英语论文网

be done by the node “b” that is to be flooded.

OLSR routing protocol allows nodes to announce willingness to act as MPRs for neighbors. There are 8 levels of willingness the lowest one is WILL_NEVER (0) which means that this node will never be chosen as a MPR, and the highest one is WILL_ALWAYS (7), which means that this node will always be chosen as a MPR. Through Hello message the willingness is spread and when calculating the MPRs this information must be considered.

3.4.2 Forwarding OLSR traffic

Relaying of messages makes flooding in MANETS possible. OLSR specifies a default forwarding algorithm that uses the MPR information to flood packets. One is however free to make ones own rules for custom forwarding of custom messages. But all messages received that carries a type not known by the local node, must be forwarded according to the default forwarding algorithm. The algorithm can be outlined as:

1. If the link on which the message arrived is not considered symmetric, the message is silently discarded. To check the link status the link set is queried.

2. If the TTL carried in the message header is 0, the message is silently discarded.

3. If this message has already been forwarded the message is discarded. To check for already forwarded messages the duplicate set is queried.

4. If the last hop sender of the message, not necessarily the originator, has chosen this node as a MPR, then the message is forwarded. If not, the message is discarded. To check this, the MPR selector set is queried.

5. If the message is to be forwarded, the TTL of the message is reduced by one and the hop-count of the message is increased by one before broadcasting the message on all interfaces.

Figure: 3-5, Node A has selected the Black node as its MPR

The fact that all received unknown message types are forwarded using this approach makes flooding of special message-types possible even if these message-types are only known to a subset of the nodes.

Figures 3-3 and 3-4 shows the paths information is passed when being spread, first using regular flooding, then using MPR flooding. The number of retransmissions in a MPR scenario highly depends on the network topology and the MPR calculation algorithm. Using the same topology as in fig 3-2a, a possible MPR calculation could lead to the black nodes in fig 3-2b being chosen as MPRs by the center node. As one can see, if the center node is to flood a message throughout the network, 4 retransmissions are done using MPR as opposed to 24 using traditional flooding.

To be able to check if a message has already been retransmitted, a cache of recently processed and forwarded messages is maintained. The data stored is the minimum needed to identify the message. This means that the actual message content is not stored, but rather just originator address, message-type and sequence number. This data is cached for a constant time of DUP_HOLD_TIME suggested to be 30 seconds in the RFC. Every received message that is processed by the local node is registered in the duplicate set. If the message is forwarded, the duplicate-entry representing this message is updated accordingly; registering on what interfaces the message has been forwarded. Based on querying the duplicate set, a node can then keep track of already processed messages and already forwarded本论文由英语论文网提供整理，提供论文代写，英语论文代写，代写论文，代写英语论文，代写留学生论文，代写英文论文，留学生论文代写相关核心关键词搜索。