There is a difference between availability and backup. Providing a workload with high availability means that if a component of the infrastructure fails, adequate measures are in place to ensure that the use of the workload continues with little or no interruption. Availability delivers the most recent version possible of data as quickly as possible. Backup, however, is for when all of those availability features have failed, when those features are the problem, or when user mistakes surpass a threshold of recoverability for the availability features.
Amazon provides excellent availability to workloads running in EC2 but there are situations when it is not enough and that’s where backup comes in. EC2 instances run on Amazon Elastic Block Storage (EBS), and Amazon provides EBS snapshots of EBS volumes. Many customers count on EBS snapshots as their backup solution. An EBS snapshot is a point-in-time copy of the EBS volume. It is a full copy, not a series of pointers dependent on the original volume, as is the case with most traditional on-premises storage systems. When Amazon executes subsequent snapshots of the volume, it stores only the data changed on that volume since the prior snapshot, in a new EBS snapshot. The service links these snapshots together, providing a point-in-time copy. In theory, EBS snapshots are a backup utility, but they don’t deliver what most organizations expect from a backup.
The Challenges of Using EBS Snapshots as Backup
Amazon stores EBS snapshots in AWS’s S3 Infrastructure, but they are not visible to the administrators from their S3 buckets. To restore an EBS snapshot, the administrator selects the snapshot they want to recover through the AWS EC2 console, and the service recreates the volume from that point in time. The restoration needs to occur on a new EBS volume. Once the administrator creates the new volume, they need to attach it to an existing EC2 instance, sign in to the EC2 instance, mount the file system, and then access files.
Because EBS snapshots are block-based, they don’t have any understanding of the underlying file-system, which means that single file recoveries are not available. If the organization needs to recover a single file or a group of files rather than the whole instance, it must follow the above process to access the files it needs. EBS snapshots also can’t be copied to less expensive Glacier or Deep Archive storage. They also can’t restore on top of an existing EBS volume.
Introducing Veeam Backup for AWS
Veeam Backup for AWS is available as a standalone product for AWS customers or as a component of Veeam Backup and Replication. It is AWS-native and addresses the shortcomings of manually managing EBS Snapshots. Veeam Backup for AWS is deployable from the Amazon Marketplace and automates Amazon EBS snapshots for frequent backups and fast restores. The Veeam backups are separate from the EBS snapshot chain and managed by Veeam.
One of the challenges of EBS snapshots by themselves is administrators generally need to perform their execution with scripts. The AWS EC2 console can trigger snapshots but doesn’t provide much automation, and navigation of the interface is cumbersome. The Veeam interface is designed for simplicity but also offers powerful policy-based backup protection.
Under the policies, administrators can set options like which region or regions they want to protect. They can also control what level of resources they want to preserve and which to exclude. Options include all, a subset, or specifically tagged resources. The administrators can also choose whether to use snapshots and can define their retention policies. They can set backup copy settings like where should the solution store the data, how often should the backup copy be made, and how long should it be retained. Finally, they can also define how and under what conditions Veeam notifies them of job status changes.
Another interesting capability within Veeam Backup for AWS is cost estimation, which the administrators also set at the policy level. It enables customers to understand the cost of executing the policy. It provides information on the backup, snapshot, traffic, and transaction costs for each policy. Administrators can export data from the cost estimator to CSV or XML.
Veeam Backup for AWS has three restore modes. It has a restore to original location mode, which enables administrators to restore to an original EBS volume without having to create a new volume and reattach the instance. It keeps all the settings of the original instance. The solution can also restore to a new location, including a new region. Lastly, Veeam Backup for AWS can perform granular file-level restores directly from the S3 bucket.
Existing Veeam Backup and Replication customers also have a fourth option. Veeam Backup and Replication integrates with S3 repositories created by Veeam Backup for AWS, allowing the restoration of any backup of an Amazon EC2 instance to an on-premises infrastructure.
While EBS Snapshots are a powerful and popular means of data protection, when used for long term data preservation they create problems. Customers need to consider tools like Veeam Backup for AWS to make EBS snapshots both easier to manage and more practical for long term data backup and recovery.