Tile Scheduling Algorithm Based on Active Cache in P2P Massive Terrain Navigation

Shaoming PAN,Zhanwu Yu,Rui LI
DOI: https://doi.org/10.3321/j.issn:1001-1595.2009.03.008
2009-01-01
Abstract:With continuous development of computer technology,3D terrain navigation is a hot application in network Geographic Information System(GIS),and massive data transmission is highly demanded.As the number of clients increases,the server performance will drop too rapidly to satisfy the client real-time rending requirement because the server loads are increasing linearly for the kind of client/server architecture.The method through Peer-to-Peer(P2P) to address the server bandwidth and performance bottleneck in large scale massive terrain navigation has attracted much effort based on P2P networks.Due to peers dynamics and heterogeneity,some peers become the hot-pots,while some others are idle.An active cache,a kind of terrain data scheduling algorithm based on grouped locality exiguity,is proposed.The scheduling algorithm,called GLRF,is proposed for the idle peers,which just helps others getting their data and does not play the tile data.When deciding which segments to get to improve the peer hit rate and peer matching rate,in addition to the rarity and uniformity considered in existing data algorithms.Compared test result in prototype system indicates that the proposed algorithm has high tile request hit ratio and match ratio,and the two metrics have high increase speed than traditional method.
What problem does this paper attempt to address?