From the Blogosphere
Disaster Recovery Ascends to the Cloud | Part 2: Deployment Considerations
Realizing an economical alternative to traditional DR
By: Nicos Vekiarides
Oct. 27, 2012 09:00 AM
As mentioned in Part I of this series, cloud technology has introduced a viable alternative to the practice of creating secondary sites for disaster recovery (DR), promising to save IT organizations hundreds of thousands or even millions of dollars in infrastructure and maintenance. While the cost reduction associated with replacing dedicated DR infrastructure is intuitive, the ability of cloud solutions to meet the recovery times (RTOs and RPOs) dictated by businesses is often less well understood.
Part I suggested two key considerations in recovering IT operations from a disaster are (1) regaining access to data and (2) regaining access to applications. Today’s cloud integrated storage or cloud storage gateways can push backups or live data sets to the cloud easily and securely, enabling the first element of a cloud DR solution. With this in mind, let’s examine two strategies for application recovery using cloud-based DR:
Strategy 1: Data copies in-cloud, application recovery off-cloud
The advantage of this approach is the elimination of dedicated secondary storage infrastructure for DR. The disadvantage is the requirement for a secondary site for application recovery.
An improvement to this approach involves leveraging a hosting provider as the application recovery site, where new application servers can be provisioned on-demand in case of a disaster. Using a hosted recovery site can be considerably faster than restoring and rebuilding the original application environment and more economical than maintaining a dedicated secondary site. However, recovery times may be impacted by the time it takes for the hosting provider to provision new servers.
Strategy 2: Data copies in-cloud, application recovery in-cloud
When using a cloud storage gateway to replicate data to the cloud, consider cloud gateways with the ability to run in the cloud. Cloud servers can then attach to the gateway to facilitate application recovery.
The process of application recovery may involve activating servers and applications via a cloud provider’s catalog. Although this process is much faster than provisioning new physical hardware, it can still be time consuming, particularly when attempting to recover tens or hundreds of servers.
Alternatively, virtual machines that resided on-premise can be reinstantiated in the cloud, similar to failover of virtual machines between hypervisors. This is possible if the same hypervisor runs on-premise and in the cloud. However, while moving virtual machine (VM) images between like hypervisors is generally straightforward, many cloud providers may not offer sufficient administrative privilege in their virtual compute environments or may not be compatible with on-premise hypervisors.
To get around these limitations and incompatibilities, an emerging option involves importing on-premise VMs into the cloud via conversion scripts and tools. An important consideration is ensuring that these conversion scripts and tools operate bidirectionally, meaning they allow a way to eventually export VMs back to the on-premise environment.
The keys to success are testing and working with a partner you trust
Keep in mind that an important aspect of any DR strategy is conducting regular testing and validation. Additionally, working with technology partners who understand the advantages and tradeoffs of DR in the cloud can be particularly helpful.
Like any major IT undertaking, DR in the cloud requires significant planning — but the payoff can be substantial if reducing disaster recovery costs and improving availability are important to your business.
November 12-13 Faculty Highlight
Digital Transformation Blogs
CloudEXPO TV Power Panels