Changes between Version 7 and Version 8 of MapGuideRfc90


Ignore:
Timestamp:
Nov 4, 2010, 1:13:29 AM (14 years ago)
Author:
uvlite
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • MapGuideRfc90

    v7 v8  
    2020== Overview ==
    2121
    22 Add support internally to the MapGuide Server Engine to render larger tiles
    23 and then slice them up into the normal tile sizes
     22Add support internally to the MapGuide Server Engine to map tile requests to larger meta tiles
     23and then slice them up into the standard tile sizes.
    2424
    2525== Motivation ==
    2626
    27 With complex maps, it's more efficent to render one large tile than many small ones.
     27With complex maps, it's more efficient to render one large tile than many small ones.
    2828There is little or no benefit for simple maps like the Sheboygan Sample.[[BR]]
    2929
    30 Using our 140 layer example map we can observer the following timing using metatilefactor 1 & 4
     30Using our 140 layer example map we can observe the following timing using metatilefactor 1 & 4
    3131and 3 different locking strategies in the attached chart MetatilingChart1.jpg.
    3232
    3333== Proposed Solution ==
    3434
    35 Add a meta tile factor to the tile service, which then will render larger tiles
    36 and then slice them up into the smaller tiles which get served to clients.
     35Add a meta tile factor to the tile service, which then will render larger tiles enlarged by the provided factor.
     36to be sliced up into the smaller tiles which then get served to clients.
     37e.g.[[BR]]
     38Client -> GetTile(0,1) -> RenderMetaTile(0,0 - 1,1) -> SplitUp -> ReturnTile (0,1)
     39       -> GetTile(1,1) -> wait................................... ReturnTile (1,1)
    3740
    3841== Issues ==
    3942
    4043Locking the threads to the implicit producer consumer scheme is complex and
    41 file locking seems not very efficient. To improve polling delay the Polling
     44the file locking seems not very efficient. To improve polling delay the Polling
    4245configuration value has been changed to ms semantics so we can specify 500ms.
    4346
    44 In addition a ace_condition has been implemented to get rid of the lockfiles
     47In addition an ace_condition has been implemented to get rid of the lockfiles
    4548all together and further improve the response time. (LockMethod == 3)
    4649In addition to the standard approach with one lock per sub tile (LockMethod == 2)
     
    4952== Implications ==
    5053
    51 When using a meta tiling factor of 4, 15 tiles will wait while the meta tile is rendered
     54For example when using a meta tiling factor of 4, 15 tiles will wait while the meta tile is rendered
    5255and sliced up. A most efficient locking scheme will have significant effect on this behaviour.
    5356
     
    5558== Test Plan ==
    5659
    57 The output of a tile map request should be indentical whether served
    58 with meta tiling enabled or without.
     60The output of a tile map request should be identical whether served with meta tiling enabled or without.
    5961
    6062== Funding/Resources ==