisco may enlarge the pie it will shrink its margins. The “private cloud” business – and I was one of the first to discuss it – will prove a fruitless distraction.
Cisco has endorsed commodity-based cloud infrastructures without offering anything competitive. At best they’ve started a race to the bottom in blade servers and networks against a company, HP, who’s just finished mopping the floor with Dell – who used to be known as a tough, low-cost competitor.
Good luck with that.
Courteous comments welcome, of course.
good news for netlogic it would seem
The site is looking great!
My input on the redesign would be to use a 2 column format. Perhaps put the ad above or below the URL bar on the right.
The new design is much more modern and streamlined. The reduction in the number of (flashing) ads is nice. Although, you need to add back that iconic head shot of yours on the right. 🙂
Sun is planning something:
http://blog.scottlowe.org/2009/03/09/is-sun-preparing-to-take-on-cisco/
http://blog.scottlowe.org/2009/03/12/this-pretty-much-answers-that-question/
“the Unified Computing System won’t unify much if the storage issues aren’t addressed”
what are the major issue Mojo thinks are issues which are posing immediate roadblocks ?
~subbukl
Benr – got the headshot back up. Does it make my head look big?
I’m still mulling the whole ad thing. Could be happier – will continue pursuit.
Sub, VMFS can run a maximum of 32 ESX nodes. Running a group of shared-storage clones? Make that 8.
Cisco announces bong-sized clouds. Frat boys cheer!
Robin
“VMware’s storage infrastructure is pathetic…” — I second that.
Robin,
I’ve read many articles on the various aspects of cloud, including storage, and still wonder if its hype or factual direction. Some aspects are compelling, while others are questionable; like keeping your personal/business critical data ‘somewhere’ out there.
BB.
Bobby,
The “cloud” does not necessarily mean public cloud infrastructure. In clouds, there exist at least these two axes:
Private cloud vs public cloud
Infrastructure style (e.g., EC2) versus computing service style (e.g., MapReduce)
Anyway, to me “cloud” is Virtualization 2.0. Amusingly, most everything in the cloud is currently vapor. For a really innovative way of looking at data centers, however, if you have not seen 3TERA AppLogic, you really need to take a look. I think that applications like theirs offer the long term vision of where this whole virtualization thing is going.
Joe.
a Few Points:
From a Networking perspective, do you really think that a “low cost” Sun/HP/IBM Switch can compete with Linksys? Hard to go lower than Cisco at some level since they already own the bottom 🙂 At the high end routing / DWDM space you have to deal with a whole host of players, some you’ve likely never heard of like Infinera. Google already makes their own ethernet switches, so I’m not really seeing a downside for Cisco networking wise.
Don’t you think that Cisco has EMC or NetApp (or both) as an acquisition target if the Three Stoogies (IBM, Sun, HP) actually get their act together? What would they do in response? Buy Brocade, Juniper & Citrix? LOL.
In any event, Cisco is targeting the low hanging fruit of windows consolidation with a fancy Cobalt like server in a can that plays nice with the data engineering teams while gracefully handling a number of oblique issues related to monitoring and security (VN-link/ nexus v1000) which will benefit VMware by removing some of the few remaining barriers to adoption related to compliance and security (extending IDS’s to the vm networking environment). This of course is a very unsexy but practical solution to existing business problems. Note that these aren’t “Storage” issues; which are really only relevant to any one but big web players.
~Max
“VMware’s storage infrastructure is pathetic…â€
Yep. And the stereo in my Porsche sucks.
Underestimating the ability of EMC and Cisco to put together end to end magic seems to be a favorite sport of folks who rally behind companies with R&D budgets the size of my mortgage. VMWare only recently learned that “storage” was not “the place where we keep our mops”.
Just like my stereo problem, that can easily be fixed.
Dave
How much of HP’s success should really be assigned to COMPAQ? Nothing revolutionary about a bunch of Wintel server sales.
Seems like battle lines are forming around the virtualization space, with IBM possibly picking up VirtualBOX in the Sun acquisition.
I’ve signed up for a Cisco analyst preso next week so maybe everything will become Perfectly Clear. Or not. I know some of my techie Valley friends think this is pretty cool, so maybe I’ll be revising my opinion – not the first time that’s happened.
Dave, IBM, HP and Sun together spend a heck of a lot more on R&D than EMC does – and since they let their PhDs publish and don’t punish them with non-competes, they have a better shot at top talent. VMware has been hiring storage engineers like crazy for the last 3 years and they still haven’t fixed the problems. Maybe they will next week, but I’ll believe it when I see it.
RC, how much of Compaq’s success should be assigned to DEC? Just asking.
Robin
COMPAQ’s success should be attributed to T.I.
AMD and Intel both owe a nod to DEC though, for all the ex-Alpha engineers that fueled advances in Athlon and various Intel processor advances. (Intel would deny that I suspect.)
And we all benefit from DIX, whether we realize DEC was a key player in ethernet or not. DLT tape? The list goes on.
Cisco has eliminated partner programs with IBM, HP, and other server providers. Now its a small channel team, not focusing on growing revenue with partners, but with its own server. Can they do it on their own?
Brocade CEO Mike Klayko’s take on Cisco’s annc. here:
http://www.youtube.com/brocadevideo
Thanks,
Wilson Craig
Director, PR, AR, Social Media
Brocade
IBM, HP, and SUN? R&D?
Last time I checked they were a bit less focused than EMC with those R&D dollars. I notice that your blog is not called “Printing Mojo”.
I certainly hope that the PhDs dying to publish go to a place where they can bask in intellectual elitism. Like Sun.
Initiailly I though this was Cisco simply jumping into server market. However, now that I have seen the technology and delved down into the components, I have to say. This anything but. The UCS changes the game. They have blended the lines such that the server, the network and the storage work as one cohesive unit rather than components that service each other.
I am a huge fan of HP (Compaq), but they have nothing that comes close and I think they will spend at least two years catching up. It requires a strong networking core compentency which they don’t currently have. I predict they are going to attempt to buy either Foundry or Entarasys in the next year.
The biggest losers are Dell and Brocade. Dell spent a lot of money trying to catch HP and still lost the blade server battle. They won’t have the cash to fight this battle. Brocade bet the farm on their version of unified fabric (FCoE), then EMC and NetApp both went with Cisco (ratified as the standard last month). They were already steadily losing market share in the FC switch market now they are starting off with one foot in the grave on the unified fabric market. Look to see them either be bought by Foundry or Entarasys.