Proactive or Reactive Routing

Proactive and Reactive are ways to classify a routing mechanism in terms of when an algorithm announces the routing route selection information. The routing mechanism mayanticipate having this information or select a route on-demand due to the high dynamicity and unpredictability of a network environment.

The proactive routing announces periodically network information generating a predictability amount of messages. In a worst case, the time to select the correct path t_s = t_period(period used to update paths). For example, formula (8) presents a possible generic function that may be adopted to calculate the number of messages generated by each node k when using proactive algorithms, where T is a set of announcement periods t independently of node or network faults. Hence, if there is large interval between path announcements, the number of messages will be small, nonetheless leading to a routing scenario practically lacking reactivity to possible network failures. Examples of pro-active routing include: DSDV (Destination-Sequenced Distance Vector) and WRP (Wireless Routing Protocol).

On the other hand, the algorithms classified as reactive do not issue their route announcement messages following any periodicity. They rely on observed failures or user demand to update routes. Consequently the number of messages announced could be much lower than that of proactive algorithms. Given that they do not need to wait a whole period to update their routing information, they may have offer better performance in unpredicted environment. They may also however suffer from low select the correct path may be given as the sum of t_s = t_fault (time needed to identify a failure)+t_request(to search for a correct path) + t_response (someone response the node with the correct path). For example, formula (9) presents a possible generic function to calculate the number of messages generated by each node k using a reactive algorithm, where F is a set of independent failures. So, one cannot know the number of message without knowing the number of failures. When some fault occurs, node k is notified, next it announces a request to find the route and others nodes announce or send back their responses. Examples that fall into this routing category are AODV (Ad-Hoc On-Demand Distance Vector), DSR (Dynamic Source Routing) and LMR (Lightweight Mobile Routing).