Page 166 - How Cloud Computing Is Transforming Business and Why You Cant Afford to Be Left Behind

MANA GEMENT STRATEGIES F O R THE CL OUD R EV OL UTION

virtual machine. New virtual machines are constantly pouring
into the data center from all directions. Do any of them har-
bor spying agents, password stealers, or other malware? Might
they end up on the physical server that you’re using? Is there a
risk of hardware failure if they do?

Maintaining Clean Operations and
Protecting Yourself from Cloud Failure

No one knows the risk of hardware failure better than Ama-
zon Web Services, supplier of the leading cloud service, Elastic
Compute Cloud (EC2). During its first two years of beta oper-

ation, EC2 did not offer service-level agreements or guaran-
tee continuous operation. During that period, servers froze
up, virtual machines died, and workloads disappeared, but
any complaints that reached Amazon Web Services were met
with a stony recommendation that you architect your software
to cope with hardware failures. Most IT pros are accustomed
to doing the opposite: they architect the data center to avoid
hardware failures.
Coping with failure in the cloud means giving your ap-

plication the capability to failover to another server. The re-
dundancy is contained and managed in the software, not the
hardware. That’s one of the major differences between oper-
ating in the cloud and operating in the traditional data cen-
ter. One of the easiest ways to do this is to direct a failover to
another nearby Amazon Machine Image (AMI), a virtual server
with the same configuration as your original. In the event of a

146