Cloud computing remains somewhat suspect when it comes to security,
and not without reason. Many noteworthy cloud services, like
Google's Gmail, are only a password away from being pilfered,
though enhanced security options are available.
Any time a business hands its data over to a third-party, some
wariness is warranted.
But cloud security in some areas goes beyond what's available using
on-premises IT. Or at least it's a better deal.
Take Google Apps, which includes Gmail, Google Docs, Google
Calendar, Google Groups, Google Sites, and Google Video, for $50
per user annually. Among its less-heralded features is automatic
backup and disaster recovery, at no extra cost.
In a blog post on Thursday, Rajen Sheth, senior product manager for
Google Apps, explains that Google Apps customers don't need to
worry about backups or disaster recovery.
Disaster recovery, he explains, is usually measured in terms of RPO
(Recovery Point Objective) and RTO (Recovery Time Objective). RPO
represents the amount of acceptable data loss in the event of an
outage -- the gap between backups -- and RTO represents the
acceptable amount of downtime before service is restored.
For large companies running Storage Area Networks (SANs), RPO and
RTO targets are often an hour or less, explains Sheth, and that
kind of disaster response usually costs a lot.
"For Google Apps customers, our RPO design target is zero, and our
RTO design target is instant failover," he says in the blog post.
"We do this through live or synchronous replication: every action
you take in Gmail is simultaneously replicated in two data centers
at once, so that if one data center fails, we nearly instantly
transfer your data over to the other one that's also been
reflecting your actions. Our goal is not to lose any data when it's
transferred from one data center to another, and to transfer your
data so quickly that you don't even know a data center experiences
an interruption."
Sheth says that Google's live replication was one of the reasons
that the City of Los Angeles decided to move from on-premises
e-mail to Google Apps.
Sheth acknowledges that no system is perfect. But imperfection is
part of Google's plan.
"We design for failure," said Sheth in an e-mail. "We operate on
such a large scale that there are always servers and server racks
that need to be maintained and upgraded. Our infrastructure is
designed to give us the flexibility to take action when we need to
without impacting users. For example, most corporations take down
their servers for a set of time to conduct planned maintenance. We
just switch people over to a new data center, and they never
notice."