LATEST INTELLIGENCE
Increasing Enterprise
Application
Performance with
Flash Storage Wide-ranging Business
Objectives Demand a
Wide Range of Flash
Solutions
Imagine a day in the life of a storage architect. Data
capacities have climbed to record levels. Meanwhile,
application users are demanding higher data throughput
and lower latencies. Easy, right? Well, yes, in some
respects. Faster CPUs blaze through I/O instructions more
quickly than ever. Multicore CPUs permit simultaneous I/O
instructions to be processed. Solid-state storage devices
can literally move data at lightning speed. So, yes, the
storage architect has many tools readily available, but
it’s still up to the architect to make the right choices, put
everything together and make it work flawlessly. Although various forms of solid-state storage have existed for
decades, the easy availability and use of affordable NAND
flash storage media is the first broadly deployable – and
deployed – iteration of the technology. During its relatively
short existence in commercial storage manifestations, flash
storage has made a lasting impact. Over the past five to
seven years, it has joined other relatively new technologies/
approaches (think of virtualization and cloud, for example) as a
key disrupter in the IT landscape.
In this paper, we examine three common considerations in
the design of storage architectures suitable for enterprise
applications: performance, management and Scale-out
capability. Although a single storage system could address
all three areas, it is often the case that one storage system
outshines the others in one aspect of enterprise
application support, while another storage system shines
in another area.
Enterprise Application Performance
When the utmost in application data performance is
desired, architects often look at industry performance
benchmarks to help determine which storage system to
use. The Storage Performance Council’s1 SPC-1 benchmark
simulates workloads from high-speed data applications
such as OLTP and missioncritical database operations.
What was in 2008 just a “vitamin supplement” to the regular
storage diet of traditional spinning disk is now available in a
whole “flash- aisle” of options, spanning hybrid implementations,
all-flash arrays, and server-side deployments. Savvy vendors are
delivering a portfolio of differentiated solutions that enable users
to integrate the optimum type and amount of flash for their
overall businesses and specific applications.
Along with the relentless growth of massive data volumes,
every organization has key applications that demand more
speed and/or IOPS bandwidth than others. Ironically, as disk
drives and systems get bigger, serving these performance
demands can get harder…which helps explain why flash is
becoming increasingly popular for its high performance, high
IOPS, and low latency. These abilities typically translate into
better ROI (think faster customer service, better profits, etc.),
and they can actually help drive down operational TCO (via
easier management, lower power consumption, etc.).
Download white papers free from www.intelligentcio.com/me/whitepapers/
www.intelligentcio.com
INTELLIGENTCIO
15