Many organizations will create
a disaster recovery plan that involves a separate facility. This is site
resiliency, where you have another physical site
that you could go to if there happens to be a problem. This means you would have
some type of preparation you would do prior
to the disaster to prepare that
particular location. This may also include
synchronizing data between your primary data
center and this backup location. If a disaster occurs
and your primary site is no longer available,
you now have the option to call a disaster, and failover
all of your business processes to this alternate location. This failover may be
something that only occurs for a number of hours. But if this is a
major disaster, you could end up using this
secondary location for weeks or even months at a time. And of course, once
the disaster is over and the primary site
is available again, we can then move everything
back to that primary location by reversing this process. One type of recovery
site is a cold site. This is, effectively,
an empty building. It provides you with power,
but it doesn't really provide much else. There's no data in
this building, which means you would need to
bring all of your data tapes or have some type of backup
available for your cold site. You would have no
people at this site, so you would need to bus people
in or somehow transport people into this facility. So it would really be
your responsibility to provide the hardware,
the applications, the data, and everything needed to get
this site up and running. A hot site is the exact
opposite of a cold site. A hot site is an exact replica
of your existing running data center that has duplicate
equipment, duplicate power systems, and duplicate data. This is often stocked
with hardware. And when you're purchasing
for your primary data center, you would almost automatically
purchase for the hot site as well. There would usually, also, be a
very high-speed data connection between your primary data
center and the hot site, so that you can constantly
back up your data and have a complete
duplicate of everything running at the hot site. Since everything is
already at a hot site and already ready to go,
it's usually a faster process to convert over to the
hot site than if you were trying to get a
cold site up and running. The problem with a
cold site is there is a lot of prep you have
to do to get that site up and running. And with the hot
site, there's a lot of expense associated with that. If you wanted to have
some type of recovery site that was in-between a
cold site and a hot site, you would have a warm site. A warm site might provide
rack space, HVAC, and perhaps, some other infrastructure,
but you would still need to provide the hardware. Or maybe you have hardware
provided at the warm site and you simply need to
bring your applications and your data. The amount of prep
that you would need to do for a warm
site would usually be associated with the amount
of money that would be charged to provide this service. If your organization needs to
get back up and running quickly but can't afford the
costs of maintaining an always-on hot
site, then a warm site might be a better alternative. Instead of moving
all of these systems to a separate
physical location, you may want to consider moving
everything to a cloud recovery site. Most cloud providers have
existing infrastructure and resources available
and instant availability if you need to move applications
or data into these cloud services. This means that you
wouldn't have to maintain a separate physical facility. And you could send
all of your data and configure all of
your systems online. The costs associated with
a cloud site can vary, but they're usually based
on either a flat fee or based on the amount of use. So if you use more
data and more storage, you'll be billed more for the
duration of that recovery. This still requires that
you document a process so you know exactly
what cloud services need to be provisioned. And you need to
understand how you will get the data from
your existing data center into the cloud site.