Strategies

We propose two strategies to determine the activity period of each network.

• Fixed Priority: Each network is assigned a fixed pre-allocated Activity Period. This time is prioritized, with some networks getting more time than the others. These priorities are specified by the user based on his or her requirements.
• Adaptive Schemes: The Activity Periods of the networks are decided adaptively based on the amount of traffic seen by the node on the network. A more active network gets more time as compared to a less active one. So, if MultiNet has to switch across $N$ networks, and network $i$ has seen $P_i$ packets in its last Activity Period $ATP_i$ , then the node stays in network $j$ for an Activity Period given by $(P_j/ATP_j)*(1/(\sum_{\forall k} P_k/ATP_k))*(\sum_{\forall k} ATP_k)$ . The first term gives the network utilization of network $j$ , the second gives the utilization across all networks, and the final term is the total amount of time the node is active across all networks. The adaptive strategies give a default time to a network that has not transmitted any traffic. This is necessary to handle the case where a node just wants to receive packets on a network.

  We study two different adaptive strategies for MultiNet. Adaptive Buffer adapts the Activity Period to the number of packets buffered on each network, while Adaptive Traffic adapts to the number of packets sent and received on each network over a predefined time window.