Page 167 - How Cloud Computing Is Transforming Business and Why You Cant Afford to Be Left Behind

D ANGERS ABOUND: SECURITY IN THE C L O UD

hardware failure, your data is transferred to the second server
and processing picks up at the point it left off.
Amazon Web Services would probably advise you to place
this failover AMI in a different zone from the original. Its data
centers are divided into different “availability zones,” each with
independent power supplies and other resources, so that a

failure in one zone doesn’t take down the whole data center.
Implementing this cross-zone model, however, imposes a new
fee that is applied to the transfer of data between zones. You
may decide that doubling what you’re paying to Amazon Web
Services for a second virtual machine is enough. You don’t
want to incur more secondary costs. You leave your backup
machine in the same zone as the original, as there’s no fee to

transfer data within the zone.
If you’ve made such a decision, consider this incident on
December 9, 2009. A component of the power supply for a
zone of Amazon’s data center in the US-East-1 region, in this
case, northern Virginia, failed, and Amazon eventually alerted
customers to this event. Several zones make up a region, so it
would be hard to tell from this information where the data
center was or whether your virtual machine was running in it.
If you did not employ either Amazon’s CloudWatch service,

which monitors your virtual machine and indicates when it
stops running, or a third-party service, such as VMware Hy-
peric’s CloudStatus service, you probably wouldn’t know that
your virtual machine had stopped running until it failed to de-
liver the expected results.
If you heard through the grapevine that an Amazon EC2
zone was down, you could go to the online Amazon Service

147