Differences between revisions 4 and 6 (spanning 2 versions)
Revision 4 as of 2005-11-29 14:29:40
Size: 1479
Editor: alpt
Comment:
Revision 6 as of 2006-11-08 21:01:17
Size: 1052
Editor: alpt
Comment: cleaned
Deletions are marked like this. Additions are marked like this.
Line 3: Line 3:
Subject: Life probability Subject: Life probability 
Line 17: Line 17:
in a multipath route we prefer to don't use very young nodes as main gateways. we prefer to avoid routes which contain very young nodes.
Line 20: Line 20:
                A
         / \
         L K
         \ /
  A
  / \
  L K
  \ /
Line 32: Line 32:
In order to implement this we penalize the new nodes/gnodes by increasing
their rtt, in this way they will be considered as slow links.

We can do this only for new nodes/gnodes 'cause if we cannot trust the uptime
value given by a node, but we can know for sure if that node is completely
new.

When the next qspn_round updates the routes the penality is removed.

Another system would be counting the number of qspn_rounds the nodes took part in.
If it has only partecipated in < 3 rounds it is considered young.
In the TP we record the uptime of each traversed node. In this way, the node
receiving the TP will be able to discriminate young routes from stable ones.
Line 45: Line 36:
related: ["Netsukuku RFC"] related: [Netsukuku_RFC]

NTK_RFC 0005

Subject: Life probability


This text describes a change to the Npv7. It will be included in the final documentation, so feel free to correct it. But if you want to change the system here described, please contact us first.


Statistically the probability of a node to continue to stay up is related to its uptime, for example if the node X has an uptime of 30 seconds and the node Y of 1200 seconds, then it is more probable that the node Y will continue to stay up.

For this reason we give higher credits to older nodes/gnodes, this means that we prefer to avoid routes which contain very young nodes.

                A
               / \
              L   K
               \ /
                N

Let's say K has an uptime of 600 seconds and L of 200. The node N, to reach A, will prefer the node K.

Implementation

In the TP we record the uptime of each traversed node. In this way, the node receiving the TP will be able to discriminate young routes from stable ones.


related: [Netsukuku_RFC]

Ntk_life_probability (last edited 2008-06-26 09:53:41 by anonymous)