One headache for information technology professionals is the constant barrage
of software upgrades necessary to avoid security threats and improve mission-critical software applications. In the past,
these upgrades meant planned downtime
and after-hours or weekend work for the IT
staff. Worse yet, some upgrades weren’t tested and IT personnel faced the wrath of the
user community on Monday morning.
The convergence of two technologies —
high-availability software and server virtualization software — has spawned a solution
to this age-old problem. A business can
build multiple virtual servers using this new
management software, creating production,
testing and development copies of its systems. Then, using the high-availability software, the business can copy or clone its production system to the testing copy. The testing version is used to load the upgrades
without disrupting the production system,
but uses a live copy of the data.
“Here is the best part,” says Martin
Sizemore, chief technology officer for
Premier Technologies. “When the testing is
complete and the upgrade is ready to go
live, then the testing virtual copy trades
places with the production version, and cutover is complete. In minutes, a tested, ready-to-run, upgraded version of the system goes
live without disrupting operations.”
Smart Business spoke with him about
these two key technologies and the productivity they offer for business operations.
What does it take for companies to implement virtualization software?
Virtualization software is generally
installed on new hardware platforms.
However, it can be retrofitted onto many
modern operating systems. Virtualization
software simply ‘carves up’ one physical
server into multiple logical servers. The key
requirement is that the disk storage is separate from the server, usually a storage area
network (SAN) and that the storage is logically assigned to the server. This approach
makes the storage virtual and allows the
hardware servers to become virtual.
Companies take the step of virtualization
to maximize their investment in hardware servers that are idle a portion of the time and
disk storage that is scattered in separate
servers. Implementing virtualization software can be done in stages to create the
environment for better controlling upgrades.
How does the high-availability software keep
the production and testing environments in
sync for the cutover?
Once the virtual servers are in place, the
high-availability software is used to set up
the replication, which is constant copying
of the data from the production server to
the testing or upgrade server, in real time.
The ability to constantly keep two copies
of production data running with the applications is the secret to minimizing disruption from the upgrade. This continuous
availability allows IT personnel to select
the proper moment in time, after testing, to
switch the production and testing servers.
After data replication is working, the IT
staff can apply the upgrades — operating
system upgrades, application software
upgrades, reporting software; the list is
endless. Because this is a controlled environment (but a copy of the live environment), quality assurance testing is more
realistic and a gap analysis is more accurate for decision-making.
What are the advantages of this upgrade
approach over traditional methods?
Traditional methods must rely on vendors to test the upgrades. However, vendors cannot anticipate the combination of
applications running at any given company.
As a result, the more operating systems or
applications that a company runs, the higher the risk of upgrades simply due to the
possible number of untested combinations. The advantage of this approach is the
ability to test one or more upgrades in a
controlled manner prior to launching the
production environment. In addition, the
testing is done with real data and real data
volumes to uncover problems that smaller
unit tests will miss.
With the speed of upgrades increasing,
the need for automating and streamlining
the upgrade process is becoming critical.
The cost of planned or unplanned down-time is dramatically increasing as well, due
to increasing dependence on computer
systems to operate the business. The simple loss of an e-mail system or downtime
due to patches on your order entry system
can have employees completely idled and
potential losses looming.
Why isn’t this solution more widely used?
This solution of upgrades without disruption needs to be more widely understood
by company leadership and move beyond
the view of IT’s chase of new technologies.
Reliability of information systems cannot
be taken for granted, especially as application or operating system vendors scramble
to address security needs and stop the next
attack.
More importantly, a sensible, repeatable
and planned approach is the best insurance of keeping productivity high for your
company while maintaining the integrity of
mission-critical systems. The key is to
begin the adoption cycle of virtualization
software and combine it with the insurance of the high-availability software.
MARTIN SIZEMORE is chief technology officer for Premier
Technologies. Reach him at [email protected] or (412)
788-8080.