Disaster Recovery Journal

Focus: Hurricane Sandy

In Its Wake, Hurricane Sandy
Left Disaster Recovery Lessons

sons learned and best practices. Many of our customers don’t have the wherewithal, time, resources, or budget to maintain a robust program that ensures an always-ready recovery posture.

By DOMINICK PAUL

It formed off the coast of Nicaragua s a tropical storm. A week later, last Oct. 29, Hurricane Sandy had become a dreaded post-tropical nor’easter as it struck Atlantic City and then its heaviest winds triggered 14-foot surges at New York Harbor. Levees burst and towns were quickly under five feet of water. In its wake, the hurricane left $65.6 billion in losses from damages and business interruptions.

IT infrastructures that safeguarded mountains of critical business data didn’t escape the storm’s effects. At SunGard AS alone, the disaster-recovery team received 342 alerts and 117 disaster declarations. SunGard AS deployed nearly one-third of its staff, five mobile recovery units, nine workgroup facilities, and 1,500 workgroup seats for its customers’ employees. Its Carlstadt, N.J., center served as an impromptu community command center for local law enforcement and medical and first-response teams.

Hurricane Sandy proved to be a major natural catastrophe; the devastation was widespread. Yet, as such events often do, the storm delivered several key lessons and best practices to follow when the next “Sandy” strikes and threatens IT data centers and facilities. This article outlines several of those lessons and best practices.

Disaster Recovery Involves
Three-Tiered Test

First, here is a bit of perspective and some viewpoints. IT disaster recovery represents a three-tiered challenge, and each layer must be addressed to recover data successfully.

The first layer relates to data protec-
tion. Let’s be clear. If your data isn’t at an
off-site, secure location, you really have
no security at all. Yet, simply having your
data protected at a second location doesn’t
in and of itself constitute a disaster-recov-
ery plan. How you decide to get your data
off-site – by tape, disk backup, storage, or
server replication – will depend on how
critical the particular applications are to
your mission and also on what your objec-
tives are for the recovery point and the
recovery time for your data.

Lessons Learned, Best
Practices

These layers served as our framework
and, from them, we recognized lessons
learned from Hurricane Sandy that we
could share at each of the three tiers of the
recovery challenge:
Several insights emerged pertaining
to the challenges our customers faced,
especially those that used tapes and
trucks as their data-protection strategy.
Transporting tapes to a disaster-recovery
provider’s facility for recovery purposes
can prove a challenge in a hurricane that
causes flooding and road closures. As a
result, many of our customers are now
considering moving to disk-based backup
or to real-time data mirroring and repli-
cation as an alternative data-protection
strategy.

Also worth mentioning is that while tape is the least-expensive medium for backing up data, it also requires the lon-gest recovery times because the data is stored in a different format and requires restoration. Recovery-time objectives less than 24 to 48 hours are difficult to achieve during a disaster if tape is your back-up method. For this reason, many customers are now considering changing to managed backup or vaulting services that leverage storage-area network, or SAN, or other replication technologies.

If you remain tape-based, take advantage of best practices in parallel processing to shorten recovery times significantly. Consider a data-recovery provider whose standby operating system offers a service where your O/S copies can be stored inside the provider. This enables the provider to configure your recovery hardware and restore your operating systems while waiting for your tapes to arrive. This can speed tape-based recovery significantly.