Dynamic Source Routing (DSR)
In the DSDV, all nodes maintain path to all the other nodes.
Due to this there is heavy traffic.
To save the battery power
DSR is considered. This DSR
divides the routing into 2 sub
problems
Route Discovery : Here a node tries to discover a route to a node (i) Only
if there is information to send (ii) and no routes are known.
Route Maintenance : If a node is continuously using this route to transmit
a packet, then the route should be without problems. But if the node
detects that the route is with problem, then it has to determine alternate
route.
Working Principle : If a node reads a route to a destination, it broadcasts a route
request with a unique identifier and the destination address as parameters.
The node which receives the (Route Request Message) does the following:
1. If the node has already received the request, it drops the request packet.
2. If the node recognizes own address as the destination, the request has reached the
target.
3. Otherwise the node appends its own address in the route, and broadcasts this
updated route request.
Principle :
The route request collects the list of address representing a possible path on
its way towards the destination. When the request reaches the destination, it
can return the request packet to the source.
When the link is bi-directional the route list is sent in the reverse order to the
destination.
When the link is unidirectional the destination does not maintain the route. It
needs to discover the route.
Example to find a route from N1 to N3 at t1,
1. N1 broadcasts the Request ((N1), id=42, target=N3). N2 and N4 receive the
packet.
2. N2 broadcasts ((N1, N2), id=42, target=N3).N4 broadcasts ((N1, N4),
id=42, target=N3). N3 and N5 receive N2’s broadcasts. N1, N2, N5
receives N4’s broadcasts.
3. N3 itself is the target.
4. N5broadcasts ((N1, N2, N5), id=45, target=N3). N3, N4 receive this
broadcast.
5. N1, N2, N5 drop this N4’s broadcast, because it has already received.
6. N4 drops N5’s broadcast.
N3 finds (N1, N2 and N5) as an alternate route but a longer route.
7. N3 has to return the path (N1, N and N3) to N1 . N3 can do the reverse
forwarding because symmetric link is assumed.
8. When the links are uni-directional the algorithm needs to be applied
again with N3 as source and. Nj as destination.
Optimization
1. To avoid too many broad cast route requests should contain a counter. For
every rebroad cast the counter is incremented. When the counter exceeds the
number of nodes in the network, the nodes can drop the request.
2. Node can Cache the path fragments from recent requests. This fragments can
be used to find other route.
3. A node can update the Cache while forwarding the packets.
4. The node can update the Cache when it overhears the transmission from other
nodes.
Maintenance of the Route: When the routes are discovered they need to
be maintained. Approaches to maintain the route are as follows :
1. If the link layer uses acknowledgement this ack can be considered as an interact
route.
2. The node can overhear the next hop which is passive acknowledgement.
3. A node can ask for explicit acknowledge.
When the links are bi-directional, no problem for maintenance. If not the
situation is complicated. If there is connectivity problem, detected by a node, it has
to inform the sender, to find a new route from the sender.
