This poster presents a probabilistic routing protocol for intermittently connected networks. In such networks, there is no guarantee of a fully connected path between source and destination, making traditional routing protocols ineffective. The proposed protocol uses a probabilistic metric called delivery predictability to determine the likelihood of a node delivering a message to its destination. This metric is updated based on past encounters between nodes and is also subject to aging, reducing its value over time if nodes do not meet frequently. The protocol also incorporates a transitive property, where the delivery predictability of a node can be inferred from the predictabilities of intermediate nodes. Decision making for routing is based on this metric, with the protocol allowing for flexibility in choosing which nodes to forward messages to, depending on the situation. In cases where no delivery predictability information is available, the protocol falls back to a variant of epidemic routing. The authors also mention future work, including the use of acknowledgments to remove messages from the network after delivery, and extensive simulations to evaluate the protocol's performance. The protocol is designed to be adaptable to different scenarios, allowing applications to tailor routing strategies based on their specific needs. The research is relevant to various fields, including Delay Tolerant Networking, sensor networking, and satellite communication.This poster presents a probabilistic routing protocol for intermittently connected networks. In such networks, there is no guarantee of a fully connected path between source and destination, making traditional routing protocols ineffective. The proposed protocol uses a probabilistic metric called delivery predictability to determine the likelihood of a node delivering a message to its destination. This metric is updated based on past encounters between nodes and is also subject to aging, reducing its value over time if nodes do not meet frequently. The protocol also incorporates a transitive property, where the delivery predictability of a node can be inferred from the predictabilities of intermediate nodes. Decision making for routing is based on this metric, with the protocol allowing for flexibility in choosing which nodes to forward messages to, depending on the situation. In cases where no delivery predictability information is available, the protocol falls back to a variant of epidemic routing. The authors also mention future work, including the use of acknowledgments to remove messages from the network after delivery, and extensive simulations to evaluate the protocol's performance. The protocol is designed to be adaptable to different scenarios, allowing applications to tailor routing strategies based on their specific needs. The research is relevant to various fields, including Delay Tolerant Networking, sensor networking, and satellite communication.