DocumentCode
2132935
Title
A study of high-performance communication mechanism for multicomputer systems
Author
Murayama, Hideki ; Yoshizawa, Satoshi ; Aimoto, Takeshi ; Inouchi, Hidenori ; Murase, Shooichi ; Hayashi, Takehisa ; Iwamoto, Hiroshi
Author_Institution
Inf. Syt. R&D Div., Hitachi Ltd., Japan
fYear
1996
fDate
15-19 Apr 1996
Firstpage
76
Lastpage
83
Abstract
Based on our analysis of the behavior of standard UNIX communication software, we propose a new communication mechanism which reduces the software overhead. This mechanism is optimized to be used in business applications, such as those of online transaction processing and database management. These applications are currently designed to operate in multiple concurrent processes, thus requiring multiprocessing environments with virtual addressing. The proposed communication mechanism, named DLA (dynamic link allocation), supports direct data transfer communication between processors without any intermediate data copying operations. The DLA mechanism adopts a synchronous non-blocking interface which can be used directly by user processes. Since the DLA mechanism provides address translation for each communication port, the sender may directly specify the virtual address of the receiver´s buffer area. With the DLA mechanism, sender and receiver CPU usages are reduced to one-third of those for TCP/IP communication, while attaining 30% higher throughput
Keywords
Unix; buffer storage; business data processing; combinational switching; message passing; multiprocessing systems; virtual storage; CPU usage; DLA; UNIX communication software; address translation; business applications; communication ports; database management; direct data transfer communication; dynamic link allocation; high-performance communication mechanism; multicomputer systems; multiple concurrent processes; multiprocessing environments; online transaction processing; receiver´s buffer area; software overhead; synchronous nonblocking interface; throughput; user processes; virtual address specification; virtual addressing; Application software; Communication standards; Communication system software; Computer interfaces; Computer networks; Concurrent computing; Message passing; Parallel processing; Protocols; Software performance; Software standards; TCPIP; Throughput; Transaction databases;
fLanguage
English
Publisher
ieee
Conference_Titel
Parallel Processing Symposium, 1996., Proceedings of IPPS '96, The 10th International
Conference_Location
Honolulu, HI
Print_ISBN
0-8186-7255-2
Type
conf
DOI
10.1109/IPPS.1996.508042
Filename
508042
Link To Document