ABSTRACT
Review Questions and Answers . . . . . . . . . . . . . . . . . . . . . . 455
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 459
While advances in wire-line and wireless transmission systems have provided ample bandwidth surpassing customer demand at least for the near future, the bottleneck for high-speed networking and enhanced service provisioning has moved to processing. Network system vendors try to push processing at the network edges employing various techniques. Nevertheless, the networking functionality is always proliferating as more and more intelligence (such as multimedia content delivery, security applications and quality of service (QoS)- aware networks) is demanded. The future Internet is expected to provide a data-centric networking platform providing services beyond today’s expectations for shared workspaces, distributed data storage, cloud and gridcomputing, broadcasting and multi-party real-time media-rich communications and many types of e-services such as sophisticated machine-machine interaction between robots, e-health, and interactive e-learning. Thus, the model of routing/switching devices has been augmented to enable the introduction of value added services involving complex network processing over multiple protocol stacks and the raw data forwarding functionality has been left only as the major task of large core switches that are exclusively assigned with this task. To cope with this demand, system designers have leaned on microelectronic technology to embed network processing functions in either fixed or programmable hardware as much as possible. This led to a new genera-
tion of multi-core embedded systems specifically designed to tackle network processing application requirements.
