Many vendors and users have been focusing on how to urge storage
capacity, as information has seen huge expansion rates in a past few years. But how does storage
performance magnitude up? What options do vendors have for improving and assessing storage
performance, and are users holding a record seriously?
In his latest Storage repository column, Jeff Boles provides us with a storage
performance analysis. Small improvements have been done to storage opening solutions,
and Boles believes that can lead to even softened improvements. At a same time, users need to be
aware of a caveats that infrequently come with these solutions.
As we proceed a second half of 2011, it looks like we’re going to see a many different
data storage attention for a rest of a year. That’s a good thing since 2010 was a minefield
for vital formulation — a drum coaster of fad with a indeterminate destiny that goes
with it. Game-changing technologies done their genuine initial forays into a market, including cloud
storage use providers, cloud on-ramps, cloud backup, cloud disaster liberation (DR) and others.
In a new rising marketplace forecast, Taneja Group identified storage products compared with the
cloud as flourishing into a $10 billion marketplace by 2014. As we see it, we’re out of a embankment with new
product innovations and now intent in a solid march.
In 2010, there was misunderstanding among a businessman ranks like we’ve never seen, with vital vendor
strategy shifts, a series of tiny businessman failures and mind-boggling acquisitions. Whew! Surely,
all that businessman misunderstanding contingency be tapering off. We need a breather and we need to demeanour during what’s
coming.
There’s really one thing that needs some-more attention: storage performance. For too prolonged we’ve
been fixated on capacity, and that engrossment has sapped some of a creation appetite that
should have been destined toward elucidate storage
performance problems. And since of that neglect, many of us are anticipating we need better
performance than we can get from today’s customary approaches. It seems many of a activity in 2008
and 2009 was centered on capacity, and in 2010 that ability contention shifted to a cloud.
Lately, it seems opening usually comes adult when a theme is solid-state
storage, an choice unsuitable to many of a storage systems many of us have in a data
centers. That competence be a future, though opening needs to be addressed now.
More on storage performance
Understanding storage
performance metrics
What’s a attribute between storage capacity, storage function and storage
performance?
But I’m commencement to see performance
solutions that have some genuine promise. Discussions around solid-state record have done us
acutely wakeful of a large stipulations of many normal storage systems. And that seems to
have desirous some innovators to do something wholly different.
Despite what some vendors are peddling, it won’t be a subsequent era of ridiculously complex,
application-specific stacks of hardware that will speed adult your database or your series crunching.
And don’t design bigger and cheaper appliances to broach on that guarantee either. Those and similar
solutions have been aversion for vital planning. How are we ostensible to build a storage performance
strategy from a proliferation of one-trick ponies that can’t adjust to changing final or the
massive common entrance compulsory within today’s consolidating information centers?
The cost of delivering scalable performance
My mandate for doing softened aren’t that stringent. Fundamentally, a resolution should give
you genuine opening and guarantee to scale good over what we need today. Simultaneously, it
shouldn’t demeanour wholly opposite than your existent storage. It should still store information in a
permanent, arguable approach and be usually as serviceable. We’re saying vendors like Alacritech, Avere,
Kaminario and others consider tough about how to broach opening that fits into today’s storage
infrastructure though a large formation effort, and with a right capabilities to deliver
future scalable
performance though compromises in storage capabilities.
One outcome of these emerging, softened architectures for storage opening is a dramatic
change in what we compensate for I/O — when assessed on a basement of dollars per I/O, such systems have
massively softened costs than rotational disk. But those sum usually tell partial of a story. Bad
architectures come during a cost approach over a disproportion in upfront costs. The upfront costs of any
performance resolution with a bad design might good be compounded by a operational costs
incurred in a march of coping with and offsetting a bad architecture, such as:
- The cost of integration. Integrating solutions that residence a subset of information can require
enormous effort.
- The cost of information protection. Few solutions are built for primary long-term storage, so
the cost of information insurance can boost significantly. Worse nonetheless are a considerations for
DR.
- The cost of replacement. Solutions built though prolonged use life as a pivotal design
criterion might strech a beforehand finish of life. The cost here isn’t usually about replacement, but
dealing with a effects of intensity failure, mostly with strategies such as progressing idle spare
equipment.
- Cost of service. Some products aren’t built for in-place service, so they might require
specialized imagination for use or disruptive sum replacements if a disaster occurs.
- Cost of scale. A product that’s attractively labelled for a singular app might simply be a
mismatch for other demands. This can outcome in apps requiring mixed storage systems, capacity
underutilization or compromising on performance.
- Cost of management. Inadequate scaling can lead to stretch and outcome in carrying to deploy
multiple inclination to prove opposite focus needs. The cost of handling removed systems can
become enormous.
And a bigger things get, a some-more costs supplement up. I’m not suggesting your primary storage vendor
doesn’t have answers that residence these issues, as some are meditative flattering innovatively about how
to extend controller capabilities. But we need to make certain we have a right primary
data storage vendor if you’re shopping into their answer for performance. For some, their current
investments or dictated new investments might never go distant enough. There might good be a significant
deficiency in both a ensuing value you’re removing per dollar and in a rival business
capabilities of your ensuing IT systems.
It might be a year of usually incremental improvement, though when it comes to performance, even small
improvements can renovate how we do storage. With a few innovators on a horizon, it looks like
the rest of 2011 could see even larger strides.
BIO: Jeff Boles is a comparison researcher during Taneja Group. He can be reached during jeff@tanejagroup.com.
This essay was formerly published in Storage
magazine.
This was initial published in Jun 2011
Article source: http://www.pheedcontent.com/click.phdo?i=e0f42b8822778abc8ed2539e30c1412d
