uml diagrams for packet loss control using tokens at network edge

Packet loss is taboo; To an Internet architect, it immediately means an inefficient design that is likely to present instability and poor performance. In this article, we argue that such involvement is not fundamental. In particular, design points exist that provide many desirable properties including near optimum performance while suffering high loss rates. We focused specifically on congestion control, where researchers have clung to the belief that loss avoidance is critical to high performance. We present a protocol that supports the exchange of resources that exist in different networks of packet switching. The protocol provides for variations in individual network packet sizes, transmission failures, sequencing, flow control, end-to-end error checking, and the creation and destruction of process-to-process logical connections. Some implementation issues are considered and issues such as internetwork routing, accounting, and waiting times are exposed. Initial TCP congestion control algorithm, the entire tradition of end-to-end congestion control has attempted to optimize network performance by tuning transmission rates in response to loss. We argue that by eliminating the unnecessary yoke of avoiding losses from congestion control protocols, by using random early detection (RED) detect incipient congestion. They may become less complex but at the same time more efficient, stable and robust.