osid: call for collaborators on protocol design (initially)

osid: call for collaborators on protocol design (initially)

Post by Dan Stromber » Sat, 17 May 2003 06:18:15



I'm thinking about trying to promote a protocol for identifying linux
systems (and perhaps other OSes) over the net.  Discussion of "Why not
just use nmap?" is at the URL below, among other things.  I think some
folks may particularly like the well-known-port I've tentatively suggested
to run it on.  :-)  I mean, as long as they seem to be going out of
business anyway.

The motivation is to come up with a decent lower bound on linux' installed
base.

Here's the URL: http://dcs.nac.uci.edu/~strombrg/osid/

I'm capable of coding this myself (if I can find the time), but n heads
are better than one, and I want to get the protocol designed well right
from the start.

If you're interested, please visit the page and mail me.  I'll collect
addresses into a mailing list if there's enough interest, otherwise maybe
we can just do this as an e-mail thread.

 
 
 

osid: call for collaborators on protocol design (initially)

Post by Jeremy Gra » Sat, 17 May 2003 07:18:44



> I'm thinking about trying to promote a protocol for identifying linux
> systems (and perhaps other OSes) over the net.  Discussion of "Why not
> just use nmap?" is at the URL below, among other things.  I think some
> folks may particularly like the well-known-port I've tentatively suggested
> to run it on.  :-)  I mean, as long as they seem to be going out of
> business anyway.

> The motivation is to come up with a decent lower bound on linux' installed
> base.

> Here's the URL: http://dcs.nac.uci.edu/~strombrg/osid/

Why should anyone voluntarily open a port and run a daemon to respond to
queries about the local OS type?  Why should anyone tell the truth if
he does so?  

Also, the phrase "It should punch through any firewall by default"
probably won't get much support for your protocol.  

--
Jeremy A. Gray

"Remember the Pueblo." -- the Fourth Law of Marvin

 
 
 

osid: call for collaborators on protocol design (initially)

Post by Dan Stromber » Sat, 17 May 2003 07:56:25




>> I'm thinking about trying to promote a protocol for identifying linux
>> systems (and perhaps other OSes) over the net.  Discussion of "Why not
>> just use nmap?" is at the URL below, among other things.  I think some
>> folks may particularly like the well-known-port I've tentatively
>> suggested to run it on.  :-)  I mean, as long as they seem to be going
>> out of business anyway.

>> The motivation is to come up with a decent lower bound on linux'
>> installed base.

>> Here's the URL: http://dcs.nac.uci.edu/~strombrg/osid/

> Why should anyone voluntarily open a port and run a daemon to respond to
> queries about the local OS type?  Why should anyone tell the truth if he
> does so?

Um, to help linux?  To help IT people who want to get a handle on how
important linux already is to them - say so they can start doing bulk
licensing with a linux vendor?

Quote:> Also, the phrase "It should punch through any firewall by default"
> probably won't get much support for your protocol.

No one said you couldn't turn it off.  And even that only means the
personal firewall.  If your border blocks it, your border blocks it.

The info's already there in a form that's good enough for attackers (nmap,
xprobe, queso, scapy, and so on).. This just legitimizes it enough that IT
people and linux PR people can start to use it.

If you're inclined to mess up a statistic that could significantly help
linux in the war with microsoft, I suppose that's you're prerogative.

 
 
 

osid: call for collaborators on protocol design (initially)

Post by Kasper Dupon » Sat, 17 May 2003 18:59:20



> The info's already there in a form that's good enough for attackers (nmap,
> xprobe, queso, scapy, and so on).. This just legitimizes it enough that IT
> people and linux PR people can start to use it.

Netcraft already does:
http://uptime.netcraft.com/up/graph?site=dcs.nac.uci.edu

--
Kasper Dupont -- der bruger for meget tid p? usenet.

for(_=52;_;(_%5)||(_/=5),(_%5)&&(_-=2))putchar(_);

 
 
 

1. calling socket protocol family from another socket protocol family ?

I am developing a new socket protocol family.

When getting into the connect routine I notice
that the request goes to a remote computer and I
would like to get this handled by AF_INET.

I was hoping to be able to call

my_connect_routing:

                struct socket *socktramp;
                int retval;
                retval = sock_create (PF_INET, SOCK_STREAM,
IPPROTO_TCP, &socktramp);
                printk ("creating new socket (%d).\n", retval);

                retval = socktramp->ops->connect (socktramp, uaddr,
addr_len, flags);

                printk ("connected ?!\n");
                release_sock (sk);

The creation of an AF_INET socket seems to work (0).

however, the call to the ->ops->connect function (which
should be the AF_INET family) fails:

   connect(): Address family not supported by protocol

Are there better ways of handling this ?

Thanks,
Mark

2. music theory servers

3. design question - new protocol handler

4. Turn off mail in crontab??

5. designing client server protocols

6. Aha1522 + NEC 2X 210 locks! Why?

7. design question - new protocol handler

8. newbie & y2k

9. Call for Papers - 5th Symposium on Operating Systems Design and Implementation (OSDI 2002)

10. hardware design or digital design

11. Call for Papers - 5th Symposium on Operating Systems Design and Implementation (OSDI 2001)

12. Chip Design on Linux with the OCEAN IC design system

13. Call for Papers - 5th Symposium on Operating Systems Design and Implementation (OSDI 2002)