Ad-Hoc Routing Protocols (Characteristics)
Classification of the Characteristics of Ad-Hoc Protocols and Routing-Strategies
Extension of
Lang2003
and
Murthy2004
.
In general, routing can be divided into two strategies:
Adaptive Routing
vs.
Not-Adaptive Routing
Adaptive
: Changes of the network-topology are adapted by the routing-strategy
Not-Adaptive
: The routing is done using fixed tables
For Ad-Hoc-Netorks, only adaptiv strategies are usefull.
Reactive Routing
/
On-Demand Routing
vs.
Proactive Routing
/
Table-Driven Routing
vs.
Hybrid Routing
Reactive/On-Demand
A route is only calculated, when it is needed
Does not try to keep routing-information everytime to all node
Proactive/Table-Driven
Routes are calculated before one is needed
Tries to keep routing-information to all nodes everytime up-to-date
Update of the tables:
Event-driven: only if a change is recoginzed
Periodically
Hybrid
Reactive and Proactive at the same time
E.g.: Intra-Zone: Proactiv, Inter-Zone: Reactiv
Distance-Vector Routing
vs.
Link-State Routing
Distance-Vector
Calculates the distance to all nodes
Exchange of these information only with the neighbours
Link-State
Mesuare the distance to the neighbours
Exchange of these information with all nodes
Flat Routing
vs.
Hierarchical Routing
/
Clustered Routing
Hierarchical Routing /
Clustered Routing
:
Trying to structure/cluster the network
Clusterhead
:
Responsible for the creation and extension of a cluster
Builds up a hierarchie of clusters
Manages the communication inside a cluster
Gateway-Node
:
Responsible for the communication between clusters
Maybe bottleneck
Flat Routing
:
Network has no hierarchy
Geographical Routing
/
Positionbased Routing
/
Direction-Based Routing
No routing-tables
Information is send in any way in the direction of the destination
No overhead to find or update routes
But:
Position required
Determination of the position via
internal search-process
external service
Uniform
vs.
Non-Uniform
Uniform
: All nodes are equal
Non-Uniform
: Some nodes have special roles, e.g.
Clusterhead
,
Gateway-Node
Full
vs.
Reduced Topology Information
Full
: All topology-information will be distributed
Reduced
: Only a fraction of the known topology-information will be distributed
Past History
vs.
Prediction
Past History
: Information of past statuses is used to make a decision
Prediction
: Expectation of future statuses are used to make routing decisions
Broadcast
Local Multicast
: Only to some nodes in transmission-distance
Local Broadcast
: Only to the nodes in transmission-distance
Networkwide Broadcast
:
Flooding
Restricted Networkwide Broadcast
: Flooding with a time-to-live
Recovery Strategy
Mechansims to keep or to restore routes
Link-Reversal Routing
Does not try to find an (somehow) optimal way
Only tries to find any way
Source-Routing
The sender specifies the way to go
Under the circumstances, a node inbetween can decide to redefine the way
Route-Selection-Strategies
Power-Aware Routing
Signal Strength
Link Stability
Shortest Path
Link-State Routing
/
Distance-Vector Routing
Direction-Based Routing
/
Positionbased Routing
/
Geographical Routing
Link-Reversal Routing
Multipath Routing
(C) 2004-2006 University of Luxembourg, SECAN-Lab
Original Version
VeryQuickWiki - HTML Export - Printable Version
Version: 2.7.1 (UniLux: 1.15.0 2006-01-19)
Modified: 2006-01-03 12:29:50
Exported: 2010-03-14 02:38:25