
Virtualization and Cloud executives share their predictions for 2014. Read them in this VMblog.com series exclusive.
Contributed article by Irfan Ahmad, Co-Founder of CloudPhysics
In 2014, Analytics will be the Only Real Differentiator Left in Storage
The game of speeds and feeds is over. It's getting harder and
harder to tell the difference between storage vendors, incumbents and startups
through the usual lens of capacity or latency or IOPs. There are no disruptive
benefits left doing things the old way.
When you peel back the curtain, most of the special sauce in
all the new approaches to storage is in the software
that lives on the devices and in the packages and the analytics performed by that software. Even amongst server-side
caching competitors, those that have (or will) separate from the pack do it
with capabilities that are rooted in the analytics software layered into their
solutions.
This software is supposed
to deliver the benefits of easier management, resource pooling, granular QoS,
predictable performance, automation, etc. At least that's been the promise of storage virtualization in the
past and of software-defined storage today. But we're not there yet.
So what's missing? Fundamentally, what's lacking is the
combination of visibility and predictive analytics that:
- Provide
you with deep visibility into what's going on
- Predict
when and where you will run out of space
- Predict
when and where, in terms of both applications/VMs/containers
and datastores, you will experience contention
- Provide
levers and automation for enforcing granular QoS for both capacity and
performance
- Enable
you to buy no more than what you need as well as no less
Long term, the only storage vendors--hardware or software--who
will succeed are those who deliver on these promises through deep visibility
and meaningful analytics.
##
About the Author
In his nine years at VMware, Irfan was R&D lead and
co-inventor for flagship products, including Storage DRS and Storage I/O
Control. He worked extensively on interdisciplinary endeavors in memory,
storage, CPU, and distributed resource management, and developed a special
interest in research at the intersection of systems. He also spent several
years in performance analysis and optimization, both in systems software and OS
kernels. Before VMware, he worked on a software microprocessor at Transmeta.
His peer-reviewed research has been featured at ACM SOCC
(best paper), USENIX ATC, USENIX FAST, and IEEE IISWC. He was honored to have
chaired HotStorage and VMware's R&D Innovation Conference (RADIO). He is an
inventor on seven issued patents and another 20 pending across memory
management, page replacement, quality of service management, I/O scheduling and
caching. He earned his pink tie from the University of Waterloo.