[SPARK-2522] set default broadcast factory to torrent

HttpBroadcastFactory is the current default broadcast factory. It sends the broadcast data to each worker one by one, which is slow when the cluster is big. TorrentBroadcastFactory scales much better than http. Maybe we should make torrent the default broadcast method.

Author: Xiangrui Meng <meng@databricks.com>

Closes #1437 from mengxr/bt-broadcast and squashes the following commits:

ed492fe [Xiangrui Meng] set default broadcast factory to torrent

core/src/main/scala/org/apache/spark/broadcast/BroadcastManager.scala

@@ -39,7 +39,7 @@ private[spark] class BroadcastManager(
     synchronized {
       if (!initialized) {
         val broadcastFactoryClass =
-          conf.get("spark.broadcast.factory", "org.apache.spark.broadcast.HttpBroadcastFactory")
+          conf.get("spark.broadcast.factory", "org.apache.spark.broadcast.TorrentBroadcastFactory")
 
         broadcastFactory =
           Class.forName(broadcastFactoryClass).newInstance.asInstanceOf[BroadcastFactory]

docs/configuration.md

@@ -419,7 +419,7 @@ Apart from these, the following properties are also available, and may be useful
 </tr>
 <tr>
   <td><code>spark.broadcast.factory</code></td>
-  <td>org.apache.spark.broadcast.<br />HttpBroadcastFactory</td>
+  <td>org.apache.spark.broadcast.<br />TorrentBroadcastFactory</td>
   <td>
     Which broadcast implementation to use.
   </td>