The Internet of Things (IoT) is an application-driven heterogeneous network, being rapidly deployed across the world. The proposed session layer for IoT is the upcoming standard for Constrained Application Protocol (CoAP). The transport layer of choice for IoT is emerging to be UDP. CoAP itself mandates usage of UDP only as the transport layer. It is well-known that practical usage of UDP requires extra employment of congestion control, generally in form of rate control. For bursty source data that is generated most of the while in IoT, usage of rate control is a must. Configuring the rate control parameters requires knowledge of the performance of the underlying network. In this paper, we first experimentally measure, analyze and model performance parameters, e.g. avg. bandwidth, round-trip delay and jitter, of a real heterogeneous wide area network. Based on the analysis, we then provide two novel, intuitive algorithms for source node rate control, and destination node timer management. These algorithms were found to improve the reception performance over that achievable using guidelines provided in RFC 5405. For delivery of urgent/critical data in IoT, we had earlier proposed usage of session-level forward error control (FEC) of the communicated data. Usage of our algorithms ensured that the deployment of the FEC scheme, which employed lightweight rate-10over11 Reed-Solomon codes, succeeded 98% of the while, while sending as many as 10 packets per burst. The algorithms can be re-used with similar high performance levels, in many other communication contexts also, that entail low-volume, bursty data traffic visualized in RFC 5405. Hence the algorithms may be considered for augmentation of CoAP standard.