Iowa State Optical Simulator (ISTOS): Design, Architectures and Features
Ong, Yana (2005) Iowa State Optical Simulator (ISTOS): Design, Architectures and Features. Masters thesis, Iowa State University.
Full text available as:
Wavelength division multiplexed (WDM) networks have emerged as the viable solution to meet the increasing demands of bandwidth due to tremendous growth in Internet and World Wide Web (WWW) usage. WDM networks divide an optical fiber bandwidth into multiple WDM channels called wavelengths. Current fiber technology allows a transmission capacity of up to 40 Gbps on a single wavelength while end-user requirement ranges from 45 to 622 Mbps typically. WDM grooming networks support this sub-wavelength bandwidth requirement by allowing multiple connections to share single wavelength. The minimum connection granularity on WDM grooming networks is called a channel. Many different routing and channel assignment schemes and network survivability mechanisms in WDM grooming networks have been proposed in research literature. The routing and channel assignment algorithms select the path and the channels on each link of the path a connection request is routed. The network survivability algorithm identifies the strategy used to protect the network against single link or single SRLG failure. A common platform to compare these schemes, against the same set of traffic, is desirable. In this thesis, we present a software tool that simulates routing and channel assignment as well as network survivability algorithms in WDM grooming networks, and serves as a framework for comparing multiple algorithms run under one simulation environment. The Iowa STate Optical Simulator (ISTOS) tool allows simultaneous simulation of multiple networks with the same topology but different routing and channel assignment and/or network protection and recovery schemes. The networks are simulated with common simulation parameters that determine the network traffic rate, the failure rate (if applicable), the total number of connection requests to be injected into the simulation, and the maximum bandwidth that can be requested by each connection (referred as request granularity). This allows parallel comparison of the schemes which could assist in identifying the most efficient traffic provisioning algorithm that best suits a particular network topology under certain traffic (and failure) pattern.
Archive Staff Only: edit this record