Highly Resilient Network Elements

Abstract

Current network traffic comprises of operations-critical data from diverse applications like SAP, multimedia over IP, collaboration, and storage. Current network interface speeds are at multiple Gigabits per second. For networks operating at this speed, operation resiliency is very important Disturbances as little as a few seconds can cause significant packet drops, which can lead to lost time and productivity for users. Network operators and administrators are constantly challenged with ensuring high availability of the network, and in the event of a failure, minimizing its impact and restoring services quickly. This is achieved by network planning/design and by using hardened and resilient networking elements. The key principles of resiliency are the (1) elimination of single points of failures, (2) early detection and failover around defective parts, (3) localization of failures and containment of their impact, and (4) restoration of failed components without impacting other subsystems. This paper addresses the architecture and design principles in building next-generation highly resilient networking equipment.