Re: [manet] DSR and unidirectional links
Sudarshan N Raghavan <snraghav@unity.ncsu.edu> Sun, 11 August 2002 23:01 UTC
Received: from optimus.ietf.org (ietf.org [132.151.1.19] (may be forged)) by ietf.org (8.9.1a/8.9.1a) with ESMTP id TAA23359 for <manet-archive@odin.ietf.org>; Sun, 11 Aug 2002 19:01:14 -0400 (EDT)
Received: from optimus.ietf.org (localhost [127.0.0.1]) by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA21332; Sun, 11 Aug 2002 18:46:41 -0400 (EDT)
Received: from ietf.org (odin [132.151.1.176]) by optimus.ietf.org (8.9.1a/8.9.1) with ESMTP id SAA21305 for <manet@optimus.ietf.org>; Sun, 11 Aug 2002 18:46:38 -0400 (EDT)
Received: from uni10mr.unity.ncsu.edu (uni10mr.unity.ncsu.edu [152.1.1.35]) by ietf.org (8.9.1a/8.9.1a) with ESMTP id SAA23085 for <manet@ietf.org>; Sun, 11 Aug 2002 18:45:20 -0400 (EDT)
Received: from unity.ncsu.edu (uni01wi.unity.ncsu.edu [152.1.1.31]) by uni10mr.unity.ncsu.edu (8.11.6/8.11.6/N.20020613.01) with ESMTP id g7BMkZl10658; Sun, 11 Aug 2002 18:46:35 -0400 (EDT)
Message-Id: <200208112246.g7BMkZl10658@uni10mr.unity.ncsu.edu>
Content-Transfer-Encoding: 8bit
User-Agent: IMHO/0.97.1 (Webmail for Roxen)
Subject: Re: [manet] DSR and unidirectional links
From: Sudarshan N Raghavan <snraghav@unity.ncsu.edu>
Cc: manet@ietf.org
Date: Sun, 11 Aug 2002 18:47:06 -0500
To: Noah.H.Miller@Dartmouth.EDU
Content-Type: text/plain; charset="iso-8859-1"
MIME-Version: 1.0
X-Originating-IP: [24.98.217.30]
In-Reply-To: <42127439@dancer.Dartmouth.EDU>
Content-Transfer-Encoding: 8bit
Sender: manet-admin@ietf.org
Errors-To: manet-admin@ietf.org
X-Mailman-Version: 1.0
Precedence: bulk
List-Id: Mobile Ad-hoc Networks <manet.ietf.org>
X-BeenThere: manet@ietf.org
Content-Transfer-Encoding: 8bit
My question was not "how DSR handles unidirectional links" but "how
DSR discovers unidirectional links". There is no auto-discovery of
unidirectional links in DSR and hence it has to be "configured" for
the possibility of unidirectional links. If it is configured for
unidirectional links, a single RREQ has the possibility of trigerring
a "RREQ storm" (in addition to RREP storms) from intermediate nodes. I
want to know whether there have been any simulations to analyze DSR's
performance when configured for unidirectional links. Most of the
simulations seem to have been done over 802.11 type links that do not
take unidirectional links into account.
Regards,
Sudarshan N Raghavan
Graduate Student, Computer Sc.
North Carolina State University
>>DSR Internet Draft 7
(http://www.ietf.org/internet-drafts/draft-ietf-manet-dsr-07.txt)
specifies how it handles unidirectional links.
>>
>>To configure DSR to allow unidirectional links, you must use the
DSR-level ACK REQ and ACK options for route maintenance, rather than
MAC layer acknowledgements. 802.11 won't allow unidirectional links,
but I suppose other radio links will.
>>
>>given the network, where arrows show directionality:
>>
>>A -> B <-> C
>> | |
>><-D<-
>>
>>For route discovery:
>>node A broadcasts an RREQ, and node B receives it. Note that NO
acknowledgement is necessary since the RREQ is a broadcast packet.
When the target C of the RREQ replies, it will piggyback its RREP on a
new RREQ back to the initiator, node A, of the route discovery. Two
routes get set up:
>>A->B<->C and
>>C<->B->D->A
>>
>>For Source Routed data transmission:
>>Node A unicasts a packet to node B. The network has been configured
for the possibility of the link A->B being unidirectional, so it sends
a DSR-level ACK either along a cached route back to A (such as
B->D->A), or initiates route discovery back to A. A thus receives
next-hop confirmation through an indirect route.
>>
>>
>>Hope that clarifies,
>>Noah Miller
>>ISTS Dartmouth College
>>
>>
>>--- Sudarshan N Raghavan wrote:
>>DSR is supposed to work over unidirectional links. But how does DSR
>>discover unidirectional links ?? Do we need to explicitly configure
>>DSR to work over uni/bi directional links ??
>>
>>Assume DSR is configured for uni-directional links:
>>1. A single RREQ broadcast can trigger multiple RREQ broadcasts from
>>intermediate nodes that have a route to the destination in their
cache
>>(assuming that the nodes are not able to find out that the source
has
>>already chosen a RREP)
>>
>>2. DSR expects each node along the source route to get an ack from
the
>>next downstream node. If the links are unidirectional and the nodes
do
>>not have a route to their upstream neighbor along the source route,
it
>>can trigger N RREQ broadcats for a source route of length N
>>
>>3. The same is the case with route maintainance if the node
reporting
>>a link failure does not have a route to the source
>>
>>This seems like an enormous number of broadcasts which has the
>>potential to bring down the network if there is lot of network
>>activity.
>>
>>Have there been any studies/simulations to analyze chain reactions
and
>>the performance of DSR when configured for unidirectional links ?
>>
>>Regards,
>>Sudarshan
>>
>>****
>>Sudarshan N Raghavan
>>Graduate Student, Computer Sc.
>>North Carolina State University
>>--- end of quote ---
>>
_______________________________________________
manet mailing list
manet@ietf.org
https://www1.ietf.org/mailman/listinfo/manet
- Re: [manet] DSR and unidirectional links Noah H. Miller
- Re: [manet] DSR and unidirectional links Sudarshan N Raghavan
- [manet] Terminal node acks Sudarshan N Raghavan