Documentation for the P2PNS live demo

For demonstration purposes we've deployed our distributed name service P2PNS on the ​PlanetLab research network. Currently our name service is running on about 450 nodes, which are monitored by a management node. Every node has registered a SIP name and can be called with an unmodified SIP UA (there's an echo service running on each node).

Please note: The current release still lacks some important security features! Therefore you can e.g. easily hijack a registered user name. This will be fixed in the next release.

Network monitor

The management node provides the current network status either on this ​webpage or as ​kml file, which can be viewed with ​Google Earth.

Each icon represents the status of a PlanetLab node. The lower part of the icon shows the general node status, the middle part show the overlay status and the upper parts indicates, if the node successfully registered a SIP name.

Things you can try

• Watch the load average of a PlanetLab node by clicking on an icon on the monitor page.
• Make a test call from the management node to any of the PlanetLab nodes: Click on a icon on the monitor page and click on the "Call this node" link. This will open a new window showing the call setup latency (from the initial SIP INVITE to the 200 OK) and the number of SIP retransmissions.
• You can try a test call with you local SIP client: Simply click on a icon on the monitor page and look for the registered SIP name sip:xxx@p2pname.org. Dial this name with your SIP UA and you should get connected to the echo test service on this node.
• Register a new SIP user in the P2PSIP overlay with your local SIP client: Configure p2pname.org:5061 as registrar and outgoing SIP proxy and choose an arbitrary user name with empty password.
• Use the command interface from the monitor page to control a P2PNS node. You could e.g. try to lookup the siblings to a key using different routing modes and compare the measured lookup latency.

Overlay parameters

The P2PNS nodes in PlanetLab currently use the following parameters:

• Kademlia with k=20 and s=8 siblings
• Iterative, exhaustive iterative, semi recursive, full recursive or source routing
• Only UDP for overlay signaling
• DHT with 4 replica
• Parallel lookups and DHT queries