Networking Reference
In-Depth Information
Reactive route discovery avoids the drawback of the proactive one by avoiding
exchanging periodic routing updates, which reduces the traffic overhead.
However, as the path is discovered only on-demand, this type of route discovery
incurs a delay overhead and longer latency for route establishment.
The category of hybrid route discovery is obtained by combining both the
proactive and reactive techniques to make use of the advantages of both and
mitigate their disadvantages. It tries to reduce the control overhead associated
with proactive route discovery and the delay incurred in the reactive one. Route Selection
As an output of the route discovery stage, there will be a set of potential routes
between a source and destination. It is the role of the route selection component to
pick the optimal path from this set that satisfies the needed performance criteria.
Most of the routing protocols are based on choosing only one path for delivering
packets from a specific source to a specific destination; however, there are some
protocols that rely on choosing multiple paths (multipath routing) [ 4 ] to provide
redundancy and fault tolerance for the routing process.
For the proactive protocols, route selection is done implicitly with the route
discovery stage. When the network topology information is shared and received by
the nodes, they update the information in their routing tables accordingly; hence,
routes available in the routing tables are the selected, best ones at that time.
Route selection in the reactive protocols is a stand-alone process. It can be
handled by the source, the destination, or the intermediate nodes. In destination-
based route selection, when the destination receives multiple route requests for-
warded by multiple intermediate nodes, it can select the path to receive data
through and sends the route reply along this path. The destination can pick the first
path through which it received the first route request, the fastest one, or it can wait
for a specific period of time. Thereafter, if it has received many route requests, it
can pick the optimal path according to some selection metrics, discussed later in
this section.
In source-based route selection, the source node may receive multiple route
replies from the destination, 2 or from all intermediate nodes that know about a
route to the intended destination. It is the responsibility of the source to pick a
route from the set of routes extracted from the multiple route replies.
For intermediate-based route selection, the intermediate nodes decide on which
route a packet should follow to reach a destination. They can either choose a route
from a set of possible routes they keep for that destination, 3 or select a next hop on
the fly. This type of route selection is involved in self-routing protocols. Since the
2 The destination may reply to all route requests it receives.
3 These routes may be discovered by them in a previous interaction with the destination or
overheard from neighbors interacting with that destination.
Search MirCeyron ::

Custom Search