Published in Wu, Q., Gu, Y., Zhu, M., & Rao, N.S.V. (2008). Optimizing network performance of computing pipelines in distributed environments. IEEE International Symposium on Parallel and Distributed Processing, 2008. IPDPS 2008, 1-8. doi: 10.1109/IPDPS.2008.4536465 ©2008 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE. This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright. In most cases, these works may not be reposted without the explicit permission of the copyright holder.


Supporting high performance computing pipelines over wide-area networks is critical to enabling large-scale distributed scientific applications that require fast responses for interactive operations or smooth flows for data streaming. We construct analytical cost models for computing modules, network nodes, and communication links to estimate the computing times on nodes and the data transport times over connections. Based on these time estimates, we present the Efficient Linear Pipeline Configuration method based on dynamic programming that partitions the pipeline modules into groups and strategically maps them onto a set of selected computing nodes in a network to achieve minimum end-to-end delay or maximum frame rate. We implemented this method and evaluated its effectiveness with experiments on a large set of simulated application pipelines and computing networks. The experimental results show that the proposed method outperforms the Streamline and Greedy algorithms. These results, together with polynomial computational complexity, make our method a potential scalable solution for large practical deployments.