Article Written by Arsalan Farooq - Vice President, Cloud Products at Accelerite
As
we enter the new year, we tend to reflect on the past and look ahead at
what needs to change with many of us declaring resolutions. Well, yes
of course, this is the year we're going to get fit physically, sort out
our personal finances and finish that house painting job we started in
2011.
On the business side, I offer these four resolutions to improve management of virtualization and cloud platforms.
1- Get Shadow IT Under Control
This is one of the biggest problems as enterprises convert from
on-premises data centers to purchasing compute power. Now, anyone in the
enterprise can go rogue and buy computing resources outside the control
of IT. The biggest issues with this practice, of course, are runaway
costs - or unmonitored, at very least - and loss of control, namely
working outside corporate compliance guidelines for security of data.
Both scenarios are unacceptable for any enterprise. If this hasn't
already been done, it's time to use available tools to monitor and
inventory computing assets on-premises, as well as the cloud. Assure
compliance and cost controls are in place. Cloud computing is a terrific
resource but must be maintained with the same rigor as on-premise data
centers.
2- Leverage Open Source and then Pay for Features, Value
Remember how companies tried to sell us "premium" speaker wire and HDMI
cables? And then we all realized that the $10 cables worked exactly the
same as the $100 cables? Wouldn't we rather put that money toward
buying components that add real value, like a better pair of headphones?
Same
goes for data center tech. Our first reflex should always be to look at
open source for the "plumbing" and then spend money on features higher
up in the stack that deliver real value. And, in fact, there is a
tremendous wealth of open source technology available that should be
leveraged whenever and wherever possible. Take core virtualization: I
can't think of any good reason to pay for a "brand-name" hypervisor when
there is open source KVM and Xen. For that matter, while Hyper-V is not
open source, it is a perfectly viable option to paying separately for
VMware. Another metaphor here is buying a Porsche to drive to the train
station. You'll get there, but there are many other great ways to get
there for a lot less money.
And
where should that money go? Management and automation are good places
to pay for features. We have one customer that made a wholesale change
from VMware to Linux (KVM) as the core virtualization technology and
uses automation software to manage everything. They're very happy and
their costs went down dramatically.
3- Centralize Management of Virtualization and Clouds
It's inefficient to use the separate management and automation that
comes with each different virtualization and cloud platform, or in some
cases with scripts and homegrown tools. Administration that is
splintered across a number of tools and technologies makes day-to-day
management inefficient, unnecessarily complex and error prone. Address
this fragmentation problem by instead using available automation and
management software that bridges platforms and pools resources. This
leverages administrative resources and scales operations for full
lifecycle management, including a centralized web-based console, along
with template-based provisioning that enables user self-service access
to computing infrastructure.
4- Ensure High Availability of All Compute Resources
This was much easier before the days of virtualized servers and cloud
computing when each physical server had its own backup. We must apply
the same diligence today with effective failover protection against
hardware and operating system outages. That includes those in the cloud.
If you haven't already, choose automation software that works across
virtualization types and cloud computing platforms and is capable of
handling live migrations and will reliably protect your VMs and server
pools with enterprise-class, policy-based backup and restore.
##
About the Author
Arsalan Farooq is vice president, cloud products at Accelerite, which he joined when Convirture was acquired in
2015. Previously, Farooq was CEO and co-founder of Convirture, started
in 2006, and before that he founded Oracle's Application Service Level
Management division, growing it from two engineers to a multi-national
organization. His career started as a software designer and in college
he founded a technology consultancy. He holds degrees in Theoretical
Physics and Computer Science from Reed College and Caltech.
