CTCP: Reliable Transport Control Protocol for sensor networks

This work presents collaborative transport control protocol (CTCP), a new transport protocol for sensor networks. It aims at providing end-to-end reliability and adapts itself to different applications through a two level mechanism of reliability variation. CTCP achieves these properties using hop-by-hop acknowledgments and a storage control algorithm that operates at each node along a flow. It was observed that distributed fault recovery increases 98% the average delivery rate and that duplication of storage responsibility minimizes the message loss. Its congestion control differentiates communication losses from buffer overflow. CTCP is called collaborative because all nodes detect and act on congestion control and also because it includes distributed storage responsibility. It is scalable and independent of the underlying network layer. The protocol energy consumption overhead was calculated and discussed for two reliability levels.