Implementation and Preliminary Evaluation of IDIPS: technical notes

The idea behind IDIPS is to provide ordered paths according to pre-defined criteria. IDIPS is composed of three entities: the client, the destination and the IDIPS server itself. The client is the entity sending request for QoS and path information towards a destination. The IDIPS server aims at providing the requested information to the client. IDIPS is operated by an ISP or an AS. A client sends a list of all the possible source and destination addresses to the IDIPS server. In addition, the client gives its criteria for evaluating paths. When an IDIPS server receives such a request, it creates a set of conceivable (source, destination) pairs, where each source and destination belongs to the initial list provided by the client. The IDIPS server then orders the pairs according to the client’s criteria so that the first pair in the list is the best choice while the last one is the least preferred. The IDIPS server responds to the client by sending back this ordered list. Note that the addresses list provided by the client might be composed of prefixes, not necessarily of complete addresses. The IDIPS protocol is defined in [1]. This technical paper aims at evaluating our current implementation of the IDIPS server. First, Sec. 2.1 presents the main concepts of the implementation. After, Sec. 2.2 gives some details about the implementation choices. Next, Sec. 3.2 describes the performances testbed and Sec. 3.2 comments the results for this testbed. Finally, Sec. 4 concludes and presents the future implementation’s milestones.