Implementation can generally be divided into two categories: do-it-yourself and recovery as a service.
Both architectures can be deployed using virtual, physical, or a combination of both at the production and recovery sites. Virtualization can be used in mirrored site architecture, but DR VMs must be constantly running and replicating.
Do-it-yourself recovery models typically take advantage of service provider infrastructure and off-the-shelf solutions, but involve designing and implementing DR sites based on the specific needs of the organization and the DR plan itself. Examples including architectures that leverage DR tools, like VMware site recovery manager and managed infrastructure like Internap’s custom hosting, AgileCLOUD, or AgileSERVERs in this category as well.
Recovery-as-a-Service (RaaS) models are prepackaged, standard DR solution services that can be bought on a subscription or pay-per-use basis. Historically, these solutions have taken the form of turnkey physical infrastructure. This has rapidly evolved into virtualized offerings given the ease of “containerizing” applications, data, and configurations on virtual machines. Virtualized RaaS vendors deploy agents that replicate data, applications, and system configurations so that they can be accessed when a disaster is declared. Persistent System’s rCloud is a good example of a proven RaaS model.
Implementation options for different failover site types
Mirrored and Hot sites are typically implemented using a do-it-yourself architecture, given that most RaaS services’ RTOs can’t practically be guaranteed at less than 30 minutes. Recovery for these site types might be best applicable to scenarios where a production site becomes unavailable, but immediate downtime is avoided because of automated site recovery capability.
Warm and Cold site implementations are most applicable when a partial production site failure takes place, and a manual intervention is necessary to initiate recovery. In these cases, RaaS may be a good complement to recovery efforts as part of an overall DR plan.
In our video, we’ll focus on a few DIY hot and warm site architectures, including:
- Dual site – Staging and production environments are separate.
- Hybrid – Colocation and/or managed hosting are used, with the ability to burst into the cloud with crossover connections.
- Content Delivery Nerwork (CDN) – Content is placed in the network, so that it is still available outside of the primary data center.
- Tokenization – For secure environments that require compliance, tokenization allows personally identifiable information to be stored in a location other than the data center.
Learn more about Disaster Recovery planning for IT organizations in our video series: