L. Carrillo, J.L. Marzo (Spain), D. Harle (UK), and P. Vilà (Spain)
Scalability, reliability, performance analysis, routing,distributed systems, adaptive routing, AntNet.
One of the key problems facing designers of Network Management Systems is that of scalability. Due to the special characteristics associated with network management, the scalability issues of such systems are different to those of traditional client/server systems. A network management system must operate continuously with high reliability (in the order of 99.999%) and within a variety of strict timescales. At the limits, responding to time critical events such as fault restoration, responses in the order of sub milliseconds are required. This paper discusses the different issues that affect scalability in distributed systems and then presents a formulation to calculate the scalability of a particular class of distributed systems know as AntNet (routing in communications networks based on ant colony systems) that can be applied in a network management role. Additionally, the paper presents the relative performance of AntNet when applied to a number of network topologies, which show that AntNet routing, based upon overall performance and resource consumption, is clearly scalable.
Important Links:
Go Back
