Many organizations find the idea of remote-only failover operations
very appealing. They view this solution as a way to cut costs while
still offering the ability to recover data systems to another
location.

However, failing over to (or replicating or
backing up data to) only a remote location isn’t sufficient as the
sole component of a DR plan. Let’s look at some of the reasons why
this is a mistake.

Potential single-server failure

Many companies fail to consider this
possibility. If only one data system fails, odds are good that the
remainder of your data center is probably still able to function
adequately–if not exactly the same as it did prior to the
failure.

The end result is that you may create a
situation where you fail over data systems to another location when
it’s not really necessary. In addition, this can cause extra work
for administrators and more downtime for users.

Impact on end users

Failing over to a remote location could change
your networking needs and how your end users do their jobs. For
example, having to suddenly access data systems across a wide area
network (WAN) could cause delays and frustration to your end-user
community.

If you experience a data-center-wide emergency,
it’s very likely that your end users will need to work from
alternate locations, affecting productivity across the board and
leaving them to deal with potential difficulties with remote
access. However, if nothing else is wrong with the environment,
you’ll likely find them far less tolerant.

Impact on data systems

If a data system fails and other systems rely
on immediate connections to the failed system, you’ll face one of
two choices: Either manually fail over the other systems, even
though they’re working perfectly fine, or else resign yourself to
dealing with the same kind of problems as your end users.

Regardless of which option you choose, it
basically means extra work. In addition, there’s the potential that
mistakes and failures will compound the original emergency.

No automatic failover

Many DR systems that involve replication or
snapshot data capture allow you to either fail over automatically
or on command. However, it’s inadvisable to fail over automatically
across a WAN, where a connection break can mimic a downed server
situation from the perspective of the DR server. That means you
lose the option to automatically fail over if you’ve implemented no
local failover option.

Implementing redundant failover systems at both
the local and remote sites can eliminate almost all of these issues
completely. And proper configuration of the systems will eliminate
the rest.

While these solutions cost more than
remote-only solutions, you can recoup those costs in the long run
in terms of avoided downtime and fewer accidental failovers.
Avoiding up-front costs and setup considerations often results in
more pain and lost profits down the road–a trade-off that isn’t
beneficial technically or fiscally.

How well can your organization deal with an emergency? Automatically sign up for our free Disaster Recovery newsletter, delivered each Tuesday, and make sure you’re prepared for the next catastrophe.