In any dynamic control plane architecture targeting connection-oriented networks, the path computation function can be defined as the one that deals with obtaining a route for a given request or, equivalently, with finding appropriate resources to be allocated to new incoming flows within the network, given the knowledge of the network status, its topology and resource availability, and the applicable constraints and policies. The path computation function can be located in a network management system or in one or more network elements. In the scope of a distributed control plane such as multiprotocol label switching (MPLS) and generalized MPLS, such a function is commonly located either in a process running on each node controller (referred to as source routing) or deployed using a dedicated entity. To this end, the Internet Engineering Task Force defines a path computation element (PCE) as a control plane functional component which is able to perform constrained path computation on a graph representing a network. This paper provides an overview of PCEs and elaborates on selected aspects regarding the deployment and use of PCEs as a functional element in the framework of control and management of optical networks, highlighting the motivations behind their adoption, common deployment models, and related testbed experimentation. The paper also briefly explores new trends such as the extensions for the so-called stateful PCEs and the role of PCEs in software defined networks.
© 2013 Optical Society of AmericaFull Article | PDF Article