BMI: A Network Abstraction Layer for Parallel I/O Export

@inproceedings{citeulike:3504877, abstract = {As high-performance computing increases in popularity and performance, the demand for similarly capable input and output systems rises. Parallel I/O takes advantage of many data server machines to provide linearly scaling performance to parallel applications that access storage over the system area network. The demands placed on the network by a parallel storage system are considerably different than those imposed by message-passing algorithms or datacenter operations; and, there are many popular and varied networks in use in modern parallel machines. These considerations lead us to develop a network abstraction layer for parallel I/O which is efficient and thread-safe, provides operations specifically required for I/O processing, and supports multiple networks. The Buffered Message Interface (BMI) has low processor overhead, minimal impact on latency, and can improve throughput for parallel file system workloads by as much as 40\% compared to other more generic network abstractions.}, address = {Washington, DC, USA}, author = {Carns, Philip and Ross, Robert and Iii, Walter L. and Wyckoff, Pete}, booktitle = {IPDPS '05: Proceedings of the 19th IEEE International Parallel and Distributed Processing Symposium (IPDPS'05) - Workshop 9}, citeulike-article-id = {3504877}, citeulike-linkout-0 = {http://portal.acm.org/citation.cfm?id=1054647}, citeulike-linkout-1 = {http://dx.doi.org/10.1109/IPDPS.2005.128}, doi = {10.1109/IPDPS.2005.128}, isbn = {0-7695-2312-9}, keywords = {filesystem, io, network, parallel, pvfs}, posted-at = {2008-11-12 03:39:39}, priority = {2}, publisher = {IEEE Computer Society}, title = {BMI: A Network Abstraction Layer for Parallel I/O}, url = {http://dx.doi.org/10.1109/IPDPS.2005.128}, year = {2005} }

TY - CONF ID - citeulike:3504877 L3 - citeulike-article-id:3504877 N2 - As high-performance computing increases in popularity and performance, the demand for similarly capable input and output systems rises. Parallel I/O takes advantage of many data server machines to provide linearly scaling performance to parallel applications that access storage over the system area network. The demands placed on the network by a parallel storage system are considerably different than those imposed by message-passing algorithms or datacenter operations; and, there are many popular and varied networks in use in modern parallel machines. These considerations lead us to develop a network abstraction layer for parallel I/O which is efficient and thread-safe, provides operations specifically required for I/O processing, and supports multiple networks. The Buffered Message Interface (BMI) has low processor overhead, minimal impact on latency, and can improve throughput for parallel file system workloads by as much as 40% compared to other more generic network abstractions. SN - 0-7695-2312-9 AD - Washington, DC, USA TI - BMI: A Network Abstraction Layer for Parallel I/O PB - IEEE Computer Society T2 - IPDPS '05: Proceedings of the 19th IEEE International Parallel and Distributed Processing Symposium (IPDPS'05) - Workshop 9 KW - filesystem KW - io KW - network KW - parallel KW - pvfs AU - Carns, Philip AU - Ross, Robert AU - Iii, Walter AU - Wyckoff, Pete PY - 2005/// UR - http://dx.doi.org/10.1109/IPDPS.2005.128 ER -