CCN Activity Session for Section 2.2 – RIP and Gradient Routing
In this exercise we’ll be thinking about routing tables and how they are constructed by protocols such as RIP. We’ll then move on to think about gradient routing.
Activity
Simplified routing tables (no distances to destination) for a network are shown below. All nodes are considered to be reachable. From the tables, reconstruct the logical structure of the network.
A
B
C
D
Destination
Next
Destination
Next
Destination
Next
Destination
Next
B
B
A
A
A
A
A
A
C
C
C
A
B
A
B
A
D
D
D
A
D
A
C
A
E
D
E
A
E
A
E
F
F
D
F
A
F
A
F
F
G
D
G
A
G
A
G
G
H
D
H
A
H
A
H
G
E
F
G
H
Destination
Next
Destination
Next
Destination
Next
Destination
Next
A
F
A
D
A
D
A
G
B
F
B
D
B
D
B
G
C
F
C
D
C
D
C
G
D
F
D
D
D
D
D
G
F
F
E
E
E
H
E
E
G
H
G
G
F
F
F
G
H
H
H
G
H
H
G
G
Ignore weights for this exercise – the structure is what we were looking for
Turn Over
Discuss
What properties of UDP does RIP rely on when discovering the routing table? What is the maximum diameter of a network using RIP? Why?
UDP relies on the UDP packets being broadcast to the network (the route to new destinations is unknown). The maximum number of hops allowed by RIP is 15 (16 is used for “unreachable”)
Discuss
Gradient-based routing approaches to routing are often used in wireless sensor networks with constrained resources. How does the use of available power alter the routes in the gradient-based scheme.
Tiers are initially assigned to nodes based on hops to the sink (normally a link to the backhaul network). Nodes with low energy raise their tier number to reduce routing (fewer tiers above). As charge recovers the node can reduce its tier is reduced back to original level, restoring the original level of routing.
Discuss
RPL uses two separate modes (storing and non-storing nodes) to allow for the possibility of very resource constrained sensing devices in the network. What is the difference between the storing and non-storing modes in terms of the routing choices available?
Non-storing nodes have minimal routing table & only pass messages forward towards the DODAG root. Storing nodes have full routing table available and may route independently of DODAG root, this results in higher power consumption and more resources needed on node.