13.2.3 Editing the web.xml File

This section describes changes to parameter values in the web.xml file that you must make for the routing engine to deploy properly. (There are also additional parameters that you can change to alter how the routing engine operates.)

  • Change the container_ds parameter to be the JNDI Name of the data source associated with the managed server. For example: JNDI/NorthAmericanDS

  • Change the routeserver_network_name parameter to the name of the Network Data Model (NDM) network built on the routing engine road network data. For example: NorthAmericanNetwork

  • If the WLS Managed Server has a Work Manager associated with it, change the wl-dispatch-policy parameter value to the name of the Work Manager. For example: NorthAmericanWM

  • Check to be sure the geocoder_type parameter is set to httpclient or None. (thinclient is no longer supported.)

    • If set to httpclient, then also set geocoder_http_url to the URL of the Geocoder servlet. For example: http://localhost:8888/geocoder/gcserver

    • If an HTTP proxy is being used, then also specify geocoder_http_proxy_host and geocoder_http_proxy_port. If no proxy exists, these two parameters can be ignored

  • If necessary, change the logfile_name parameter value. By default, the logfile_name parameter is set to log/RouteServer.log. This default relative path includes a subdirectory named log, relative to where the routing engine is installed. The logfile_name parameter can also be set to an absolute path, for example: /scratch/logs/RouteServer.log.

  • Change the partition_cache_size parameter. The default value for this parameter is 70, but it will probably need to be changed depending on the amount of memory allocated to the heap on the managed server. The following formula can be used to get a good starting point for a cache size.

    partition_cache_size = (NodesPerGigabyte/AvgNodesPartition)*UsableMemory

    Where:

    • NodesPerGigabyte is the number of nodes per gigabyte. (This value should not change. In the data sets as of December 2013, this value is 15000000, that is, 1.5 million.)

    • AvgNodesPartition is the average number of nodes per local partition. This does not include the highway partition 0. The memory for the highway partition is accounted for in the 1 gigabyte subtracted from the allocated heap size. For the North American data set, the AvgNodesPartition value is around 26000. You can check the actual average nodes per partition by using the following query:

      SELECT AVG(COUNT(node_id))
     FROM node 
     WHERE partition_id>0
     GROUP BY partition_id;

    • UsableMemoryis the managed server allocated heap size in Gigabytes minus 1 Gigabyte.

    This formula generates a safe number for the partition_cache_size parameter. Depending on the types of user information being used and the average number of concurrent requests being processed, it may be possible to add another 15% to 20% to this number. Use the WLS console to monitor the heap usage before changing this number.

    The heap can then be monitored while the routing engine is running to tune this number up or down. However, setting this value too high may cause the managed server to run out of memory.