[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [manet] AODV Question

Hello everybody,

I tried to ignore this thread from the beginning because I have better things to do. But now I cannot help it. ;)

1) IPv4 headers have a 8-bit field called TTL (Time To Live).
("TTL (time to live): each router must decrement by 1 and discard if zero. Also decrement if router holds for longer than 1 sec.")

2) IPv6 headers have a 8-bit field called Hop Limit.
("hop limit (8 bit): Decremented by one by each node that forwards the packet. When the hop limit field reaches zero, the packet is discarded.")

3) AODV is a UDP protocol. So The AODV message goes wrapped in an IP header. Even if AODV message would have a bigger hop count, the IP packet has only a 8-bit field. Therefore, it will die after the 255 hops.

4) Point number 3 is not a big problem since I have seen some AODV implementations that reset IP's TTL or Hop Limit to 255 at each hop. Anyway, once the route is constructed, when you would send the data, the IP packets would die before reaching the final destination. (Yep, at the 255 hops :) ) (And there is no much point in to constructing a route that cannot be used ;) ).


Regards,

Manel Guerrero Zapata

P.D.: This is my new email address. The previous one:
(Manel.Guerrero-Zapata@nokia.com) does not work anymore.


Jitesh Shah wrote:
Hi Daniel,

I completely agree with you when you suggest that one
must take care of the future needs when we are in the
development stage.
We might have to face a situation in the real life
when we might actually need more than a byte. The
things being in development stage, I believe perhaps
this is the right time to incorporate the changes
rather than evolving with AODV 1.1 at a later stage
:-)

Rgds
Jitesh



--- DANIEL BYRNE <daniel.byrne@adtran.com> wrote: >
Agreed. A branched chain configuration would be

difficult to support with only a BYTE. My point is
that although most everyone involved in development
and testing of this protocol have envisioned a
'cluster of connected nodes' arrangement, the
reality is the sum total of all possible topolgies
in a wireless network of this type is potentially
limitless. Limited in effect by only the range of
the radios in the network. How are we to know how
some future person or organization will use this
protocol? Will they need the extra BYTE? Should we
place artificial limits on the protocol by
purposefully limiting the hop count, address range,
etc.?

One solution is the "extensions" field in the latest
draft of AODV could potentially provide a means to
expand upon the protocol in the future if
neccessary. 
http://moment.cs.ucsb.edu/pub/draft-perkins-manet-aodvbis-00.txt



-----Original Message-----
From: Jitesh Shah [mailto:jiteshshahin@yahoo.com]
Sent: Thursday, November 13, 2003 7:36 AM
To: manet@ulfius.com; manet@ietf.org
Subject: RE: [manet] AODV Question


Hi,

It depends on how you see the network. If you are
refering to a network which has a CHAIN topology
then
the byte used for RREP and RREQ might be way
insignificant. But when you refer to a well
connected
network then the byte seems sufficient.

Rgds
Jitesh

--- Robert Cain <manet@ulfius.com> wrote: >
Connectivity is the number of other nodes each node


can reach with 1 hop.
Therefore a connectivity of 10 means there is 10
nodes that a can reach
directly. Obviously in an actual network this

would


vary from node to node.


Rob Cain :-)




What does the connectivity parameter specify?  Is

it saying in 100,000
nodes only 10 nodes would be connected?


-----Original Message-----
From: Robert Cain [mailto:tgm@ulfius.com]
Sent: Wednesday, November 12, 2003 11:13 AM
To: DANIEL BYRNE
Subject: Re: [manet] AODV Question


1 byte seems OK to me.

With some simple analysis assuming a circular

homogeneous network of node


count K and a reasonable connectivity C then the

radius of the network in


hops can be derived to be the square root of K/C.

Which for a network size


of 100,000 and a connectivity of 10 gives a hop

count radius of 100.
Seeing


as a network of size 100,000 will be probably at

practical limits of size


due to performance degradation then I think 1

byte


may well be enough.


hope that helps :-)

Rob Cain
if the
"AODV routing protocol is designed for mobile ad
hoc networks


 with populations of tens to thousands of

mobile


nodes"
why is the HopCount limited to only a BYTE? 
This


effectively limits the


number of hops to 255.  Why not expand the hop

count to two bytes into the


reseved region of the RREQ packet.  Something

similar could be done with


the RREP packet as well.


_______________________________________________
manet mailing list
manet@ietf.org
https://www1.ietf.org/mailman/listinfo/manet




_______________________________________________

manet mailing list

manet@ietf.org

https://www1.ietf.org/mailman/listinfo/manet 





________________________________________________________________________




Yahoo! India Mobile: Download the latest polyphonic

ringtones.

Go to http://in.mobile.yahoo.com



_______________________________________________

manet mailing list

manet@ietf.org

https://www1.ietf.org/mailman/listinfo/manet 


________________________________________________________________________
Yahoo! India Mobile: Download the latest polyphonic ringtones.
Go to http://in.mobile.yahoo.com

_______________________________________________
manet mailing list
manet@ietf.org
https://www1.ietf.org/mailman/listinfo/manet





_______________________________________________
manet mailing list
manet@ietf.org
https://www1.ietf.org/mailman/listinfo/manet