ARTICULATING USER PREFERENCES ABOUT SURVIVABILITY IN A NETWORK PLANNING DSS

摘要

Decision support systems for planning today’s high capacity networks, like those in telecommunications, must include features that model “survivability”. This property assures that traffic is protected from the potential damages associated with some disruption, like a break in a network link. When a network provides “100% survivability”, the property is fairly well defined: all of the traffic is protected against any network failure. However, if a planner is willing to accept some smaller percentage of survivability, e.g., to lower the cost of the network, then some ambiguities arise. There are a variety of ways that partial survivability can be interpreted, measured, and provided. Currently available decision support systems for network planning are limited in their ability to model these different options. In this paper, we proposed some features that could be added as enhancements to a DSS, so that it can better satisfy a planner’s needs regarding survivability. We describe features to model the alternative schemes available to measure partial survivability. These would give the planner the opportunity to articulate more meaningful objectives regarding network survivability. The enhanced DSS could then invoke the appropriate models and algorithms to help the planner identify what she would consider to be superior network solutions. In addition, such a system could help the planner evaluate the trade-offs between network survivability and network cost.