Okazaki, Takumi (2005) Retrieving subset of result before completing top-k query. Technical Report. Stanford.
As the power of PCs is getting better, Peer-to-peer (P2P) architectures are becoming popular. Pure P2P does not have servers, and peers (PCs) share their resource. A P2P strategy can reduce the cost of the system. One of the basic problems is to search by keyword in P2P system. To enable the efficiency of P2P, search information should be managed in a distributed fashion not in central servers. One approach to reducing query cost is to search only the top-k elements, in order not to calculate the complete results. We focus on the top-k algorithm and propose a method to recognize some objects are certainly among top-k before completing top-k algorithm.
|Item Type:||Techreport (Technical Report)|
|Subjects:||Computer Science > Distributed Systems|
Computer Science > Query Processing
|Related URLs:||Project Homepage||http://infolab.stanford.edu/peers/|
|Deposited By:||Import Account|
|Deposited On:||24 Aug 2005 17:00|
|Last Modified:||22 Dec 2008 18:34|
Repository Staff Only: item control page