Cisco Systems will introduce “Project California,” a blade system specially tuned for virtualized environments, at an event in San Francisco on March 16, InternetNews.com has learned.
The news comes as talk surfaces around Cisco’s other strategies to ramp up its virtualization portfolio: namely, through an acquisition of VMware or its parent, EMC. So far, those talks never moved beyond an initial phase, although industry observers and Wall Street continue to watch for signs of renewed discussions among the three.
For the time being, however, Cisco’s efforts in virtualization focus primarily on the launch of its new blade system. According to a source familiar with the products, the blades will be based on Intel’s Core i7 processors and come with up to 192GB of memory, well above the maximum capacity of 128GB in today’s blades. Intel (NASDAQ: INTC) recently announced it would begin shipping Core i7 Xeon processors, codenamed Nehalem-EP, as part of its Xeon 5000 series.
The blades include a PCI-Express connection, allowing them to connect to Cisco’s (NASDAQ: CSCO) high-speed Unified Fabric architecture. These connections also give the blades very fast Ethernet access to both the network and storage devices and eliminate the need for a storage-area network (SAN). Instead, the blades would talk directly to the storage servers.
The blade servers are believed to come with Cisco’s Nexus
5000 switches embedded in the chassis, which support the Unified Fabric and is built to be virtualization-ready. The servers will also feature tight integration with and support for VMware software.
This would put computing and networking power all in a single box.
“It’s more of making the computer part of the network, thus Unified Computing,” said the source, turning Sun Microsystems’s infamous “The network is the computer” slogan on its ear.
The term “Unified Computing” was first floated by Cisco CTO Padmasree
Warrior in a January blog post, where she described it as “the advancement toward the next
generation data center that links all resources together in a common
architecture to reduce the barrier to entry for data center virtualization.
“In other words, the compute and storage platform is architecturally
‘unified’ with the network and the virtualization platform,” she added.
A spokesperson for Cisco declined to comment on the story.
Next page: Built for virtualization
Page 2 of 3
Built for virtualization
The source went on to say that Cisco has no interest in the broad, general purpose market for blades, only for virtualized systems. Analysts noted that most blades are virtualized anyway, making them de facto mass market products.
“On any blades run now, people are running virtual workloads,”
said Andi Mann, senior analyst with Enterprise Management Associates.
But he added “this does not sound like just a copycat move and trying to
get into a market someone else owns. This is Cisco looking at the market
and extending it in their own way.”
Gordon Haff, principal IT advisor with Illuminata, felt the same way,
and is less sold on the idea. “If you leave aside high performance
computing, that’s like saying you’re going after where all servers are
going,” he told InternetNews.com. “The direction for mainstream
servers will be to be virtualized. I don’t see that as a tiny niche of
servers.”
The vast amount of memory would make it ideal for putting up to 100
virtual machines on the server, a problem blades currently have in that
due to memory constraints, they can only run so many virtual machines.
By using Core i7 instead of Xeon, Cisco gets DDR3 memory, which is
much faster than DDR2 and draws less power. It’s also a much lower power
draw than FBDIMM used in current Xeons, which draw much more power than
DDR memory and run hotter.
Also, the Core i7 has the new QuickPath Interconnect, which allows
for much greater bandwidth than any prior x86 processor, greatly
improving memory performance. After I/O, memory performance is the
second biggest issue vexing virtualized servers.
The goal of Cisco’s Project California is to remove the bottlenecks of
virtualization, at the memory and adapter levels. The faster Core i7 and
large amount of memory will improve performance within the blade, and
the connection to Unified Fabric will make it easier to move data in and
out of the blade, as well as moving virtual machines around with
VMware’s VMotion.
The two analysts were split in their view of the features and strategy.
“This makes those blades extremely agile on how they can move resources
around,” said Mann. “The bottlenecks on blades are network traffic.
Saturation of the network is much more of a problem then saturation of
memory. This gives enterprises an extremely efficient way of using
computer resources that utilize network connections and storage
connections.”
Hoff is not sold. “I can appreciate the general point Cisco is
making, going after a specific class of servers, probably relatively
higher-priced and higher margins. But I will point out that a niche
strategy has not been an effective strategy for server makers for the
past decade,” he said.
“I can’t think of a good example of a large storage vendor who has
gone into storage servers in a niche way.”
Next page: Returning fire at HP?
Page 3 of 3
Returning fire at HP?
Why Cisco would enter a new market, one dominated by HP, IBM and
Dell, three companies quite capable of putting up a fight, seems
unusual. What may have stirred all of this might very well be HP nudging
into Cisco’s territory first.
HP has its own networking equipment, the ProCurve line. The previous
CEO of HP, Carly Fiorina, reportedly encouraged HP’s sales force to promote Cisco products over ProCurve during her tenure. However, the new chief, Mark Hurd, made ProCurve a part of the company’s growth strategy and is directly taking on Cisco in this area.
So is this Cisco’s way of telling HP to step off? The source familiar
with Project California thinks so, and so does Haff.
“I’m sure the people involved in it would see it that way,” he said.
“It sounds like a toe in the water, an exploration, a competitive
counter and so forth, than it does a broad based foray into the server
market.”
He added “HP’s doing well with ProCurve, I don’t see them backing off
where they are having success.”
Mann agreed. “It’s not necessarily a shot directly at HP, but it’s a
shot across their bow,” he said. “They can’t afford to put aside someone
like HP or IBM. The hardware in a datacenter is predominantly HP and
IBM. They’ve got to work with those companies. If they shut either one
out they dig themselves a hole.
“But it could be a message to HP that there’s overlap and Cisco will be aggressive in the datacenter market.”
