Palantir can be built from source code.

• Build Palantir service.

1. Prerequisites

   Java JDK. (Our tests are based on sun JDK 1,6, other version e.g. 1.7 or openjdk should work as well, but not tested). You can download JDK 1.6 from http://tinyurl.com/7w2jeyl, and follow the install instruction there to get Java ready.

   Apache Ant. Install by: apt-get install ant

2. Compile

   Go to elf/floodlight dir

   Compile by:

   ant clean; ant
   

3. Start Palantir service

   Go to elf/floodlight dir

   Start as normal floodlight controller:

   java -jar target/floodlight.jar
   

• Build Hadoop-Palantir patch

1. Prerequisites: the same as 1.1

2. Compile

   Go to elf/hadoop dir

   Compile by:

   ant clean; ant
   

As a user case, we have built a multi-datacenter aware Hadoop. To enable this functionality, you need to:

• Set up Hadoop cluster as normal, e.g. configure conf/*.xml and conf/masters conf/slaves. This is the basic configuration of Hadoop.

• In conf/core-site.xml, add the following property:

      <property>
              <name>topology.node.switch.mapping.impl</name>
              <value>org.apache.hadoop.net.SDNTopologyManager</value>
     </property>

• Use REST API to register network topology with Palantir service:

1. Register root switch of the network:

curl -d '["the dpid of root switch"]' http://controller ip:controller port/wm/hadooptopology/root/json.

For example the root dpid is 00:00:00:00:00:00:00:01, your Palantir service (floodlight controller) is running on localhost:8080, then this registration should be like:

curl -d '["00:00:00:00:00:00:00:01"]' http://localhost:8080/wm/hadooptopology/root/json

2. Register the gateway switches

curl -d '{"datacenter id":["gateway1 dpid", "gateway2 dpid" ... ], ... }' http://controller ip:controller port/wm/hadooptopology/gateway/json

For example, there are 2 datacenters, named dc1 and dc2. the gateways of dc1 is 00:00:00:00:00:00:00:02 and 00:00:00:00:00:00:00:03. The gateway of dc2 is 00:00:00:00:00:00:00:04, then the registration would be like this:

curl -d '{"dc1":["00:00:00:00:00:00:00:02", "00:00:00:00:00:00:00:03"], "dc2":["00:00:00:00:00:00:00:04"] }' http://localhost:8080/wm/hadooptopology/gateway/json

3. After these setting up, you can verify them by querying a host, e.g. 10.0.0.1 using:

curl -s http://localhost:8080/wm/hadooptopology/?host=10.0.0.1

This should return the network location according to your registration. Our patched datacenter aware Hadoop will use a similar API to conmmunicate with the Palantir service