You can find a visualisation of the test network we have set up here :- Freenet Network Structure∞.

Explanation

Most of the page consists of a graph of the nodes in the test network, how they are interconnected, and some information about each node. At the bottom of the page there is also a histogram showing the number of connections of varying "lengths". The length of a connection is determined by the difference between the "locations" of the nodes it connects. The location of a node is a number between 0 and 1, the location of each node is displayed on the graph.

Connections are fixed, but locations can change. The algorithm we use aims to achieve a large number of short connections, and a small number of long connections, by swapping the locations of nodes in the network. Ideally the probility of a connection existing between two nodes should be proportional to 1/D where D is the distance between the nodes.

Each node in the graph looks something like this:

sandos	the name of the node
0.04941	the node's location
R:0 I:0 TR:0	the number of requests, inserts and requests that are currently transferring
Ver. 321	the version of the node

the color of the node indicates:
yellow: the node has less then 3 connection
blue: the node uses the version smaller then last good build and can not connect therefore to the up to date nodes
green: everything ok

linkcolours:
red: there is a backoff, the node backing off is marked , and the backoff time is shown
black: indicates ocurence of locations swap, points to the nodes which swaped locations recently + addresses which were swaped

font:
font is normally rather small when the node is inactive.
It gets bigger when the node make inserts or requests or is transferring requests


The histogram at the bottom of the page makes it easy to see how successful the algorithm has been in the test network.