Pat Gelsinger had recently come to EMC last fall after almost 30
years at Intel when he started to take a look at the projects going on
inside the data storage giant.
It was then that Gelsinger, now the president and COO
for EMC’s Information Infrastructure Products group, saw the work the
company’s researchers were doing into using a caching technique to help
grow the idea of federated data and conquer the distance quandary
that’s always vexed IT professionals.
“He said, ‘That’s it. That’s going to be big,’” Brian
Gallagher, divisional president of EMC’s Symmetrix and Virtualization
product group, said in a recent interview at the company’s Hopkinton,
Mass., headquarters.
Several months later, Gelsinger and other EMC
officials introduced their Virtual Storage strategy March 11 during a
conference with industry analysts and in a lengthy blog post by Chuck Hollis, vice president of global marketing and CTO at EMC.
What EMC introduced at the time was more a concept
rather than a product or roadmap, but the idea has generated a lot of
interest among analysts and others in the industry. And EMC officials
are working to keep that interest going as they enter the final
month before the EMC World 2010 show in Boston May 10-13.
Gallagher said Virtual Storage will be a key theme at
the show, and speculation is the first product tied to the Virtual
Storage strategy could be announced there.
The idea of Virtual Storage has grown out of the work that EMC, its VMware subsidiary and others have done in virtualization and cloud computing, particularly private clouds, Gallagher said.
Gelsinger described the private cloud as an
environment where businesses can seamlessly access pools of computing
and data resources in a secure fashion regardless of where they’re
housed—whether it’s in their own data centers or external compute cloud
environments hosted by others.
The pooling of these IT resources creates a federated
storage environment, and it’s an environment that the IT department
still has to have control over.
Key hurdles to this vision of a private cloud has
been the limitations of storage—in particular, the limitation of
distance. Even in storage virtualization environments, the data is
still tied to physical systems, and currently is traditionally limited
to individual data centers. Virtual Storage should not be confused with
storage virtualization, officials said.
“Data distance is among the most obstinate of data
center problems,” Charles King, an analyst with Pund-IT Research, wrote
in a March 17 report.
The biggest issue is that the farther away data is,
the large the problems that arise around latency, bandwidth and data
coherency, King said.
“Why? In part, because you can’t ‘fix’ the speed of
light—even in the best circumstances, the movement of data faces
inherent physical limits,” he wrote. “In addition, bandwidth is not
free and is also distributed unequally, particularly beyond urban and
suburban centers.”
Like server virtualization—where applications and
operating systems are separated from physical systems—EMC’s Virtual
Storage strategy will separate the data from physical storage devices.
The vendor is using technology gained through its acquisition of
YottaYotta in 2008 to create a “distributed cache coherence” to
facilitate and manage data across distributed computing environments.
The combination of distributed cache coherence and
global federation of storage will negate the challenges brought about
by distance, according to Gallagher.
“We’re eliminating distance as a problem and making distance an enabler,” he said.
“With global federation, multiple compute elements
see a shared local and global environment with resource pooling and the
illusion that petabytes of storage reside locally,” Jane Clabby, an
analyst with Clabby Associates, said in a March 17 report. “The same
set of technologies manages both local and global storage resources.”
In his blog, EMC’s Hollis said the possibilities of Virtual Storage are great.
“You'd be able to seriously consider moving thousands
of VMs over thousands of miles,” Hollis wrote. “Or you could easily
play the market on energy costs by moving workloads to where they're
most efficient to run.
“It'd be much easier to aggregate larger data centers
from smaller ones—or put information closer to users, if that was
needed. Applications could move to alternate locations on demand
and we'd eliminate major sources of downtime.”
It also is something that would benefit smaller business as much as larger enterprises, he said.
EMC officials up to this point have talked about
Virtual Storage more as a concept and direction rather than outline the
technologies, IP or products that will be involved. Gallagher described
the discussion around Virtual Storage as the first step in what will be
a journey.
“You’ll hear more and more about it, but it is real,” he said.
Gallagher did say that the first product will be a
physical appliance, but declined to elaborate on what it will be or
when it will be released. However, he added that Virtual Storage will
be discussed further at EMC World.
David Hill, an analyst with the Mesabi Group, said in
a March 17 report that Gelsinger said he was confident EMC has been
able to solve the traditional problems around storage and distance.
“EMC is now publicly committed to being able to build
practical storage federation using virtual storage on top of
distributed cache coherence,” Hill wrote.
Despite the lack of details, analysts were intrigued
by the possibilities of EMC’s vision. Pund-IT’s King said noted that
data centers could be built in places were energy is cheap and
plentiful, regardless of how far it is from end users. King and Clabby
also talked about the ability to move resources to where workers are in
what they called a “follow the sun, follow the moon” approach.
The analysts also outlined a scenario where data can be moved when an environmental threat—a hurricane, for example—arises.
Gallagher also said that businesses will be able to
run some lower-level tasks—such as batch processing—in data centers
were energy costs are lower.