If you had to sum up the sales pitch that Cisco Systems is polishing up as it prepares to deliver its "California" Unified Computer System, it would look something like this: California will save you money, even if we charge a premium for server capacity. Cisco, of course, has not admitted that it will in fact charge a premium …
So, let me get this straight. Cisco, who operates in a high-margin networking hardware business wants to muscle in on IBM/HP/Dell/Sun who operate in a cut-throat low-margin world (at least for HW) with an integration play?
Got to give them credit for trying to monetize their customer base by hurting the other players, but in the process there's a good chance they'll expose their own cozy market economics to increased scrutiny and increased competition from those same players.
I can't see this at all. They are risking a very valuable networking business to break into the Blade market and upset the big boys? What cretin came up with that idea? Does the Cisco board really think IBM and HP will gladly give up market share and not seek some sort of revenge?
With the management server gone, does that mean admins are no longer needed?
Personally, it looks to me like Cisco is trying to shoot themselves in the foot for the folks that won't fall for their nonsense. I can't see why anyone will jump in and buy a Cisco blade server otherwise. And if they allow you to buy their hw, but with another vendors server, they're also shooting themselves in the foot. Am I the only one confused at their logic/timing?
The comparisons in this article definitely aren't against other Blade enclosures
Lets look at the HP c-7000 enclusre, which takes 16 half-height blades (460c) which have two CPUs. To allow 320 servers, it would take 20 c-7000 chassis'
Each of the c-7000 chassis are 10 RU, so you can fit 3 per rack with some network equipment and patch panels, or 4 per rack with extra cooling. Lets take 3 per rack.
That's only 6.6 racks, let's call it 7.
Each of the c-7000 blade enclosures feature Network ports, where you can add Cisco 3020 ethernet switches, Brocade FC switches or passthrough modules.
It seems to me that Cisco have created a great product definitely worthy of further investigating, however their lies, damn lies and statistics are misleading at best.. It would be unwise to read these Cisco statistics and swallow it whole.
Agree with Dave although unsure what you really mean by Econimics?
Suppose mimiced economics?
Cisco although entering a mature hardware commodity server market where price/performance ersosion is the norm and with fierce competition especially in the X86 standard server space, appear to be taking a slightly different approach . . .
The initial hardware pricing doesn't seem great value, they are concentrating on the integration savings and power savings rather than promoting cheap servers like the other players, and they may have some possible success here.
The pricing may limit the overall market appeal but they seem bound to get a degree of low hanging fruit from their integrated value proposition, and their approach appears to have reduced complexity savings and maybe performance gains by coupling the server blades and switches ever more closely, tightly and simply.
The look set for some rich pickings along the way.
The simple integrated approach could speed market take-up and yield faster implementation and installation times as more of the configuration and set up is preconfigured in the box and rack and it may prove to be a great way to rapidly generate private and public cloud configurations and make it far easier to design and deploy so called transclouds.
Because of the integrated and standardised approach it could be very readily scaled out to adapt to changing business needs.
A standardised and integrated network, compute, store building block could be safer and simpler and very much faster and secure.
It could make it very much easier to deploy the VM parts.
It would be interesting to know more about the performance per volume space, given that they appear to have a good handle on reducing power consumption.
Over a three year cycle energy cost is still a massively high component so there is a long way to go in the industry to improve this.
It would be good to know more about reliability/availability and this will only really come from monitoring large scale field deployments over time.
Dave is right, by muscling in on the server turf they draw the spotlight onto their existing cozy margins and the exisiting server centric vendors will accelerate a competitive response in return.
This is good news for us customers though!
Cisco may have success with modular flexible storage in this approach as well and it may well be that they are approaching things from a useful switch centric direction.
In architecting datacenters and reliable IT infrastuctures, although application deployment and software platform are key and are so often seen as start points, looking at communications bandwidth and the switch first and then mapping out is an interesting approach as maybe is better balance between comms compute and store with this approach.
The networking and management software wrap including VM is a great source of value here and you can imagine that this integrated approach could help with faster and thinner deployment and very much better system resource allocation.
There are possible security improvement considersations also with integrated server storage switches.
Its going to be highly intesting to see what happens with early deployments following general availability in the June/July 2009 time frame.
IT supplliers all seem focussed on getting to the same end point so there is going to be much fun and perhaps more vendor consolidations ahead.
The integrated California approach could help Cisco become harder, better, faster, stronger triggering competitor improvement.
Wrong. Cisco is fighting Sun
Sun is working in the integration of it's networking gear, that is kind of "standard" because it connects to standard servers but is awesome because it offers features and performance only possible with dedicated equipment until now, with standard servers, not only from Sun but from the other players.
When you can buy a "router" with several 10+GbE NICs that can be virtualized, talk with the other NICs in the server directly, multiplexed "in hardware", etc, for $2,000 USD, the high cost, dedicated, sole-purpose routing equipment will be obsolete. That time is now and Cisco has to expand it's business.
In a few years Cisco gear will be running in a AMD or Intel CPU, with NICs from Sun at first and then they will make their own to plug to standard servers.
I can't wait! :-)
For Beta Interindependent NeuRealIT World Orders *.
The SMARTer Servers,[Intelligence Enabled and Enabling] do not require Integration within a larger and/or Proprietary Legacy System/Organisation/Business Model . IT will remain Independent as a Spontaneous Core MetaData Supplier and Virtual Machinery Sharing InterNetionally Novel Information/Structured Intelligence, which Legacy Dependent and Driven Systems will Require of themselves to React to ....and to Support, in Order to Survive in an Environment which Total Information Awareness and their Servers can choose to Render, Destructively Chaotic.
Econimics 101 ... By Dave Posted Tuesday 24th March 2009 21:55 GMT = Nail on the Head.
That's what I was thinking too, Dave.
*New World Orders a la BIN?
If you like to gamble I'll tell you I'm the man
Ace idea Cisco...
All you have to do is to persuade big business that you, a network hardware manufacturer, are in this for the long haul and that your expensive modular system you've just pulled out from up your sleave is actually going to be still marketed by you next week/year.
Having been stung by such ideas before, and not even during a recession, I'm going to fold on this one.
What next from Cisco... games consoles? mobile phones? kitchen appliances?
Paris, because no doubt she will have a unified whatjimmywhojimflip too next week
Now, if they can ...
... get the network card to autonegotiate with the switch it will be worthwhile :-)
Now, if they can ... just do the next tricky dicky bit.
Now, if they can ... ... get the network card to autonegotiate with the switch it will be worthwhile .... and credit anonymous team driver accounts with account specific code-sharing wishes, aka Covert Stealth Support of Virtual Drive Direction ..... would be AIMagic, and you would not believe what that can do for them, with them in Proxy Control.
Quite who them is, in Control, is always going to be the Eternal Question, to keep everything Moving On nicely. :-)