In this paper, we study the all-to-all personalized and non-personalized broadcast problems in a star graph. This work is motivated by the observation that existing works in the literature only try to optimize the data transmission time, but ignore the start-up time required to initialize communication. As a result, existing algorithms, although claimed to be optimal, are only so when the start-up time is negligible. In this paper we try to optimize both costs at the same time. We develop an all-to-all personalized broadcast algorithm that is more efficient than existing results. We also present an all-to-all non-personalized broadcast algorithm that outperforms existing results in most cases except when the start-up time is very small (in a ratio of O(1/(n - 1)!), where n is the dimension of the star graph) compared to the transmission time. Extensive simulations have been conducted, which show 10% approx. 80% improvement over existing results.
|Number of pages||9|
|Journal||Proceedings of the National Science Council, Republic of China, Part A: Physical Science and Engineering|
|State||Published - 1 Nov 1998|