Changes between Version 11 and Version 12 of MapGuideRfc90


Ignore:
Timestamp:
Oct 14, 2011 3:29:49 AM (6 years ago)
Author:
uvlite
Comment:

align with revision 6174

Legend:

Unmodified
Added
Removed
Modified
  • MapGuideRfc90

    v11 v12  
    99||RFC Template Version||(1.0)||
    1010||Submission Date||(Date/Time submitted)||
    11 ||Last Modified|| Uv Nov 4 2010||
     11||Last Modified|| Uv Oct 14 2011||
    1212||Author||Zac Spitzer, Uv ||
    13 ||RFC Status||(draft)||
     13||RFC Status||(ready for review)||
    1414||Implementation Status||(testing sandbox)||
    15 ||Proposed Milestone||(2.2)||
     15||Proposed Milestone||(2.3)||
    1616||Assigned PSC guide(s)||(when determined)||
    1717||'''Voting History'''||(vote date)||
     
    3939e.g.
    4040{{{
    41 Client -> GetTile(0,1) -> RenderMetaTile(0,0 - 1,1) -> SplitUp -> ReturnTile (0,1)
     41Client -> GetTile(0,1) -> GetMetaTile(0,0 - 1,1) -> RenderFromMetaTile-> ReturnTile (0,1)
    4242       -> GetTile(1,1) -> wait................................... ReturnTile (1,1)
    4343}}}
     44
    4445== Issues ==
    4546
    4647Locking the threads to the implicit producer consumer scheme is complex and
    47 the file locking seems not very efficient. To improve polling delay the Polling
    48 configuration value has been changed to ms semantics so we can specify 500ms.
     48the file locking seems not very efficient.
    4949
    50 In addition an ace_condition has been implemented to get rid of the lockfiles
    51 all together and further improve the response time. (LockMethod == 3)
    52 In addition to the standard approach with one lock per sub tile (LockMethod == 2)
    53 we also experimented using a single file per metatile (LockMethod == 1).
    54 This is a variable in the TileService section of the serverconfig.ini together with the
    55 UseMetaTiles variable specifying the multiplication factor.
     50The default locking method is 0 which uses a single file per metatile (LockMethod == 0).
     51LockMethod is a config value in the TileService section of the serverconfig.ini together with the
     52UseMetaTiles value specifying the multiplication factor.
     53
     54In addition an attempt to use ace_conditions has been implemented to get rid of the lockfiles
     55all together and further improve the response time. (LockMethod == 1)
     56There are still some stability issues in the release version.
    5657
    5758== Implications ==
     
    5960For example when using a meta tiling factor of 4, 15 tiles will wait while the meta tile is rendered
    6061and sliced up. A most efficient locking scheme will have significant effect on this behaviour.
    61 
    6262
    6363== Test Plan ==
     
    6767== Issues ==
    6868
    69 There are extra tiles calculated to fill up the grid of the meta tiles. They are empty small and don't take long to compute. Any algorithm to avoid that would cause a lot of extra complications due to the loss of the currently orthogonal algorithm. This would impact tile generation and locking strategies. So we currently think its not an issue.
     69There are extra tiles calculated to fill up the grid of the meta tiles. They are empty tile and don't take long to compute. Any algorithm to avoid that would cause a lot of extra complications due to the loss of the currently orthogonal algorithm. This would impact tile generation and locking strategies.
     70Using the MgCooker is adding even more empty tiles around the map boundaries. So we currently think its not an issue.
     71
     72There seem to be some leaks somewhere which could not be found also with some tools. However, they are small and considering the use case of the metatiling mode - namely filling the tile cache - they might be tolerable.
     73
     74[[Image(MapGuideMemoryMetatilingLockfiles.png)]]
     75
     76[[Image(MapGuideMemoryNoMetatiling.png)]]
    7077
    7178== Funding/Resources ==
     
    7582== Sandbox ==
    7683
    77 http://trac.osgeo.org/mapguide/browser/sandbox/rfc90 OBSOLETE Due to many trunk changes
    7884http://trac.osgeo.org/mapguide/browser/sandbox/rfc90.2  branched from Rev 5924 and merged in changes
     85
     86so http://trac.osgeo.org/mapguide/browser/sandbox/rfc90 is now OBSOLETE