Disaster Recovery to Microsoft Azure using Arcserve UDP 6.5
The featured video below is a demonstration of a low-cost Disaster Recovery solution that leverages Microsoft Azure as a Disaster Recovery platform and Arcserve UDP 6.5 Update 2 as the Disaster Recovery / Backup Solution.
During this demo, I am using Amazon EC2 as my “on-premise” production environment. However, keep in mind that this solution could work just as well for physical or virtual servers located in a corporate datacenter or cloud.
This particular demo is utilising two recovery point servers, one “on-premise” in Amazon EC2 and the 2nd hosted in Microsoft Azure. The Recovery Point Server hosted in Azure is an optional component that gives you the benefit of off-site backup retention eliminating the need to store backups offsite by other means, such as using a Tape Vaulting company. The Azure-based RPS also reduces the amount of data sent over the network because only globally deduplicated data is transmitted over the network. Global Deduplication can drastically minimise data transfer and storage. Having an RPS local to the Recovery Environment can also speed up the recovery process, and is essential for Linux Recoveries using “Instant VM Recovery to Azure.”
The deployment topology I have used in this demonstration is that of two Arcserve UDP Consoles, each with the Recovery Point Server role enabled. This topology is a simple way to configure backups to replicate to an isolated network such as a remote cloud that does not have visibility of the on-premise subnets. I often refer to this type of deployment as an “MSP Style” deployment as it is commonly used by Managed Services Providers to store multiple clients backups on a single shared Recovery Point Server.
The Backup Plan on the Amazon AWS hosted Arcserve Console has two simple tasks. Backup Windows Servers using the Arcserve UDP Agent, then replicate those changes to a “Remotely Managed Recovery Point Server.”
The Azure-based Arcserve Console has a reciprocal “Backup Plan.” The primary task of this plan is to “Receive Data from a Remote Recovery Point Server”. The secondary task is a Virtual Standby Task – to Microsoft Azure. When the Virtual Standby Task runs for the first time, a full restore of the backup image completes. The restore process creates virtual disk files (.vhd) stored in Azure Blob storage. Subsequent Virtual Standby Tasks will complete incremental restores in a series of snapshots. The Recovered VM can be powered on in Microsoft Azure using any of the available recovery points.
An incremental backup of the Amazon EC2 instance takes just a few seconds, and the subsequent replication to Azure takes a matter of minutes. In the demo, the actual data backed up is reduced by 93.5%. The replication to Azure completes in 3.5 minutes with a 94% data transmission reduction.
I hope you enjoy the demo of this low cost, elegant Disaster Recovery Solution leveraging Microsoft Azure. If you know anyone that might be interested in this article, please do pass on the link. If you have any questions or comments, then please do comment below or get in touch with us directly at firstname.lastname@example.org.