Slashdot Log In
Cisco Barges Into the Server Market
Posted by
kdawson
on Mon Mar 16, 2009 06:45 PM
from the unsheath-your-blades dept.
from the unsheath-your-blades dept.
mikesd81 was one of several readers to write in about Cisco's announcement of what has been called Project California — a system comprising servers made from 64-bit Intel Nehalem EP Xeon processors, storage, and networking in a single rack, glued together with software from VMWare and BMC. Coverage of this announcement is everywhere. Business Week said: "The new device, dubbed Project California, takes servers into new territory by cramming computer power into the very box that contains storage capacity and the networking tools that are Cisco's specialty. Cisco's approach could help companies use fewer machines — saving money not only on hardware, but also on power and IT staffing — in building data centers. ... Cisco is well-girded to take this step. It has more than $30 billion in cash, more than any other tech company. The company is moving into no fewer than 28 different markets, including digital music in the home and public surveillance systems." The Register provides more analysis: "Microsoft is, of course, a partner on the California system, since you can't ignore Windows in the data center, and presumably, Hyper-V will be supported alongside ESX Server on the hypervisors. (No one at the Cisco launch answered that and many other questions seeking details). ... The one thing that Cisco is clear on is who is signing off on these deals: the CIO. Cisco and its partners are going right to the top to push the California systems, right over the heads of server, storage, and network managers who want to protect their own fiefdoms."
Related Stories
This discussion has been archived.
No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
Full
Abbreviated
Hidden
Loading... please wait.
Misread title... (Score:5, Funny)
Re:Misread title... (Score:5, Funny)
Thought title said: "Costco Barges Into The Server Market". If so, I would've renewed my Costco card to get some cheap servers.
Yeah, but you've gotta buy a whole datacenter if you want one.
Parent
Huh!?!? (Score:5, Insightful)
Cisco...saving money?!?! Right.
Re:Huh!?!? (Score:4, Informative)
Cisco is expensive, but their stuff is incredibly well supported, both by them and third parties. Our entire network infrastructure is Cisco, and I'm a fan of the "one throat to choke" approach. Engineers are more worried about solving my problems than blaming someone else.
Even their lowest level of SMARTNet offers next day hardware replacement...
You're correct in that Cisco is expensive up front, but in the big picture (at least for our organization) it's really not a bad deal.
Parent
Redundancy (Score:5, Funny)
A server blade in a switch works WAY much better than a switch blade in a server...
Why use bleeding edge intel chips? (Score:5, Interesting)
I have to ask : why Nehalem EP Xeons? Those are the absolute bleeding edge chips that Intel manufactures, and as such as the most expensive by a significant margin. Newegg doesn't even have the chip listed on their website, yet carries 91 different server CPU models. While space inside the data center does cost money, and so does electricity, is it really so expensive as to be worth paying for a chip that is probably 10 times as expensive per MIP as cheaper alternatives? The motherboards are more expensive as well, especially when you factor in the huge markup for server grade parts.
The only advantage of the Nehalem is that it is SLIGHTLY faster per processing thread, but networking is usually an "embarassingly parallel" problem.
Re:Why use bleeding edge intel chips? (Score:5, Insightful)
haven't read the details, but intel probably put more virtualization logic into the CPU like they have been doing for the last few years. the price isn't that big a deal if you can put more VM's per CPU core than on the older chips
Parent
Re: (Score:2)
Re:Why use bleeding edge intel chips? (Score:4, Insightful)
There are limits. Running more VM's on a CPU costs power, and makes that physical device a single point of failure for multiple environments. Balancing such environments turns out to be much, much trickier than a lot of people like to admit: the very clever and sophisticated software to swap around live environments or do load balancing becomes its _own_ point of failure, bringing down entire racks of equipment in intermittent or even complete failures.
And yes, I've had this happen with servers with "five nines" uptime lauded and promised but somehow, never actually written into the contract. It would have been a lot cheaper to simply do a regular backup schedule and have a second rack of more capable, cheaper, failover equipment.
Parent
Re: (Score:3, Interesting)
This.
I work for a company that resells an application that recommends DoubleTake for hot failover. While it does indeed work, it is an administrative nightmare and very difficult to set up PROPERLY. Plus, failback never works: it's much easier to just fail "forward".
Now, the fault of a lot of this is the application, not DoubleTake. However, the solution always appears brittle, and the cost of "false failovers" is very high.
Re: (Score:3, Interesting)
You dont seem to be too familiar with vmware, and it's lack of single points of failure when implemented correctly. Sure, something can fai, but everything else should be able to pickup the slack.
Also, when you're paying per CPU 3K for Vmware licenses, another 3k for MS datacenter licenses, and who knows how much for each license on on each virtual server instance.... that extra 30 watts you're worried about is NOTHING if you can cram 2 more virtual servers onto a CPU.
Re: (Score:3, Interesting)
Re: (Score:3, Insightful)
In the first corner we have...
VMWare ESX + HA, dual power supplies, fibre channel (read: real) SAN, shared SCSI(FCP) storage, centralized management console
and the new challenger...
Xen, single power supply, SATA, swappable USB storage for failover, command line
Those are from two totally different divisions... planets... galaxies, what have you. You cannot compare the two.
You don't actually talk someone out of a SAN and into buying USB storage, or from ESX+HA to physically swapping external storage. It do
Re:Why use bleeding edge intel chips? (Score:5, Insightful)
I have to ask : why Nehalem EP Xeons? (...) the most expensive by a significant margin. The motherboards are more expensive as well,
Expensive - Cisco, so what's the part you don't understand?
Parent
Re: (Score:3, Insightful)
I have to ask : why Nehalem EP Xeons?
Because they are there to make a splash with ridiculous specs. Specs that wouldn't be possible without Nehalem. They're innocent looking enough (we have a few), but they are here to make an impact.
Re:Why use bleeding edge intel chips? (Score:5, Interesting)
One of the main things that would make Nehalem attractive is a few things for blade servers. The process 45nm moving on to 32nm later will provide the smallest footprint while not giving up any CPU power. Also, the Nehalems (Bloomfield and Core i7) that I have tested don't seem to offer much in terms of better performance, but the power usage is considerably lower. Also, FB-DIMMS (and DDR2) were a bit too consumptive of power, the newer memory technology is an attempt to reduce power consumption. Also, the CPI (formerly CSI) offers the Intel CPUs a Hypertransport-like interconnect system allowing system builders to scale. The footprint of 16+ core systems with Nehalem will be far smaller than the previous generations of Xeon MP processors.
I think the idea is more memory, more CPU processing power, less power and heat and scalability with a given architecture.
I noticed in my testing the L1 Data and instruction caches (32KB per core) in the Core i7 is one cycle more latent than the core2 (4 cycles vs 3), the L2 cache (which is 256KB per core rather than the 4-6MB per two cores in Core2) is faster, down from 17 cycles to 12 cycles. With this boost in L2-speed came a cut of 3.75MB-5.75MB in size. The way they mitigated that loss was to give a "large" 8MB L3 cache that runs on a slower clock. This new system, along with a hyper-threading implementation that, unlike the previous one, seems to genuinely enhance performance in nearly every test, allows Intel to make top-performing chips (see CPU 2006 @ spec.org for the latest results) that scale better via QPI, use less power and fit into smaller spaces than previous chips.
See:
Sorted SPEC CPU2006 Integer and Floating Point [spec.org]
CINT2006
Hardware Vendor System Result Baseline # Cores # Chips # Cores Per Chip Published Disclosure
1) YOYOtech Fi7EPOWER MLK1610 (Intel Core i7-965) 36.0 32.5 4 1 4 Jan-2009
2) ASUSTeK Computer Inc. ASUS P6T WS PRO workstation motherboard (Intel Core i7-965 Extreme Edition) 35.2 31.5 4 1 4
3) ASUSTeK Computer Inc. Asus P6T Deluxe (Intel Core i7-965 Extreme Edition) 33.6 30.2 4 1 4 Nov-2008
4) ASUSTeK Computer Inc. Asus P6T Deluxe (Intel Core i7-940) 30.8 27.8 4 1 4
5) Dell Inc. Dell Precision T7400 (Intel Xeon X5492, 3.40 GHz) 30.2 27.6 8 2 4
CFP2006
Hardware Vendor System Result Baseline # Cores # Chips # Cores Per Chip Published Disclosure
1) ASUSTeK Computer Inc. ASUS P6T WS PRO workstation motherboard (Intel Core i7-965 Extreme Edition) 39.3 37.4 4 1 4 Feb-2009
2) YOYOtech Fi7EPOWER MLK1610 (Intel Core i7-965) 35.7 33.6 4 1 4 Jan-2009
3)ASUSTeK Computer Inc. Asus P6T Deluxe (Intel Core i7-965 Extreme Edition) 33.6
Parent
Re:Why use bleeding edge intel chips? (Score:5, Informative)
I work for Sun, and have been beta testing Nehalems for almost a year now.
Sun is also the PRIMARY Nehalem vendor for Intel. We got the special treatment (I don't know how), but we get to be the first real Tier-1 vendor shipping Nehalems, and let me tell you that Intel has helped us a lot in hardware integration and software tuning.
The end result is that Nehalem EP (which are dual-socket systems) is significantly faster than any of the Core2 series, and spanks even the AMD Shanghais. They've gone to the on-chip memory controller ala Opterons, and it's helped considerably. In addition, they've redone the interconnect bus to make is much more HyperTransport-ish (though HT 3.x is still superior) - it's called QuickPath Interconnect. The overall result is much better performance under load, even for single-threaded apps.
For an application such as Virtualization, Nehalems are well worth the $$$. You get considerably better loaded performance than previous Intel CPUs, and with VMs, high system utilization is the GOAL. Up until now, AMDs were considerably better than Intel chips under high load, but the Nehalems just stole the dual-socket crown back.
I'm still waiting to get my hands on the EX series (quad-socket), so I don't know how they'll compare to AMD's 8000-series. Be interesting to see.
-Erik
Parent
Re:Why use bleeding edge intel chips? (Score:4, Funny)
Wow... you work for Sun, you've got a five-digit UID, and it's binary? You're the ultimate Slashdot trifecta. :-)
Parent
Re:Why use bleeding edge intel chips? (Score:5, Insightful)
I have not put it through the ringer with other tests...
It's wringer. You put things through a wringer [wikipedia.org] to squeeze the water out of them. This is very stressful. If the item in question isn't clothing, and isn't strong, it's likely to break. If it's a person, it's likely to punch the shit out of you when you're done.
Parent
Re: (Score:3, Interesting)
I have to ask : why Nehalem EP Xeons?
Because if you are a networking hardware vendor looking to crash the server hardware vendors' party, you don't engineer your product with yesterday's technology. Cisco is in a cash position where they could sell this as a loss leader to get their foot in the door, or at least until the price on the components come down as the Nehalem becomes more competitive.
Not to mention, the cost of the cpu components themselves is a relatively small portion of the overall cost of a system this size. Maybe in a low-e
Re: (Score:3, Interesting)
Well, the part was only released a month ago. There's a significant speed boost per processing core versus the older core 2/core quad line, and a new socket type/new ram type. As of right now, there's nothing faster that money can buy in the x86 architecture. Heck, for generalized processing it's probably the fastest chip money can buy. A new supercomputer would likely run best with a massive array of thousands of these things.
Re: (Score:3, Informative)
Sounds expensive (Score:5, Insightful)
Sounds like the right architecture, but at a price.
It amazes me that so many "enterprise" IT companies can sell what are essentially just Linux servers [networkworld.com] with their brand name tacked-on, at a 5000% mark-up.
Re: (Score:2)
Don't knock it, it's just the Enterprice [slashdot.org]!
Now more than just hardware.. (Score:5, Insightful)
I used to think along the same lines as you, that with 'reasonably competent' administration, it's all a wash.
And now after a stint in the industry, I've realized a lot of the industry is unable or unwilling to invest what is required to make effective use of hardware. The stuff in general can be complex and many companies are content to pay a premium to the vendor to tap into their aggregated skills rather than probably pay even more to have architects of their own with the experience and skills to match the vendor.
In this case, they are dressing up some core technologies that are pretty well understood, wrapping up it all with a lot of buzzwords, and pushing forward. The technical cynic in me shrugs, but I recognize what they *claim* to be trying to do may be valuable to some people.
That said, after years of struggling with Cisco's repeated decisions to support their proprietary standards to the exclusion of industry standards make me not want to touch their equipment or embrace any 'full management' stack they would want to give me. Some of it does the job sufficiently, but buying into a platform that makes it difficult to entertain competing product is something I like to avoid.
Parent
Re: (Score:3, Funny)
yeah, but will it be 32bit only? (Score:4, Insightful)
it's 2009, years after we were supposed to have flying cars, most new computers are 64bit, and Cisco refuses to release a 64bit IPSec client For x64 (64-bit) Windows support, you must utilize Cisco's next-generation Cisco AnyConnect VPN Client." [cisco.com]. So umm...we're supposed to think they have any clue what's going on above layer 4 these days? What are they going to be installing on these servers, Windows2000?
Re:yeah, but will it be 32bit only? (Score:5, Informative)
Parent
Apple sitting on $28 billion (Score:4, Interesting)
Not sure that Cisco is such a lone cash giant as suggested. Apple has $28 billion in reserves as of Jan 22, 2009 [wsj.com]. With the recent economic fiasco, both Cisco and Apple might be in different positions.
It has more than $30 billion in cash, more than any other tech company.
Seth
I've read about servers in shipping containers (Score:2)
This "barges" idea sounds like the next logical step.
Microsoft, of course ? (Score:4, Informative)
Microsoft is, of course, a partner on the California system, since you can't ignore Windows in the data center
Microsoft is supposed to have about 30% of the server market [netcraft.com], so I am not sure I get that of course.
Re:Microsoft, of course ? (Score:5, Insightful)
There are many different reasons you might want a server, web presence is only one.
Parent
Blah Blah Blah (Score:3, Insightful)
Re:Blah Blah Blah (Score:5, Insightful)
They suck at everything else.
Not everything they do is perfect, but they broke into the Fibre Channel switching business quite effectively. They can, and do, break into new markets. Servers are a logical step for them since there's a huge advantage to providing a vertical stack of networking, servers, and whatever else they can muster.
Parent
Re: (Score:3, Interesting)
Yes, at prices that will make high-end Server 2008 enterprise installs look cheap.
Re: (Score:3, Informative)
But from what I've seen, their server technology appears relatively weak. I.e. their blades appear less dense than 1U servers.
Not true. It's 6U for 8 blades. I just took a look at the chassis that I have access to. Now, HP c7000 is better density than this at 16 blades in 10U. But I just want to be clear, Cisco's chassis is not less dense than 1U servers.
Re: (Score:3, Insightful)
We're just seeing what the graphics folks would call "one more time around the wheel of karma": a prototype of something new is done in software, implemented in proprietary hardware for speed, and soon afterwards the next generation of main CPUs catches up to the point where it can be done in software once more, usually with some small amount of custom silicon on the interface board.
--dave
They go for the "soft" target (Score:5, Insightful)
The one thing that Cisco is clear on is who is signing off on these deals: the CIO. Cisco and its partners are going right to the top to push the California systems, right over the heads of server, storage, and network managers who want to protect their own fiefdoms.
Presumably, they are doing this because they know that the CIOs, on average, are less well informed than their technical subordinates. It is a classic salesman's tactic: go straight to the "decision maker." I'm not saying that CIOs are not well qualified and intelligent people (I'm sure that most are). However, at the CxO level in a large company, you are a strategic thinker. You are most likely not going to be on the bleeding edge of the latest hardware trend.
To put it another way, the CIO is the "soft" target. You always go for the soft target.
Naturally, Cisco (and other vendors) know this. Hence, you go after the CIO and dazzle him with fancy presentations and wine and dine him and viola, you get a big sale. This how MS does it, and how other big tech companies do it.
If you are fortunate enough to have the ear of your CIO, make sure to warn him about snake oil peddlers.
Re: (Score:3, Insightful)
Despite what your ID number says; you are new around here, aren't you?
Re: (Score:3, Funny)
This 'blade' computing, is that VME bus or S-100?
Re:They go for the "soft" target (Score:5, Interesting)
This is very true. I am currently evaluating a forklift upgrade of one of my POPs, and we're looking at the Cisco vs. Juniper proposition.
While I'm a VP level operational head at an ISP, the Cisco rep told me straight out that he doesn't typically engage technical people like me when he comes in. He typically talks to the C level people, and it shows, because he's not keeping up with the Juniper rep. The Juniper team has already put me in front of many technical product development people, and the depth of the conversations have been truly refreshing. I'm feeling more and more comfortable with going Juniper as the days go by.
Parent
A good move (Score:4, Interesting)
For large networks and satellite office, you have a server or 2, a phone system, network gear, maybe some video surveillance gear. They'll walk into the CIO's office and say:
"you have all this gear from different vendors, with different support contracts and different departments finger pointing when problems arise."
"Now here is the cisco way, one box, one department, one vendor to call. Stick it in a closet and forget about it. Let us show you all our management tools which show everything in a single pane of glass"
If they do it right, it'll make for a very slick demo.
This is their attempt to do the same in the datacenter.
Re: (Score:3, Insightful)
Exactly. The grandparent's argument falls on its face as soon as you hook up your do-everything-box to a telco circuit.
WOW. Innovation at every step! (Score:5, Funny)
Amazing stuff.
I'm stunned. One box, with processor, storage, and networking -- ALL TOGETHER in one package. Who would have thought that would be possible?
If you can't make the sale, move up the org chart! (Score:5, Interesting)
Yep. That's the Cisco I know and loath. If you can't convince the literate, just move up the org chart.
Years ago, at my institution (150+ buildings, about 15K active IP addresses,) we did a cost analysis of our Cisco addition and decided that it was unnecessary. We could do everything we needed with cheaper, commodity devices.
So, for the next couple years, all upgrades/replacements were to simpler structures. To non-proprietary protocols. And to non-Cisco equipment. We have been Cisco-Free for about 4 years.
The hardest part was beating off the attacks from Cisco Sales. These attacks were vicious. They lied (even more than usual for Cisco sales droids.) They tried their best to discredit us. First they approached the head of IT. Then the VP for Business. Then the president.
Finally, they went to the Board of Regents. They said we were incompetent. They said our actions were endangering the future of our institution. Fortunately, the Regents decided to let us try it.
It has worked out great for us. Our capability is up. Our reliability is way up. Our security is up. Our costs are down (about 1/2 the price of equivalent Cisco.)
But, it only happened because upper management was willing to trust us. I get the impression that most management would fold under the pressure we saw.
Miles
Re: (Score:3, Interesting)
At the small ISP I worked at, we pretty much bought into Cisco for several years. We had an AS5200 PRI for our full 56k PRI lines, and a 3000 series model (can't recall which one) as our gateway router. This worked fine until we started rolling out some more advanced networking, such as proprietary 900mhz and 2.4ghz wireless. Suddenly we were faced with either having to upgrade this equipment (some of it not so young), and the costs were not insignificant.
I asked my boss to give me a couple of weeks to s
All smoke and no fire.... (Score:5, Interesting)
OK. My bias up front - I work for Sun.
That said, there were several pre-Cisco-announcements from HP, IBM, and Sun about how the California system is a no-go. Admittedly, they're the competitors for Cisco, but after having looked at the existing rack blade/switch systems from those three vendors, I really don't see any difference worth mentioning from current product lines.
Here's some thoughts:
Overall, this looks like a stupid move. I realize that Cisco needs to look for more revenue streams in the face of the commoditizing of most network gear, but this seems like an '80s solution to a 2010 problem.
-Erik
Re: (Score:3, Informative)
OK, my bias up front, I work for Cisco, though not directly in the California project.
Cisco doesn't have integrated solutions. All the others provide storage, network, and compute integration with large, well-trained Professional Services orgs. Cisco has CCIEs in piles, but what do they know about anything but network gear?
Cisco has people like me and I'm far from alone.
I'm very picky about where I accept employment from and I CHOSE to work at Cisco even though I despise Northern California. For whatever it's worth.
Brought to you by the authors of IOS (Score:4, Insightful)
Any interface or management tools for such virtualized pool systems, brought to you by the company that wrote and maintains IOSS, already has a serious hurdle to jump. That language is very, very clever and flexible. It is also absolutely awful for attempting to do simple, straightforward configuration tasks that _look_ like they should work from the documentation, but which extra steps to actually provide and can only be done directly from the text interface, not from _any_ of the GUI's. (I spent some time trying to configure proper failover behavior last year on a set of switches. It was painful.)
If their virtualization management tools are similarly powerful, flexible, and utterly useless to anyone not a highly trained Cisco technician, then these systems will be very expensive and under-utilized doorstops in most environments. If, however, they've fired the fools who wrote the IOS control language and replaced them with people being fired by Juniper and thus raised the IQ at both companies, then they have a chance to fill a serious niche in environments like Wal-Mart, where a centrally managed and flexible system could manage inventory, sales registers, their highly automated security and power management, and personnel records. Putting all those on different servers provided by different vendors is nightmarish to manage: having a consistent, virtualized hardware environment makes hardware repairs and upgrades far, far more efficient, in my experience.
And let's face it: when I've looked at the stock room at such a facility, I've often seen a server or desktop stashed under someone's desk, lying on top of a card table in a corner where it never got mounted but is still in use, or miscabled and unscrewed down with only one plug in on top of a rack with an overloaded UPS. Simplifying and modularizing that for such an environment would be a big headache that I wouldn't have to have for dealing with such partners and clients. (I highly recommend taking a cell phone and taking pictures of such setups to let their head office know there is a problem.)
Because.. (Score:4, Interesting)
There is exceedingly fat margins in storage, *even* by cisco standards, and that's a high benchmark. The SAN market is a vendors dream, where nickel and diming every little feature and even every little port on every switch is the status quo, except the nickles and dimes are more like 5 and 10 thousand dollars.
As far as technical reasons, they are mostly not there. One exception is that Cisco is pushing to replace FC with Ethernet, presumably with the promise of an escape from the painful FC market practices. Though assuredly they will bring some of the market behaviors over, they will make it somewhat easier to make the sale. They tried to just release a product into that market against the likes of Brocade and QLogic, but I think Cisco has realized their only substantial chance to stay vital is to suck in storage infrastructure into their fabric they have some reputation in, ethernet.
People are already starting more and more to consider other vendors 'good enough' for traditional networking needs. Cisco wants to own the whole mess so that people will be more afraid to move off.
Parent