DocumentCode :
2999449
Title :
Middleware Support for RDMA-based Data Transfer in Cloud Computing
Author :
Ren, Yufei ; Li, Tan ; Yu, Dantong ; Jin, Shudong ; Robertazzi, Thomas
Author_Institution :
Dept. of Electr. & Comput. Eng., Stony Brook Univ., Stony Brook, NY, USA
fYear :
2012
fDate :
21-25 May 2012
Firstpage :
1095
Lastpage :
1103
Abstract :
Providing high-speed data transfer is vital to various data-intensive applications in cloud computing systems. We design a middleware layer of high-speed communication based on Remote Direct Memory Access (RDMA) that serves as the common substrate to accelerate various data transfer applications in cloud computing, such as FTP, HTTP, file copy, sync and remote file I/O. This middleware offers higher end-to-end bandwidth than the traditional TCP-based alternatives, while it hides the heterogeneity of the underlying high-speed architecture. In this paper, we describe the design of this middleware, including resource abstraction, and task synchronization and scheduling. We provide a reference implementation of the file-transfer protocol over this RDMA based middleware. Our experimental results show that it outperforms several TCP-based FTP tools, such as GridFTP, while maintaining very low CPU consumption on a variety of platforms. Furthermore, these results confirm that our middleware achieves near line-speed performance in both LAN and MAN, and scales consistently from 10Gbps Ethernet to 40Gbps Ethernet and InfiniBand environments.
Keywords :
access protocols; cloud computing; middleware; transport protocols; CPU consumption; Ethernet; GridFTP; HTTP; InfiniBand environment; LAN; MAN; RDMA based middleware; RDMA-based data transfer; TCP-based FTP tool; cloud computing system; data-intensive application; end-to-end bandwidth; file copy; file-transfer protocol; high-speed architecture; high-speed communication; high-speed data transfer; line-speed performance; middleware layer; middleware support; remote direct memory access; remote file I/O; resource abstraction; scheduling; sync file I/O; task synchronization; Computer architecture; Message systems; Middleware; Protocols; Semantics; Servers; Data Center Networks; Middleware; Protocol offloading; Remote Direct Memory Access;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Parallel and Distributed Processing Symposium Workshops & PhD Forum (IPDPSW), 2012 IEEE 26th International
Conference_Location :
Shanghai
Print_ISBN :
978-1-4673-0974-5
Type :
conf
DOI :
10.1109/IPDPSW.2012.133
Filename :
6270759
Link To Document :
بازگشت