General

This entry will explain you how to interpret the "Search Details" Graphin the Kad dialog (if you don't see it, click on the "Contacts" button). However before doing so, there is one important thing mention: You don't need to understand, monitor or look at this graph at all to have Kad work properly. This graph provides insights into Kads working mechanisms, but if you are not interested in it you can safely skip this page and eMule won't even be angry with you ;)

What this graph is about

Kad is a decentralized network and database. Every action consists of searching and contacting other nodes (users) to retrieve or store information about something. This graph visualizes this process, allowing you to watch in real time how eMule searches the proper nodes and contacts them.

Details

There are basically three types of queries you will see: Storing data, retrieving data and finding other nodes to fill your routing table. The last one is the most common action, but also the most boring one, so lets rather start by explaining how eMule finds a keyword your are searching for.
If you want to follow the explanation on your own eMule: Be sure to be connected to Kad, go to the search dialog, enter a keyword to search for, select "Kad" as method and click start, switch back to the Kad screen and click the entry which has your keyword as name in the "Current searches list".
You will now see how the graph develops and in the end it should look kinda like this:

So what does it mean? Each dot represents a node. The closer a dot is to the bottom, the closer it is to the keyword we are searching (not physically), so Kads first task is to find nodes which are close to the keyword, because those are most likely to know something about our keyword.
To archieve this Kad asks other nodes if they know nodes which are closer to the keyword than themself and if they do they will send us those. In the graph this is shown by arrows. An arrow pointing from one node to another means that the node where the arrow start sent us the node where the arrow ends. Of course several nodes can (and will) report nodes which we already know, this is why you will see nodes which have several arrows pointing at them. Nodes which have no arrow pointing at them at all were taken out of our local routing table - mostly those are the first three top left nodes.

The color of the nodes shows how the nodes responded to our request:

• : This node answered our request and sent us nodes which are closer to the keyword than itself (when retrieving information like in the example above most nodes send us 2 results, when storing information nodes send 4 results)
• : This node was asked, but we are still waiting for an answer
• : This node didn't answered at all, it is probably dead
• : This node answered, but wasn't able to find nodes which are closer to the keyword than itself

Once Kad has found enough nodes, it will start to ask them if they have any files indexed for the actual keyword we are searching. If they do they will send us the result and when we have enough of them or after we searched for a given time, Kad will stop the query and the search is finished. That is why searches for a common keyword will result in only a few nodes shown on the graph and end after a few seconds, while searches for a rare keyword might go on for quite some time.

The keyword request to other nodes is shown on the graph with those symbols:

• : The node had files indexed for our keyword and sent us the result which will now be listed in the search dialog
• : The node had no files stored for our keyword or didn't answer

Finally, you can watch all details about any node by hovering the mousepointer over it and waiting for the tooltip.

This concludes the explanation about how a keyword search is shown on the Kad graph. As written above, there are two other types of actions: Storing files, keywords or notes works pretty similar to the explained example, except that we don't expect results when we store our keyword or file. Finding nodes for our routing table ("Node Lookups") on the other hand looks always like one green node (the one we asked) pointing to 11 yellow nodes (those we added to our routing table) - there is nothing much to see here.