Disaster Recovery in an IP WorldDisaster Recovery in an IP World

UC offers a more resilient and flexible infrastructure, but requires a different approach to recovery.

Migration from traditional TDM PBXs to IP-based unified communications platforms represents a significant shift in the communications network architecture. As organizations embrace this next generation of communications platforms, however, a reassessment of existing disaster recovery and business continuity (DR/BC) plans must also be undertaken. Compared to legacy solutions, the technologies in a modern UC environment offer a more resilient and flexible infrastructure, but require a different approach to recovery.

My colleague Michael Brandenburg has recently published a market insight on ensuring DR/BC in an always-on, dynamic UC infrastructure. In DR/BC terms, there are two key metrics by which all decisions are driven: recovery time objective (RTO) and recovery point objective (RPO). RTO represents the amount of time needed to recover to a reasonable level of functionality; RPO defines the time period between the last secure storage of data, such as voice mail archives, chat logs or call recordings, and the moment of the disruptive incident.

The current technologies driving enterprise communications, including VoIP services, server virtualization, and hosted or cloud-based solutions, can directly lower the time it takes to recover from a major event, as well as minimize the data loss as a result of the event. For instance, SIP trunking offers a number of DR/BC capabilities that are not possible or are cost-prohibitive on legacy circuits, since SIP trunk sessions are carried over a business data network rather than over the dedicated physical connections that analog or PRI circuits are delivered on. Voice calls are delivered around points of failure, whether on the network or the customer premises, even in case of a metropolitan or regional outage. The SIP trunking services delivered by most service providers offer multiple tiers of failover options to match the particular needs of businesses of any size and scale.

Likewise, within a virtualized server environment, the stack of UC applications can now run on a shared pool of generic compute, memory and storage resources while overseen by a hypervisor, which monitors and allocates resources to each virtual machine running in the environment. In the event that a particular resource, such as a single processor blade in a server, experiences a failure, the hypervisor will move the affected virtual machines to another resource within the environment, effectively eliminating the single point of failure found in hardware-based platforms. Similarly, because the UC applications are now simply virtual machines within this shared environment, multiple instances of the same application can run within the same environment, allowing for redundancy and high availability even if a software instance crashes.

Finally, Hosted UC solutions put the onus of uptime assurance and rapid disaster response on the service provider rather than on the IT department. When architected properly, hosted solutions provide failover and geographic diversity on a scale that only the largest of enterprise customers could even begin to match. Because many hosted UC solutions are accessible either across carrier networks or the public Internet, affected users remain connected from remote and small branch offices, home offices, and in many cases, from their mobile devices. By their very nature, hosted UC solutions are highly resilient to localized or even regional disasters.

But IT managers need to lay the groundwork to make such benefits possible, in all three situations mentioned above. Michael has outlined several best practices and customer scenarios to make the most of the DR/BC capabilities inherent in hosted, virtualized and IP-based communications. Clients can download the full report at www.frost.com.

Follow Melanie Turek on Twitter and Google+!
@melanieturek
Melanie Turek on Google+