2. Literature Review

Many network applications are built on top of TCP. And many are continuing to be built on TCP in the future. Up to date studies show that several alternatives have been proposed to lessen the poor end-to-end performance of the unmodified TCP in the wireless medium. The protocols in this section will point out the advantages and disadvantages of each approach.
• Congestion Avoidance: It is assumed that all losses are due to congestion avoidance, which becomes quite problematic over wireless links. Congestion avoidance is a way to deal with lost packets. In packet-switched networks, packets move in and out of the buffers and queues of switching devices as they pass through the network. A packet-switched network is often referred to as a “network of queues”[W. Stevens 2]. A characteristic of packet-switched networks is that packets may arrive in bursts from one or more sources. Buffers help routers absorb bursts until they can catch up. If traffic is excessive, buffers fill up and new incoming packets are dropped. Increasing the size of the buffers is not a solution, because excessive buffer size can lead to excessive delay. Congestion can occur when data arrives on a fast LAN and gets sent out a slower WAN [Allen 2]. Congestion can also occur when numerous input streams arrive at a router whose output capacity is less than the sum of the inputs. Now, any host injects packets at anytime and those packets are variable in size, which makes predicting traffic patterns, and providing guaranteed service impossible.
• Slow-Start mechanism: A similar congestion detection mechanism is applied during slow-start to decide when to change to the congestion avoidance phase. A slow-start mechanism works by limiting the rate at which data is transmitted into the network. To have valid comparisons of the expected and the actual throughput, the congestion window is allowed to grow only on every other RTT (real time technology). When a router begins to transmit data into a Frame Relay network the swift increase in the amount of traffic may cause congestion in the network. Slow start minimizes the impact of increased network traffic and reduce the severity of any congestion. Older TCP’s would start a connection with the sender injection numerous segments into the networks, up to the window size advertised by the receiver ( Teltrend 4). While this is fine when the two hosts are on the same LAN, if there are routers and slower links between the sender and the receiver, problems can occur [Allen]. Some intermediate router must queue the packets, and it’s possible for that router to run out of space. Congestion avoidance and slow start are independent algorithms with different objectives. But when congestion occurs TCP must slow down its transmission rate of packets into the network, and then call upon slow start to get things going again. In practice they are implemented together. Congestion avoidance and slow start require that two variables be maintained for each connection: a congestion window (cwnd), and a slow start threshold size (ssthresh). • The Fast-Retransmit And Recovery Approach: This approach addresses the issue of TCP performance when communication resumes after a handoff. Unmodified TCP at he sender interprets the delay caused by a handoff process to be due to congestion (since TCP assumes that all delays are caused by congestion) and when a timeout occurs, reduces its window size and retransmits unacknowledged packets. Often, handoffs complete relatively quickly (between a few tens to a couple of hundred milliseconds), and the mobile host requires long waits before timeouts occur at the sender and packets start getting retransmitted. This is because of coarse retransmit timeout granularities (on the order of 500 ms) in. The fast retransmit and recovery approach lessens this problem by having the mobile host send a certain threshold number of duplicate acknowledgments to the sender. This causes TCP at the sender to immediately reduce its window size and retransmit packets starting from the first missing one (for which the duplicate acknowledgment was sent). The main drawback of this approach is that it only addresses handoffs and not the error characteristics of the wireless link [R.Caceres and L. Iftode 2].