Don’t Let CPU Power Prevent You Delivering QoS to All Your Customers

Quality of Service (QoS) used to be a simple matter. In today’s IP-enabled world with a profusion of applications and vastly differing bandwidth requirements, it’s become a lot more complicated.

This makes life difficult for network operators – and the fact that demand is surging complicates matters further. Network operators are faced with the twin challenge of both massively increased demand and increasing application requirements. Delivering QoS to all users fairly and equally is tough.

Until recently, they have relied upon packet inspection solutions based on CPU power to ensure that problems can be identified and resolved. But keeping pace with demand makes this increasingly expensive. With core packet transport moving to 10Gb/s, 40Gb/s and beyond, the load on CPU-based systems will grow. This will dramatically increase the cost of such systems, increasing both CAPEX and OPEX and reducing profitability.

A better approach is to offload the pre-processing task to dedicated hardware resources. This reduces the load on the CPU and ensures that the system can scale effectively to cope with increased bandwidth in the future. There can be significant cost benefits – we estimate by as much as 50% at transmission rates of 40Gb/s. What’s more, the adoption of hardware acceleration solutions can ensure the QoS delivery across an entire network.

This is essential, as QoS is the key to customer satisfaction. If customers can’t access the services they want, or if they don’t perform as expected, they will migrate to networks that offer better performance. QoS is going to become a differentiator for network operators and increasingly selective users will search for solutions that deliver what they need.