Continuing from yesterday’s post about disaster recovery (DR), today I want to highlight some of the types of DR services that are available and some of the benefits and drawbacks of each service. So without further ado, Allons-y!
The first DR service, and one of the more common entry-level types of service is replication to VM. This is a cloud-resident service that can be used for cloud-VM-to-cloud-VM or on-site-to-cloud-VM data protection. This type of service is based on continuous data protection solutions such as EMC Atmos, and is best suited for applications that have strict recovery time and recovery requirements, as well as application awareness as it can be used to protect both on-site and cloud production instances.
The second main style of DR offering is in back-up and restore to the cloud. This utilizes both cloud storage and cloud computing resources to allow organizations to restore data to virtual machines which are located in the cloud and run the environment remotely until the on-premise infrastructure is restored. It can also be used for pre-staging where the restores are currently updated on a scheduled basis to ensure that the recovery time objectives are met. In some cases, the enabling of the cloud infrastructure is included as part of the DR plan.
Similar to this offering is backup to and restore from the cloud style services. As you can imagine, this service uses the cloud to replicate data on VMs which are restored to on-premise infrastructure in the case of a disaster. This is an evolution of tape backup services which are common IT practices. The key to ensuring this service meets your objective needs is to make sure that your IT team understands the backup and restore processes to avoid any potential issue with the restoration of services. You should also look for services which offer compressed or data dedupe and encrypted movement of data alongside customized retention options to ensure the service works with your network and security requirements. Bandwidth will be a large concern during the restoration stage, and if you are pushing terabytes of data through an unoptimized pipe, you run the risk of meeting your recovery time objectives. Some cloud providers offer a service whereby they courier over the data in disk format, or customers may opt to keep business-critical resource backups in an on-site cache.
The last major DR solution is the fully-managed DR model, an increasingly popular option for organizations who don’t want to have to deal with the risks and processes associated with DR. In this model, the customer uses a managed service provider to manage both the production and the backup instances. The benefit of this type of model is that the customer leverages the cost savings associated with a cloud infrastructure and the usage-based costing of the DR solution. The service provider handles everything from beginning to end, including ensuring the recovery time objectives are met. The risk in this case is pretty obvious, the service provider is responsible for everything and so it is absolutely important that the SLA is reviewed very carefully and critical business assets identified with the appropriate recovery time objectives assigned. If not, should a disaster occur, the customer will be the one at risk while the provider points to the SLA and says “It wasn’t part of the SLA”. The SLA will determine whether you have access to your applications and resources should an emergency occur.
There are clearly many different options for DR that leverage the cloud. The cost savings and technological benefits can be huge for SMBs who do not have the in-house expertise to provide DR services. The key is to ensure whatever plan you decide to implement is tested routinely and provide the correct training to all users in advance to avoid any complications in the event of a disaster.