Databases for Spatial Graph Management

Please scroll down to the bottom of the page for a graphical illustration.

T1.gif T2.gif

T3.gif T4.gif

T5.gif T6.gif

T7.gif T8.gif

T9.gif T10.gif

T11.gif T12.gif

T13.gif T14.gif
Paging Using Rival technique: Sorting by Depth First Search Traversal. Road intersections and segments within a page have a common color.
Lots of white (cut) edges => WORSE clustering.
CRR = 0.609 for
Minneapolis Major Roads (1079 nodes, 3067 edges)
Stored in 40 disk pages Paging Using Our CCAM (Connectivity Clustered).
Road intersections and segments within a page have a common color.
Has fewer white (cut) edges => BETTER clustering.
CRR = 0.827 for
Minneapolis Major Roads (1079 nodes, 3067 edges) Stored in 40 disk pages.
Effective paging is the key to minimize disk I/O cost which dominate the total response for many applications.


T1.gif	T2.gif

T3.gif	T4.gif

T5.gif	T6.gif

T7.gif	T8.gif

T9.gif	T10.gif

T11.gif	T12.gif

T13.gif	T14.gif
Paging Using Rival technique: Sorting by Depth First Search Traversal. Road intersections and segments within a page have a common color. Lots of white (cut) edges => WORSE clustering. CRR = 0.609 for Minneapolis Major Roads (1079 nodes, 3067 edges) Stored in 40 disk pages	Paging Using Our CCAM (Connectivity Clustered). Road intersections and segments within a page have a common color. Has fewer white (cut) edges => BETTER clustering. CRR = 0.827 for Minneapolis Major Roads (1079 nodes, 3067 edges) Stored in 40 disk pages. Effective paging is the key to minimize disk I/O cost which dominate the total response for many applications.