With a New Server, Cisco Pushes "Comm-puting" Strategy

9 thoughts on “With a New Server, Cisco Pushes "Comm-puting" Strategy”

  1. Wow! That is one huge step towards the unification of such a new balanced unit ready for virtualized software and services, I’d love to deploy a data center using these bad boys, the power detailed in the specs looks good, with a 8-core nehalem powerhouse and 384gb of ddr3 goodness. Virtual servers would love this platform. Cannot wait for the official release and details on this unit.

  2. This is all well and good, the world needed better carrier class hardware for compute platforms – lets call them superplatforms. The other half of the equation is the software that is sufficiently general yet comprehensive, such as Openlink’s “superserver” app server, database, id managers, linked data repositories, etc.

    So, we now graduate from comm boxes at enterprise class to servers at enterprise class. It’s still all the services that define the system.

    1. I agree with Alan that services define the system. Cisco revolves around this concept, which is probably why John Chambers, quoted in a ZDNet blog post, commented that he has always been “Technology-agnostic.”. As the Manager of Network Systems Networking and Services Switching for Cisco, I can tell you that Cisco acknowledges the increasing demands of the network, but we also realize that customers need full service and full functionality. For example, we recently improved our technology with a new line of Nexus switches and redesigned the 6500 Catalyst switch to work as a virtualized service node. The Unified Computing announcement shows how we take our existing and new technology, put it to better uses, and provide full service so that our customers can make the most of enterprise network virtualization.

      The move into Unified does not demonstrate a conflict with other companies, but rather reflects the increasing demand from our customers for this kind of technology.

  3. Some more details in our post at Virtualization.com: http://tr.im/ciscofornia

    For example, the part about the switches (8th blade slot for a Nexus) etc is not what we’ve heard from Cisco.
    –> 4 or 8 blades (4 with lots of memory and redundant I/O connections, 8 for higher CPU density but without redundant I/O paths)
    –> 2 fabric extenders
    –> FEX connects to UCS6100 switch (Nexus 5000 with management ports)

  4. Are you kidding me?

    Cisco plans to take everyone back about 25 years to a high-margin, single, all-in-one closed “solution”. How can they be innovating here when they touting open standards and using closed protocols between the blade system and Nexus switch. In short, you can’t use the blade system unless you use the nexus switch. More $$ for greazy Crisco.

    Cisco is trying to own everything in the datacenter…which sounds good to them $$. How many datacenter’s do you know run entirely on one or two vendor’s gear??

    If this is the Empire (or Enterprise) Strike’s Back, shouldn’t be long before Return of the Jedi (underestimated little guy) puts the Emperor (Chambers) back to building overpriced death stars with IOS hacks posted on the Internet and delivered via an R2 unit.

    The idea of network virtualization sounds good, but should have partnered with someone that has blade system experience, instead of spreading yourself too thin.

    Should fizzle out in about the same time as their last blade venture did back in 2005.

    And as for the Sun guy (Phil) …is your company building the blades for them? Looks like Sun OEM gear to me – which I highly doubt it is.

  5. Its the comm-odity revilution, not the comm-unication and not the comm-puting !

    Amazon and Google are building huge data-centers made out of the cheapest computers they can buy. As a monopsony they force lowest margins on suppliers. They don’t pay for expensive RAID ( poor storage companies:) ) and they don’t pay for expensive boxes ( It is all the same intel inside anyway ).
    The secret of their success is software level load balancing, high availability and storage.
    Network vendors have been able to avoid the problem so far because one still needed expensive switches.
    With virtulization you need 4% of the switches you needed before because most of the traffic is inside the server and you get the switch for free.
    This is very bad news for everyone who used to sell expensive equipment in a fancy chassis.

    Cisco is trying to keep the 60% margins it has by starting to sell in the 15% margin commodity server world.
    It still remains to be seen why this server is so unique.
    It is not as if IBM and HP have not been making excellent blade servers that run VMWARE for 10 years now.

    I think the really big story here is Cisco would need to move to a software company model, which is quite a big change.
    Obviously a lot of their intellectual property is in software, but the whole thinking of the company is more “hardware” oriented.

  6. Oh well, whatever the motive or the benefits, things have become a bit boring in the industry with all the economic downturn. So as a techno enthusiast, this is an interesting inflection point. I’m enjoying the marketing dynamics here.

This site uses Akismet to reduce spam. Learn how your comment data is processed.