Don’t Forget About the Network in your UC Plans

Since the very first VOIP deployment, IT engineers and architects have struggled with performance management. Whereas “throw bandwidth at the problem” was a viable solution when the only sensitive application was voice, the introduction of video as part of the UC equation, coupled with increases in other latency sensitive traffic means that a successful UC implementation requires paying close attention to network performance metrics including latency and jitter. And now simple “QoS” is evolving into a new term – Application Delivery Optimization(ADO), which not only includes prioritization of traffic, but also bundles tools such as caching, flow control, compression, and rate shaping to provide network managers with a comprehensive suite of tools at their disposal for managing the entire spectrum of applications on the network. Engineering networks for application delivery is challenging because the target moves continually. Where a few years ago primary focus might have been on moving files around, or optimizing delivery of Outlook/Exchange traffic, today’s enterprise is moving rapidly into newer and less-well-understood waters. New applications are coming on quickly, many needing lots of bandwidth and real-time performance. They are highly visible to users (there’s nothing more visible than the quality of rolling video, for example) and because the apps are critical to new modes of collaboration and customer service, they are not optional. New architectures and applications that dramatically change performance requirements include Voice Over IP (VoIP), voice and video conferencing, and other collaboration tools; data-center consolidation; Software as a Service (SaaS); and desktop virtualization. Moreover, organizations need not just capacity and conditioning but control: the ability to ensure capacity is going to applications according to their priority. They must prioritize traffic in fairly granular ways, especially with media sharing a continuing concern and recreational use of bandwidth-intensive social media and other sites still on the rise. Video conferences should not suffer because there is a new hot video on YouTube. Access to critical SaaS applications must take priority over basic Web browsing or even over other, less-critical SaaS services. And there’s a cost component. With 21% of organizations paying IT for WAN bandwidth, at an average of $60/Kbps—the business lines have a strong incentive to demand the best performance for their money. They will press for meaningful performance guarantees in their SLAs, and so IT staffs will have ever-stronger incentive not only to deliver the performance but to be able to verify they are doing so for any and all applications (i.e. that the network is not the problem, if there is one). To deliver new services to users in an evolving branch and telework environment, IT may need to compress traffic to conserve bandwidth, but must condition network traffic to guarantee acceptable performance. The organization must also control traffic to ensure capacity goes to applications according to their organizational priority, and not to recreational or social media sites. As business lines increasingly have to pay IT for their bandwidth, IT having visibility into traffic, control over it, and the ability to condition traffic will be essential not just to meeting user expectations but also business line SLAs. Now is the time for organizations to take a close look at their application optimization systems and plans. They should: -Evaluate new application requirements and performance to gauge the suitability of current WAN and optimization. -Be prepared to upgrade, replace, or supplement current solutions for new use cases or traffic types. -Expect an increase in real-time and fault-intolerant traffic, if not an increase in bandwidth. -Optimize for the shift in traffic flows as a result of IT centralization and data center consolidation. -Optimize for the shift of traffic associated with desktop virtualization. -Control, monitor, secure, and optimize traffic with SaaS providers and other direct-to-Internet branch traffic. -Plan on providing optimization and security to progressively more numerous and smaller endpoints: big branches become mini branches become teleworkers. Cheap hardware or robust, affordable soft clients are key.