Fuzzy Bandwidth Broker: Machine Learning Based Approach to Resolve Architectural Issues

His paper proposes a novel idea of using fuzzy logic for architectural and resource management aspects of the bandwidth broker. The scalability problem of bandwidth broker, being a centralised resource manager in a domain, can be solved by employing a distributed architecture. The decisions regarding the distributed architecture, namely, number and location of distributed entities can be best solved using computational intelligence. This paper focuses on the fuzzy logic based approach for resolving architectural issues of bandwidth broker. In addition, we also propose two phase resource allocation algorithm for bandwidth broker. In first phase, when large amount of resources are available, fuzzy logic is used for decision making to reduce processing overhead. In case of low resource availability, the resource allocation algorithm transitions to second phase, where crisp values are used for decision purpose.his paper proposes a novel idea of using fuzzy logic for architectural and resource management aspects of the bandwidth broker. The scalability problem of bandwidth broker, being a centralised resource manager in a domain, can be solved by employing a distributed architecture. The decisions regarding the distributed architecture, namely, number and location of distributed entities can be best solved using computational intelligence. This paper focuses on the fuzzy logic based approach for resolving architectural issues of bandwidth broker. In addition, we also propose two phase resource allocation algorithm for bandwidth broker. In first phase, when large amount of resources are available, fuzzy logic is used for decision making to reduce processing overhead. In case of low resource availability, the resource allocating algorithm transitions to second phase, where crisp values are used for decision purpose.