Since the early days of the Internet, TCP has been the premier transport protocol for most if not all applications. TCP predates IP and was designed to run over any packet-switching wireline network environment where errors due to the underlining medium are an exception rather then a rule. This complex protocol, in addition to many services such as flow control, congestion management and reliability, attempts to provide both an efficient utilization and a fair share of network resources. Wireless and mobile networks generate an environment in which TCP may suffer severe performance degradation due to possible misinterpretation of random losses that come from high BER, low and variable bandwidth, and frequent disconnections. The reason is the triggering of the congestion management mechanisms in response to the losses since TCP assumes by default that a lost packet is an indication for network congestion. The results are a decrease in throughput, poor resource utilization and eventually an unacceptable performance of the applications. There is no alternative in making TCP understand heterogeneity through modification and enhancement of the algorithms that the protocol consists of. Naturally, some of the options are enumerated in the article, but in the model we propose the preservation of the E2E semantics is an imperative. In addition to an overview of the �mainstream� research area, this article present the deployment of the loss pair method as a possible solution to some congestion management issues and thus making heterogeneity transparent to TCP.