1 Click Cloud Disaster Recovery Orchestration: How It Works


Hi, this is Mike with Actifio. Some of you may have seen another video introducing our cloud dr orchestration where we can orchestrate recoveries from an on premises data center into the cloud. This video is going to go a little bit deeper into how it works because some of you may have seen the video and had more questions about that. So starting out with Actifio has had for a couple of releases now the ability to take physical servers and virtual machines that are backed up into an Actifio Sky and recover them into cloud instances. And there’s really a couple of features about this that make it special.

The first one is that the Actifio Sky running that recovery into the cloud instance has the option of whether or not it should copy all of the data for that VM into the block storage for that new cloud instance, or if it should run from the mount persistently presented from the Actifio Sky Just like Actifio sky can do instant mounts on premise back to the original servers, all of the data can run over that same style of connection all within the cloud. And in doing this, we can provide very short recovery times, very short RTOs. Only the boot volume really needs to be copied from the Actifio Sky storage onto that block storage.

The second thing that this feature enables is it enables the recovery not just from block storage that’s allocated to an Actifio Sky that maybe was being replicated from the source directly onto the block storage, it also works when that sources object storage. So that regardless of block or object on the backup copy, you can still do that same short RTO recoveries directly into the cloud. Now with that said, if you had to specify for each VM over and over again which cloud, where should it go, what network should it go into, what IP address should it have? That becomes very tedious and that’s where the Actifio resiliency orchestration comes in.

The other place that it comes in is in performance and load balancing. So what resiliency director allows you to do is to spin up as many Sky appliances as you want in the cloud where you’re running the disaster recovery at the time that you need them and through an intelligent load balancing algorithm will choose the best appliance to run each new recovery. And it makes sense. Each individual Sky appliance in the cloud can only handle so much throughput, so much workload, so much IO. And so it will be limited on how many servers or VMs it can recover per hour for example. But this level of horizontal scaling allows you to have effectively any recovery time that you want because we can scale out to two or 10 or 50 or a hundred different Sky appliances.

And if those recoveries are copying all of the data onto persistent block storage, for example, on a permanent recovery as opposed to just a test at the end of it, these appliances are no longer needed. You can actually take them down and no longer consume any costs, any CPU or memory with them. So this level of orchestration and scalability provides a lot of benefit for the customers. For you.

Everything gets pre-configured. Of course, in Resiliency Director, you will predetermine the cloud destination. Supported destinations include GCP and AWS, with the first release of Resiliency Director for the cloud in 10c. We’ll be adding other clouds, including Azure pretty quickly on the heels of that. You predefine the cloud as the destination. You predefine the networks, you predefine the IP addresses everything that you need very simply and efficiently in Resiliency Director, so that at the time you run it, there’s no decisions to be made except do you want to run it. And then when you execute, you get the results you’re looking for.

Hopefully that helps answer a few extra questions that have been coming up. See you next time.

Read more about cloud disaster recovery from an Actifio Customer.