As a SaaS offering built on and for the public cloud, Cloud Expo is the natural home for Druva. It’s exciting to see other companies join the fray, as it validates the approach we decided to take several years ago. We look forward to presenting our solution to attendees of the show.
Several years ago, our founder Jaspreet Singh saw the tide was heading towards cloud, specifically the public cloud. At the time Druva was another on-premises backup and recovery offering. It was a solid product, but it was not that much different than many other products in the space. That all changed when we pivoted toward the cloud.
Other vendors wanting to move toward the cloud simply made sure their product could run well in the cloud. They certified what type of VM and storage different configurations of their product needed, and then began helping customers “lift and shift” their product into the cloud. We wanted to do something different – re architect the product from the ground up to work within a cloud environment.
Jaspreet knew, for example, that backup workloads are very transient in nature. Backup loads are very high at night, of course, with a few other jobs running during the day time. What if Druva used Amazon’s automatic scaling system to increase and decrease compute resources to match this elastic demand?
He also looked at how all the other vendors were using EBS storage, and felt that it was too expensive. He wondered if there was a way to use the less expensive S3. Other backup vendors using S3 tended to use it as a dumping ground, placing large backup images on it when they were no longer needed for active recoveries. Instead, Druva opted for a design where every backup of every file or image was split up into tiny chunks and stored individually as objects. This works hand in hand with global deduplication, and it makes S3 much faster from a recovery perspective. During a restore, Druva can request tens of thousands of objects, rather than a single large blob. Using S3 the way it was meant to be used made it the perfect place to store backup images.
Finally, we needed databases to store data and metadata, and we again opted for the AWS services of DynamoDB and RDS. DynamoDB stores our file and deduplication metadata, and RDS stores all the configuration data. The scalability of DynamoDB is why Druva is the only vendor able to do global deduplication across an entire customer’s environment, no matter how large.
As the reader can see, Druva uses the cloud the way as it was meant to be used. Treat the cloud like an operating system and use the various resources it offers, and your system design will thank you for it. We look forward to getting a chance to tell our story at Cloud Expo this week.
Want to harness the power of the cloud? Learn how Druva can hep.