-
公开(公告)号：GB2498620A
公开(公告)日：2013-07-24
申请号：GB201221370
申请日：2012-11-28
Applicant: IBM
Inventor： ZHANG YUE ZH , LIN HAIBO , ZHENG KAI , LI YAN
Abstract: The disclosure of the present invention relates to a method and system for network transfer adaptive optimization in a large-scale parallel computing system. In the method a message to be transferred through the network is formed based on information related to a task executed by computing node within the system (204). The message includes identification information of the computing nodes to perform data transfer, corresponding sink nodes, and an amount of data to be transferred. Preferably the message is created at the application layer. The message is then transferred to a network layer (208) where a new data transfer pattern for data transfer between said computing nodes and said corresponding sink nodes in accordance with the received message is created and executed (302). In optimizing the transfer pattern the sink node location may be adjusted. (304) The advantage of the method allows parallel frameworks such as MapReduce to be work with Software Design Networks (SDN) or Openflow frameworks.
-
2.发明专利
公开(公告)号：GB2498620B
公开(公告)日：2014-06-04
申请号：GB201221370
申请日：2012-11-28
Applicant: IBM
Inventor： ZHANG YUE ZH , LIN HAIBO , ZHENG KAI , LI YAN
Abstract: A method and apparatus for performing network transfer adaptive optimization in a computing system. The method includes forming a message to be transferred through the network based on obtained information related to a task executed by computing nodes, wherein the message includes identification of the computing nodes to perform data transfer and corresponding sink nodes, and an amount of data the computing nodes to transfer to the corresponding sink nodes; transferring the message to a network layer; and forming a data transfer pattern for data transfer between the computing nodes and the corresponding sink nodes in accordance with the message.
-
Search Results
2
records
(took 0.016 seconds)
