gmane.network.peer-to-peer.netsukuku

Re: Another question.

Alpt — 2008-11-21T15:11:04

On Thu, Nov 20, 2008 at 07:50:54PM +0100, : ~> Don't you think that all this ETP flying around the net, could overload the ~> network? ~> If all the nodes start an ETP in the same time, there could be some problem ~> right? As you can read in the previous email (DART, Netsukuku) you shouldn't worry about this problem too much (see Distance Vector routing). However, let's do a simple counting. Worst case: you have n nodes that join at the same time. Then each node creates an ETP. In total there are n ETPs flying around. So if you have 256 nodes, you'll get 256 different ETP. Now, since the topology is hierarchical, you must count in each level. So, suppose you are a node. In the level 1 there 256 nodes (including you). So here you see 256 ETPs flying around. In the level 2 there 256 gnodes (including your). The ETP treats each gnode as a single node. So, even in level 2 you see only 256 different ETPs. The same for the other levels. So, if you have 4 levels, then 256 + 256 + 256 + 256 = 1024

DART and Netsukuku comparison

Alpt — 2008-11-21T13:49:33

_______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku

Re: Another question.

Matteo — 2008-11-20T20:49:10

_______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku

Re: Another question.

Alpt — 2008-11-20T20:29:18

On Thu, Nov 20, 2008 at 08:37:00PM +0100, : ~> Ok I think I've found it in the svn... Let's try it! WARNING: the q2sim.py doesn't simulate the ETPs. It was developed before them. It simulates the CTPs. The CTPs aren't utilised anymore. You can find detailed information about CTPs only in the old versions of qspn.pdf. Probably, you've found the link to q2sim.py in the Reference section of the current qspn.pdf. You can ignore it. I'll answer to the other questions later.

Re: Another question.

Matteo — 2008-11-20T19:37:00

_______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku

Re: Another question.

Matteo — 2008-11-20T19:34:43

_______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku

Re: Another question.

Matteo — 2008-11-20T18:50:54

_______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku

Re: Another question.

xDxD — 2008-11-19T21:16:35

Matteo ha scritto: hello there! i contacted alpt a while ago, as i'm starting to develop a couple of things on netsukuku. (for alpt: i have not disappeared: i've just got my personal financial crisis, and i'm getting over it: as soon as i have something to eat, i'll have time to develop and share some of the stuff i obscurely told you about in the email) while starting up the development i also started to get the docs organized a bit, figuring out it could be useful for some other people as well. i'll be happy to help out in getting the docs together on a wiki. let's coordinate on this. bye! xDxD _______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku

Re: Another question.

Alpt — 2008-11-19T21:02:52

On Wed, Nov 19, 2008 at 09:50:08PM +0100, : ~> In all this discussion I saw so much "U should read this", "u should read ~> that" and so on, [...] ~> so many people asking for undocumented ~> stuff and confused about where to look for... "You should read this here" is a proof that what people are asking _is_ documented ;) There's a general index here: http://netsukuku.freaknet.org/index.php?pag=documentation&file=main_documentation/netsukuku It should help navigate the docs. (It is accessible from Documentation->Main documentation in the site). Then here we are to help. The answers to the undocumented requests will be inserted in the docs. (f.e., see http://dev.hinezumi.org/changeset/1458 )

Re: Another question.

Matteo — 2008-11-19T20:59:54

_______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku

Re: Another question.

asbesto — 2008-11-19T20:50:08

In all this discussion I saw so much "U should read this", "u should read that" and so on, that I think documention is too sparse, confuse and somewhere obscure and incomplete: so many people asking for undocumented stuff and confused about where to look for... also, part of the docs are text files, other are in the wiki. I think is better to join docs in the wiki, it helps their development :)

Re: Another question.

Alpt — 2008-11-19T19:50:08

On Wed, Nov 19, 2008 at 04:49:42PM +0100, : ~> I've already read it, but now I got it. ~> ~> What about the counter node? It's a fisical node or is more simply more like ~> a "virtual" one? You should read ntk_rfc/Ntk_p2p_over_ntk.pdf . It is an implemented NTK_RFC (in pyntk). With p2p_over_ntk it is very easy to describe ANDNA: """ 2. As a further example consider ANDNA[6]. In ANDNA there are two different P2P services in one: the service that stores the IPs and the service of the counter nodes, that stores the number of hostnames registered by a node. """ Btw, a counter node is as physical as it is a hash node.

Re: Another question.

Matteo — 2008-11-19T19:47:54

_______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku

Re: Another question.

Alpt — 2008-11-19T19:44:11

On Wed, Nov 19, 2008 at 08:22:09PM +0100, : ~> You say: "quality of the link A<-->B changed, so it's time to rediscover new ~> optimal routes ... ETP send..." ok I undestand it, but I can't undestand how ~> you discover that the wireless link from A to B is changed in quality..... ~> (Maybe you launch an ETP packet every X time?) Nope, we use the Radar. If I recall correctly the Radar isn't described in details in any .pdf. However it is very simple: The radar sends in broadcast a bouquet of BQUET_NUM packets and waits for the reply of the alive nodes. It then recollects the replies and builds a small statistic. By watching the previous statistics, it can deduces if a change occurred or not. If it is, one of the following events is fired: 'NEIGH_NEW', 'NEIGH_DELETED', 'NEIGH_REM_CHGED' In this way, the other modules of pyntk will be noticed. A radar is fired periodically by Radar.run(), which is started as a microthread. Note: the current statistic is based on the RT

Re: Another question.

Matteo — 2008-11-19T19:22:09

_______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku

Re: Another question.

Matteo — 2008-11-19T15:49:42

_______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku

Re: Another question.

Ricardo Lanziano — 2008-11-19T03:14:02

You might want to check topology.pdf[1], specially section 7. Hooking phase. The whole document gives you an insight of how the topology works and why is it structured as it is, explaining how the route map is composed and the benefits of having a hierarchical topology and the relationship with the current Internet addressing system (IPv4, IPv6), the different levels of organization of gnodes, and such. [1] http://netsukuku.freaknet.org/2html/documentation/main_documentation/topology.pdf

Re: Another question.

Matteo — 2008-11-18T22:23:07

_______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku

Re: Another question.

Alpt — 2008-11-18T22:17:05

On Tue, Nov 18, 2008 at 05:35:00PM +0100, : ~> I need to know, how can I obtain an ID numeber when I join a network? Are you talking about the IP? See http://permalink.gmane.org/gmane.network.peer-to-peer.netsukuku/1386 ~> Because I'm studing ANDNA, and when you made the registration for the ~> hostname of a new node, you have alreafy the ID (ok, it's right) but how do ~> you obtain it? What ID are you referring to? Can you give more details?

Re: Another question.

Alpt — 2008-11-18T22:10:43

On Mon, Nov 17, 2008 at 06:49:02PM +0100, : ~> When I start an ETP packets? is there always an ETP that is flying around or ~> I start a message like it only in hooking phase? ~> Maybe I can send an ETP every X time? In qspn.pdf 4.1 Extended Tracer Packet you can see all the cases: 1. Changed link 2. New link 3. Broken link 4. A new node joins 5. A node dies ~> If the node X goes down, how can I notice that? (if I'm a neibourg, and if ~> I'm a node in other gnode) Yes, the neighbours of X notice its death and act according rule 5. (In detail, the neighbours use the Radar to check their links status)

Re: Another question.

Matteo — 2008-11-18T16:35:00

_______________________________________________ Netsukuku mailing list Netsukuku-pma9X3FYtpzZ+VzJOa5vwg< at >public.gmane.org http://lists.dyne.org/mailman/listinfo/netsukuku