Redundancy + Agility Doesn't Have to Equal Complexity

Posted 20 October 2015 8:22 PM by Tony Lombardo

(Tony Lombardo is Commvault's Director, Digital Marketing and Technology)

I think we can all agree that 'the web' is an important part of our lives. It’s expected at this point that websites are available 24x7x365. We’re not looking at a percentage uptime, we’re looking at always-on, highly available. You can’t be highly available without also being scalable. Remember the days when websites were 'slashdotted?' I barely even remember slashdot these days, but to put it into perspective for you - imagine your website suddenly became the focus of an article on reddit. Guess what happens to the site you built in anticipation of 10,000 page views/day with 30 concurrent users? Spoiler Alert! 504 Gateway Timeout, 503 Service Unavailable, 500 Internal Server Error…

The cloud thankfully changed the early days of the web, and created a solution for a highly scalable, highly available site. Only, it isn’t quite that simple, as anyone who has gone this route surely knows. The redundancy in a cloud protects you from failure on one of the nodes in that cloud. If you setup your redundant nodes in a separate availability zone, you can get protection from a datacenter outage in a rack. Set up in different geographies and you have redundancy in case an entire datacenter goes down. But there are still ways for your site to drop offline. An administration mistake, a DDOS, a buggy release, etc. So now we need redundancy across clouds. And redundancy across clouds means you need synchronization/replication across clouds as well. You could 1) hire someone to manually do this each day, 2) set up database replication and a separate robocopy script, and maintain the servers at known intervals, or 3) configure the entire backup replication and recovery process through a single console, across multiple clouds – from vm, to database, to storage. I’m lucky enough to be working for a company whose product does option 3.

Here’s a simple prescriptive plan for setting up a redundant site:

  • Deploy your site to 2 separate cloud providers (Azure, AWS, Rackspace, or any other)
  • Set up redundancy across availability groups for each cloud
  • Set up Commvault software to backup databases, files and any VMs on your primary cloud
  • Set up Commvault software to restore databases, files and any VMs to your secondary cloud

Once you have the high level setup down, you can start to tweak and tune. Incremental backups can help you significantly save on compute and network resources. Using live sync features can completely eliminate backup windows. Retention policies can ensure that a failed update that made its way into your backup hasn’t destroyed your ability to recover. But the biggest benefit comes from knowing that whichever direction you move in the future, having Commvault backup and recovery solutions in place will work – across clouds, storage arrays, and even hyper-converged environments.