The problem of topology aggregation (TA) for scalable, quality-of-service (QoS)-based routing in hierarchical networks is investigated. This problem has been a hot research topic since the early 1990s. TA deals with the process of summarizing the topological information of a subset of network elements. This summary is then advertised throughout the network and used by network nodes to select the appropriate routes based on the QoS parameters. Summarization introduces “loss” of the topological information, which could affect the routing performance. The goal of TA design is to balance the compaction and the routing performance.
This paper proposes a new source-oriented TA approach. Sharply different from conventional approaches, this approach eliminates redundancy in the advertised information by advertising only the relevant information to a source node for path selection. Three new source-oriented TA schemes are proposed. Another contribution of this paper is the extension of the source-oriented approach to a multi-QoS-based TA. Two new approaches are introduced for determining the appropriate values for the multiple QoS parameters associated with a logical link. These two approaches overcome the overly aggressive or conservative tendency in determining the actual value in the conventional approaches. Extensive simulations on sparse and dense network topologies under both the static and the dynamic scenarios illustrate the merit and improvement of routing performance of these new approaches.
This paper is well organized and easily understood. The comprehensive presentation of conventional approaches and new approaches in TA make it valuable to both advanced researchers and general interested readers.