Addressing Disaster Recovery Challenges of the Hybrid Datacenter
In discussing the applicability of Disaster Recovery as a Service (DRaaS) to their requirements, we’ve seen firsthand that DRaaS can be simultaneously appealing and frustrating for enterprise customers with complex heterogeneous computing environments. They’ve discovered two primary challenges to IT service continuity in the hybrid model: avoiding DR solution fragmentation and assuring network resiliency.
Avoiding DRaaS Solution Integration/Vendor Fragmentation
Hybrid deployments often involve fragmentation that needs to be addressed creatively from a DR viewpoint. That’s because the recovery solutions for each of the diverse and distributed production resources — public/private cloud-based processing, colocation resources and the customer’s own remaining datacenter assets — must be meshed into a holistic, integrated solution that can deliver an orchestrated enterprise recovery from catastrophic events impacting any or all of the hybrid datacenter components.
We find enterprise prospects and customers looking for ways to implement different DRaaS solutions for discrete platforms and components while, at the same time, trying to avoid contracting with multiple services providers. Implementing a DRaaS solution for VMware with one provider, physical Intel servers with another, iSeries with another and AIX with yet another may deliver benefits to each component but fragmentation to the whole. Customers are becoming aware that vendor management, such as the coordination of multiple managed services-based DRaaS solutions, adds complexity to recovery event orchestration, when what they need during an actual declaration is speed and simplicity.
We see prospects looking for single-source managed availability services delivered across a broad range of hybrid environments. This can best be achieved by selecting a single vendor capable of implementing and managing DRaaS solutions and successfully orchestrating recovery across diverse platforms.
Assuring Network Resiliency
Like any cloud-based offering, DRaaS depends on the design and performance of the recovery network. With their computing resources distributed, rather than centralized, hybrid production deployments add network complexity to a recovery strategy. When the focus shifts from DRaaS-based virtual machine recovery for a limited set of needs to recovery of the enterprise, network issues loom large. Unfortunately, they are sometimes ignored.
For network purposes, hybrid datacenter environments are more like flexible grids than traditional, fixed hub-and-spoke environments. To achieve application and business process resiliency, the recovery network’s design needs to address any combination of losses in the hybrid infrastructure. Everything that “talks together” in production needs to continue to do so if one or more elements are ultimately compromised.
The ultimate goal is a “self-healing” recovery network grid designed to transparently shift network access to DRaaS recovery solutions protecting any component of the hybrid production environment at appropriate performance levels so that any impact to users is minimized. A DraaS recovery solution is only as good as the network it rides on.
Since this is not a capability typically offered by most pure play DRaaS providers, the challenge for enterprise clients is to optimize selected DRaaS solutions by being sure to design and embed a comprehensive network recovery solution to support different scenarios up to, and including, a catastrophic loss of services across multiple elements. This means looking at more than Internet bandwidth. If a real site disaster happens and you have not pre-established corporate network connectivity that mimics your production environment at the DRaaS provider’s site (MPLS, ATM, dedicated facilities etc), your recovery will be orphaned in the cloud. DraaS and recovery network design and implementation are separate disciplines.
A well-designed and properly integrated DRaaS solution can effectively back-up an organization’s entire hybrid IT environment. That’s why IT executives transitioning their enterprises to a hybrid datacenter believe that DRaaS offers the best approach to meeting their recovery time and recovery point objectives. The more complex an enterprise’s hybrid datacenter, the more the success of its DRaaS solution depends on the experience of a vendor in understanding, preparing for and overcoming the recovery challenges inherent in the hybrid model.