发明名称 |
Data locality in data integration applications |
摘要 |
A computer-implemented method includes identifying a logical node. The logical node is associated with one or more source stages. The computer-implemented method further includes identifying one or more file block components. The one or more file block components include a retrieval target for at least one of the one or more source stages. The computer-implemented method further includes identifying one or more physical nodes and determining, for each of the one or more physical nodes, a degree value. A corresponding computer program product and computer system are also disclosed. |
申请公布号 |
US9462058(B1) |
申请公布日期 |
2016.10.04 |
申请号 |
US201615050565 |
申请日期 |
2016.02.23 |
申请人 |
International Business Machines Corporation |
发明人 |
Bonagiri Krishna K.;Jacobson Eric A. |
分类号 |
G06F17/30;H04L29/08 |
主分类号 |
G06F17/30 |
代理机构 |
|
代理人 |
Ghalehshahi Pooya Shoghi |
主权项 |
1. A computer-implemented method comprising:
accessing a configuration file; identifying a logical node, said logical node being associated with one or more source stages; identifying one or more file block components, said one or more file block components comprising a retrieval target for at least one of said one or more source stages and being stored on a distributed file system; identifying one or more physical nodes; determining, for each of said one or more physical nodes, a degree value; identifying one or more qualified physical nodes from said one or more physical nodes having said degree value of one or more; creating a preferred physical node table, said preferred physical node table comprising, for each of said one or more qualified physical nodes, an identifying indication and an indication of said degree value; sorting said preferred physical node table based on said degree value associated with each of said one or more qualified physical nodes; determining a candidate preferred physical node based on each said degree value; determining whether said candidate preferred physical node is available for allocation to said logical node; responsive to said candidate preferred physical node being available for allocation to said logical node, allocating said candidate preferred physical node to said logical node; and responsive to said candidate preferred physical node not being available for allocation to said logical node:
marking said candidate preferred physical node as unavailable for allocation to said logical node; anddetermining an alternative candidate preferred physical node based on each said degree value. |
地址 |
Armonk NY US |