Using AWS for backups and archiving is very common and an easy entry point for organizations new to the cloud. The main reasons for the popularity cloud-based backup solutions are AWS’ global infrastructure, data durability SLAs, a rich ecosystem of partners and vendors, and compliance with regulations such as HIPAA, PCIDSS, and so on. Taking a phased approach that begins with using cloud storage as a backup data store is a reasonable and common approach. However, taking this further to create your application environment in the cloud can be a good business continuity strategy. It is also easier to track the actual usage of the backup data on the cloud, thereby presenting further opportunities to reduce your overall costs.

In most cases, the enterprise already has a well-established backup strategy, policy, and technology solution in place. They are not looking for a complete replacement. The primary motivation here is to leverage the cloud as a lower cost destination. However, for most early stage start-ups, the cloud may represent their first and only backup solution.

There are third-party solutions such as Commvault that are natively integrated with Amazon S3 and Glacier. Other backup solutions can also be integrated using storage gateways. These approaches can help evolve your backup solution to embrace cloud storage with the least disruption while you continue to use your existing processes. Cloud storage represents "unlimited" capacity so you do not have to worry about closely tracking your tape usage or rotating through them constantly.

It is important to carry out a data classification exercise for all the data in your application. For example, you might classify your data as long-term data to be kept for compliance reasons and unlikely to be restored, very high volume data to be transferred from on-premises storage to the cloud, data shared by multiple applications (document shares, image repositories, and so on), highly available data, data related to online collaboration applications, and so on. This classification can help you choose appropriate solutions for storage and backup. AWS provides different storage classes, that is, S3, Glacier, and EBS, to meet varied data life cycle management requirements. All these data storage services are scalable, secure, and reasonably priced.

It is also important to define certain guiding principles for your backups. For example, you could choose to back up only the data and not the entire VM. This would mean you are choosing to rebuild (using a service such as AWS CloudFormation) instead of restore. Other guidelines may recommend building stateless services and storing all data on Amazon S3 and leveraging services such as SQS, assuming that all instances are temporary or fail sooner or later. You might also want to take snapshots of your EBS volume on other EBS volumes to recover faster from instance failures. In the DevOps environment, owners of applications have increasing responsibilities in terms of their data.

While evaluating the cost of a cloud-based backup strategy, ensure you do not restrict your TCO calculations to hardware and software alone. For a good comparison, ensure you take into consideration costs associated with facilities, maintenance, people and professional services, storage utilization, transportation, cost of capital, and so on. In larger backup environments, the cost of a cloud-based solution compares very favorably versus the total cost of physical media, robot systems, and other costs mentioned earlier.