headers
Komяpa | 13 Aug 2008 11:33
Picon

ntk and thoughts about it.

Hi all.

I'd like to know something about ntkd development status. Is python 
implementation working at the moment? Will it have all the features of old C 
ntkd? Are there plans to finish / stabilize the code / protocol by certain 
date? 

Some thoughts about ntk itself:

 - It lacks support for messengers. NTK network in first will fall apart very 
frequently (AP switched off, dead internet connection for Viphilama). It could 
leave a region without an irc/jabber/mail server. In my opinion it's possible 
to write distributed irc/mail server implementation that will be able to 
survive after LHC startup. :)

 - Windows support. I know that linux is better, but here in Belarus m$win is 
installed on 90% of computers. Something should be done. :)

 - APs. Old C implementation was able to run on APs. New one depends on 
stackless python. There's only 1 mb of flash and 1-2 mb of free ram available 
on most APs, which is defenetely not enough. That's why I think that routing 
core should be rewritten in C.

Thanks for attention. Waiting for comments. :)

Kom komzpa@... http://komzpa.net

_______________________________________________
Netsukuku mailing list
Netsukuku@...
(Continue reading)

Permalink | Reply | 8 comments |
headers
Ricardo Lanziano | 13 Aug 2008 15:38
Picon
Gravatar

Re: ntk and thoughts about it.

On Wed, Aug 13, 2008 at 4:33 AM, Komяpa <komzpa <at> gmail.com> wrote:
> I'd like to know something about ntkd development status. Is python
> implementation working at the moment? Will it have all the features of old C
> ntkd? Are there plans to finish / stabilize the code / protocol by certain
> date?

That's been an old question since very long time ago: we don't know.

It depends entirely to the free time the devs have, and there is only a few
bits we can do about it:

 * Either code and submit patches
 * Or help with everything that isn't code (documentation, web site, mailing
   lists, IRC, but mostly documentation)

pyntk is evolving, so it will have the same features that the old C
implementation
had and more. In fact, the features are documented in the main_doc category.

>  - It lacks support for messengers. NTK network in first will fall apart very
> frequently (AP switched off, dead internet connection for Viphilama). It could
> leave a region without an irc/jabber/mail server. In my opinion it's possible
> to write distributed irc/mail server implementation that will be able to
> survive after LHC startup. :)

The p2p bits are already on the code, it just needs to be finished. However,
there is also a project to bring decent p2p bindings that use the ntk network
at it's glance, but it will come later when the new pyntk code stabilizes.

IMO, it will be a good idea, but I rather see pyntk working first :>
(Continue reading)

Permalink | Reply | 5 comments |
headers
Eriol | 13 Aug 2008 17:36
Favicon
Gravatar

Re: ntk and thoughts about it.

On Wednesday 13 August 2008 11:33:26 Komяpa wrote:
> ntkd? Are there plans to finish / stabilize the code / protocol by
> certain date?

On Wednesday 13 August 2008 15:38:12 Ricardo Lanziano wrote:
> That's been an old question since very long time ago: we don't know.
>
> It depends entirely to the free time the devs have

I have a good news: I'll work on pyntk for a class exam of mine, so I'll can 
spend much more time on it! :)

Right now I'm studying for another exam, but optimistically I'll start 
coding at the beginning of October.

--

-- 
 Eriol - *p = NULL; - EIBTI
 GPG Key ID 0B7C8A19
 http://mornie.org
_______________________________________________
Netsukuku mailing list
Netsukuku <at> lists.dyne.org
http://lists.dyne.org/mailman/listinfo/netsukuku
Permalink | Reply | 2 comments |
headers
Komяpa | 13 Aug 2008 19:19
Picon

Re: ntk and thoughts about it.

On 13 August 2008 16:38:12 Ricardo Lanziano wrote:
> It depends entirely to the free time the devs have, and there is only a few
> bits we can do about it:
>
>  * Either code and submit patches
Ok. I'll try to, but I need at least something like a plan what to do first :)

>  * Or help with everything that isn't code (documentation, web site,
> mailing lists, IRC, but mostly documentation)
What are the problems with documentation?

> "Q: Will Netsukuku be ported to Windows?
> A: Short answer: if you code the port, yes.
>    Answer: We need coders for that. There are a lot of things to be done
> and the Windows port is what we care less."
So, nothing changed since October of 2007? 

> >  - APs. Old C implementation was able to run on APs. New one depends on
> > stackless python. There's only 1 mb of flash and 1-2 mb of free ram
> > available on most APs, which is defenetely not enough. That's why I think
> > that routing core should be rewritten in C.
>
> You can get a stackless python binary under the 1mb limit. You just need to
> strip off everything you don't need from the original distribution.
Even if it is possible to make a distribution of stackless in 700kb (we need 
~250-300 kb for ntkd itself), i think we will be out of RAM. At the moment my 
htop shows me that pyntk listening to dummy interface uses 7 mb of ram. And 
there are 8 on devices at all in most cases. 

_______________________________________________
(Continue reading)

Permalink | Reply | 1 comment |
headers
Mike Zink | 13 Aug 2008 19:38
Picon
Gravatar

Re: ntk and thoughts about it.



On Wed, Aug 13, 2008 at 10:19 AM, Komяpa <komzpa <at> gmail.com> wrote:
> >  - APs. Old C implementation was able to run on APs. New one depends on
> > stackless python. There's only 1 mb of flash and 1-2 mb of free ram
> > available on most APs, which is defenetely not enough. That's why I think
> > that routing core should be rewritten in C.
>
> You can get a stackless python binary under the 1mb limit. You just need to
> strip off everything you don't need from the original distribution.
Even if it is possible to make a distribution of stackless in 700kb (we need
~250-300 kb for ntkd itself), i think we will be out of RAM. At the moment my
htop shows me that pyntk listening to dummy interface uses 7 mb of ram. And
there are 8 on devices at all in most cases.

Python may be a bit too resource heavy at the moment, but on the other hand ntk is an extremely new technology, and by the time it takes off you can bet that most APs will have more than enough resources to run stackless python and pyntk. Even now, nearly all newer APs have upwards of 4 MB flash memory and 8 MB RAM, and those numbers are just going up. I really don't think this will be a problem by the time ntk gains popularity.

--
- Mike Zink
_______________________________________________
Netsukuku mailing list
Netsukuku@...
http://lists.dyne.org/mailman/listinfo/netsukuku
Permalink | Reply | 0 comments |
headers
ZioPRoTo (Saverio Proto | 13 Aug 2008 19:47
Picon
Gravatar

Re: ntk and thoughts about it.

> I'd like to know something about ntkd development status. Is python
> implementation working at the moment?

It is not working
I opened this ticket long time ago: http://dev.hinezumi.org/ticket/1

But ticket is still there :(

>  - It lacks support for messengers. NTK network in first will fall apart very
> frequently (AP switched off, dead internet connection for Viphilama). It could
> leave a region without an irc/jabber/mail server. In my opinion it's possible
> to write distributed irc/mail server implementation that will be able to
> survive after LHC startup. :)

I'm sorry but your statement is confused.
Netsukuku aims to build a distributed network. Now I can easily
understand that server-based services are not well suited in a
spontaneous distributed network architecture, but that's another
story.

Maybe what you are trying to say is that when netsukuku will be
working there will be a need for distributed applications ?

Saverio Proto
_______________________________________________
Netsukuku mailing list
Netsukuku@...
http://lists.dyne.org/mailman/listinfo/netsukuku
Permalink | Reply | 1 comment |
headers
ZioPRoTo (Saverio Proto | 13 Aug 2008 19:50
Picon
Gravatar

Re: ntk and thoughts about it.

> I have a good news: I'll work on pyntk for a class exam of mine, so I'll can
> spend much more time on it! :)

Once again, before coding towards nowhere ... may I suggest to have a
deep look into DART ?

http://dart.cs.ucr.edu/

I already pointed this out, and also AlpT said it has a lot in common
with Netsukuku.

Maybe some good idea (or optimization) may come out from a deep look.
Maybe they have already some implementation ??

Saverio Proto
_______________________________________________
Netsukuku mailing list
Netsukuku@...
http://lists.dyne.org/mailman/listinfo/netsukuku
Permalink | Reply | 1 comment |
headers
soreiser | 13 Aug 2008 21:05
Picon
Favicon

a help for the english Wikipedia's page

Hi everybody,
today i noticed that the english Wikipedia's page about netsukuku
reports an advice to fix or improve the page, preventing possible
deletion, 'cause does not meet guidelines or something.
Practically, it says that is not so clear that netsukuku is an
EXPERIMENTAL network (?! it's explained in the -first- line of the page!
Bah...), and suggests to make changes, add/merge/improve some things.
I'll be happy if someone could help me out to fix that problem, 'cause i
am not so expert with editing Wikipedia "in a correct way", so maybe
someone could do it for me PLEASE or helping me with that.
Here's the link: http://en.wikipedia.org/wiki/Netsukuku
greetings.

_______________________________________________
Netsukuku mailing list
Netsukuku@...
http://lists.dyne.org/mailman/listinfo/netsukuku
Permalink | Reply | 10 comments |
headers
Ricardo Lanziano | 13 Aug 2008 22:22
Picon
Gravatar

Re: ntk and thoughts about it.

On Wed, Aug 13, 2008 at 12:50 PM, ZioPRoTo (Saverio Proto)
<zioproto@...> wrote:
>> I have a good news: I'll work on pyntk for a class exam of mine, so I'll can
>> spend much more time on it! :)
>
> Once again, before coding towards nowhere ... may I suggest to have a
> deep look into DART ?
>
> http://dart.cs.ucr.edu/
>
> I already pointed this out, and also AlpT said it has a lot in common
> with Netsukuku.
>
> Maybe some good idea (or optimization) may come out from a deep look.
> Maybe they have already some implementation ??

Well, it might be a good idea; new concepts might come up and probably
ntk could learn something. But it's also true that Netsukuku by definition
evolves, so can the theory. Maybe is something that we might need to
change in the future?

Pretty much justifies the effort, B.A.T.M.A.N did it this way. The theory
evolved as so did the implementation, now they are in the thirth version.

--

-- 
Ricardo Lanziano
6838 E880 7CA4 D4B6 8283 8F65 96AD EC0B F113 AB62
Unix is simple. It just takes a genius to understand its simplicity.
_______________________________________________
Netsukuku mailing list
Netsukuku@...
http://lists.dyne.org/mailman/listinfo/netsukuku
Permalink | Reply | 0 comments |
headers
Ricardo Lanziano | 13 Aug 2008 23:36
Picon
Gravatar

Re: a help for the english Wikipedia's page

On Wed, Aug 13, 2008 at 2:05 PM, soreiser <inspheration@...> wrote:
> Hi everybody,
> today i noticed that the english Wikipedia's page about netsukuku
> reports an advice to fix or improve the page, preventing possible
> deletion, 'cause does not meet guidelines or something.
> Practically, it says that is not so clear that netsukuku is an
> EXPERIMENTAL network (?! it's explained in the -first- line of the page!
> Bah...), and suggests to make changes, add/merge/improve some things.
> I'll be happy if someone could help me out to fix that problem, 'cause i
> am not so expert with editing Wikipedia "in a correct way", so maybe
> someone could do it for me PLEASE or helping me with that.
> Here's the link: http://en.wikipedia.org/wiki/Netsukuku
> greetings.

Thanks for letting us know. I've fixed it a bit and cleaned a bit of the
latest misconception about (ab)use of terms in the media and added
a bit more of real information. My battery won't last long so I will update
the page later.

Btw, what's the current position about ntk in the ninux community?, Just
for a few shortnotes on the wikipedia.

--

-- 
Ricardo Lanziano
6838 E880 7CA4 D4B6 8283 8F65 96AD EC0B F113 AB62
Unix is simple. It just takes a genius to understand its simplicity.
_______________________________________________
Netsukuku mailing list
Netsukuku@...
http://lists.dyne.org/mailman/listinfo/netsukuku
Permalink | Reply | 9 comments |

Gmane