- Since the advent of networking, customers have always weighed the cost of throughput vs. the effort of traffic optimization, creating a pendulum effect.
- The market has reached a point where both sophisticated traffic management and performance are required.
In the modern enterprise, the average IT manager has many goals, but a few in particular have been coming up frequently: alignment of the function with business needs (IT acting as a business partner); agile application and solutions deployment; and an infrastructure that will scale and grow with the customer over time.
However, there are also many challenges facing the IT manager. Foremost among these are expense management and user experience (satisfaction, support, etc). In the network domain of IT, specific to the service experience, it has often been a bandwidth vs. management debate: robust, granular QoS vs. “fat pipes.” Throughput has always seemed to increase dramatically as the previous speed reached a congestion point: from 10 Mb to 100 Mb, 1 Gb, 10 Gb, and 40 Gb, with 100 Gb now on the horizon. The prices have fallen rapidly enough that when weighed against the cost and hassle (platforms, labor, training, etc.) of deploying sophisticated edge-to-edge traffic management, the cost of the bandwidth often wins. However, as the virtual age continues to accelerate and applications grow more splintered across several compute nodes simultaneously, it is no longer sufficient simply to “throw bandwidth” at the issue. Virtual desktop technology, simultaneous video streams, and cloud-based storage, to name a few applications, are highly dependent on both high-quality connections and sufficient bandwidth. Based on my research and conversations with end users, I do not believe that either bandwidth or traffic management/optimization will ever be sufficient to meet the expectations of the enterprise user again. Therefore, as enterprises assess and review their next network project or solution, consideration should be given to both the scale of the system (access, box to box, and core) over time and the overall solution’s QoS control and management capability, whether through the switching vendor or via a third party. Without both, enterprise end users will not be satisfied and some projects will be doomed to fail before they even begin.