1. Zone Routing Protocol (ZRP)
Traditional Routing Schemes
Motivation
The Zone Routing Protocol
Query Control Mechanisms
Conclusions

2. Traditional Routing Schemes
Proactive Routing
• Maintains an up-to-date view of the network
• Example protocols: Distance Vector, Link State
Reactive Routing
• Acquire routes on demand
• Example protocols: TORA, AODV

3. Motivation The disadvantages of both traditional routings.
As seen, the amount of update traffic is quite high in proactive routing.
In reactive routing, route query is delayed because route information may not be available
ZRP aims to address the problems by combining the best properties of both approaches.

4. The Zone Routing Protocol
Based on the concept of zones
A zone is defined for each node separately
• Zone radius r given as number of hops
• The zones overlap
Peripheral nodes
G, H, J, and I
Neighbor nodes
B, C, ….
We depict zones as circles, but they are not.
Center of S-zone
r = 2 (hops)

5. The Zone Routing Protocol
Based on the concept of zones
A zone is defined for each node separately
• Zone radius r given as number of hops
• The zones overlap
Peripheral nodes
G, H, J, and I
Neighbor nodes
B, C, ….
We depict zones as circles, but they are not.
Center of E-zone
r = 2 (hops)

6. Proactive Intra-zone Routing
ZRP refers to the locally proactive routing component as IARP
IARP is not a specific routing protocols. Instead, IARP is a family of proactive routing protocols.
Each node maintains the routing info to all nodes in its zone
The updates are only local in the zone.
Reduces the maintenance costs to a limited zone

7. Reactive Inter-zone Routing
The globally reactive routing component is named IERP. IERP is responsible for discovering routes to destinations beyond a node’s routing zone.
For example, when S has a packet to send to x, IERP is responsible for discovering this route.
How does IERP work? I will explain in the next few slices.

8. Interzone Routing (cont.)
IERP has two phases for route discovery:
Route request
Route reply
IERP don’t use standard flooding search.
IERP border-casts a route request.
Border-casting is a packet delivery service that allows a node to efficiently send a message to its peripheral nodes.
Route query packet is uniquely identified by a combination of the source node’s ID and request number.
Upon receipt of a route query packet, a node adds its ID to the query.

9. Interzone Routing (cont.)
If peripheral nodes don’t find the destination, they continue the border-casting process.
If the destination is found, a route reply is sent back to the source by reversing the accumulated route.

10. A Concrete Example of How IERP works.
S has a packet to send to x
X is not in S’s routing zone. S issues a route request using IERP. The request is bordercasted to the peripheral nodes (the gray ones). These gray nodes search their routing table for the destination. They don’t see the destination. So the border-casting process continues (at node I, J, G , and H).

11. An Example (cont.)
Node I bordercasts the request to its peripheral nodes (gray ones). Due to query control mechanisms, the request is not passed back to nodes D, F, and S. I will discuss query control later.

12. An Example (cont.)
Finally, the route request is received by node T, which can find the destination in its routing zone. A rout reply is sent back to node S.

13. Query Control Mechanisms
• The zones of neighboring nodes overlap
• Route requests forwarded several times
• More traffic than in flooding
Uses three types of query-control mechanisms
• Query detection
• Early termination
• Random query-processing delay

14. Query Detection
To be able to prevent queries from reappearing in covered regions, the node must detect local query relaying activity.
QD1 allows nodes to detect queries as they relay them to the edge of the routing zone.
In a single-channel networks, it is possible for queries to be detected by any node within the transmission range of a relaying node (QD2)
Eavesdropping (QD2)
Ex. B can detect the query using (QD2)
K is unaware of the query

15. Early Termination
Uses information obtained from Query Detection and topological information
Termination of queried entering already covered regions
Node x receives a query message to relay. It takes advantage of its extended zone and QD to identify all Y’s interior nodes as being covered.
Later, X receives a second copy of the query to relay on behalf of border-casting node Z. Again X identifies all Z’s interior nodes as being covered.
X should relay the query to two of Z’s peripheral nodes. However, X recognizes that both peripheral nodes have been covered.

16. Random Query Processing Delay
When a node initiates a border-cast, a node’s routing zone is instantly covered by the query. However, it takes time for the query to make its way along the border-cast tree and detected through QD.
During this time, the zone is vulnerable to query overlap from nearby nodes. To prevent this problem, RQPD can be employed.

17. Example
Node X and Y share common neighbors. Assuming that X and Y will receive the route query at approximately the same time. Without RQPD, X and Y will both proceed to bordercast. Later they determine that their bordercasts were redundant.
With RQPD, X and Y each backs off a random period of time. So RQPD allows one of them to detect other as being covered.

18. Conclusions
ZRP provides a flexible solution to challenge of discovering and maintaining routes. It combines two different methods of routing in one protocol.
ZRP reduces the traffic amount compared to pure proactive or reactive routing. The amount of intra-zone control traffic increases with the zone’s size. However, using the knowledge of the routing zone to reduce the amount of inter-zone control traffic through border-casting.
QD, ET and LT provide significant improvements compared with purely reactive (r = 1) and purely proactive schemes (r ® ¥)
Zone routing is targeted for large networks. The amount of control traffic does not depend on network size because proactive updates are only local.