Improving performance of Ad Hoc TCP in MANETS Seminar Topic

Performance of TCP over MANETS:

The transmission protocol that is used extensively across the internet and file transfer data services is the Transmission Control Protocol TCP. Now-a-days the usage of TCP is also spread across the MANETs also. Generally across a MANET, there may be several TCP connections and so it is very difficult to estimate the performance of TCP in multiple connections which is also considered as a complex task. Generally, some of the attributes like mobility, traffic load and scalability are considered while estimating the performance of MANETs.

A wide research is made in the direction of MANET’s performance through the “Transmission Control Protocol” TCP, “User Datagram Protocol” UDP and “Constant bit rate” CBR along “File Transfer Protocol” FTP. As discussed earlier, across the MANETs scalability of the routing protocols is based virtuously on the attributes. In case of failures in the routes MANETs are not supported by the present form of TCP particularly in terms of the techniques of congestion control. The performance of TCP across the Mobile Adhoc Networks can be discussed by several techniques that already exist, but yet a lot of hard work is needed for this research to estimate the performance of TCP across the MANETs as it is difficult to do it with the currently available techniques.

TCP-F or TCP-Feedback is the feedback mechanism that can be used in route failures for improving the performance of TCP, in this case the route failures or link failures are identified by the intermediate nodes and then a notification message is passed to the original source of TCP and this notification message is called as the “Route Failure Notification” RFN. When this RFN is received by the source of TCP then it automatically ends the process of communication by stopping the sending of packets that are required. And then if in case a route to the destination node is identified by the intermediate node then it will send a “Route Re-establishment Notification” RRN for the source of TCP and once this RRN is received by the source then it will restart the process of data transmission by reactivating the state. Therefore, it makes clear that the original throughput of TCP is purely based on the average time taken to repair the route failure or link failure along the mobile Adhoc Networks.

Later another notification technique called as “Explicit Link Failure Notification (ELFN) is introduced for improving the TCP performance across the MANETs as indicated earlier. When the TCP source is intimated about the link failure or route failure by the intermediate node, then the TCP source will stop the timer and state across the window of transmission. The re-establishment of route is not done in this particular context, instead of re-establishing the route the TCP source will verify for the possible route in the MANETs and forwards packets continuously.

When compared to the RFN notification technique ELFN is considered as the best notification technique across both OPNET and NS2 simulations and this is also regarded as the best option for increasing the TCP performance across the Mobile Adhoc Networks. Generally TCP is regarded as an effective resource for the transmission of data and for the delivery of end-to-end information in case of the wired networks but in case of the wireless networks, TCP fails abruptly. Mostly in case of MANETs TCP fails and the reason for this is in certain conditions like delay and packet loss TCP cannot work where certain conditions such as packet losses and delay of packet delivery are generally encouraged by the MANETs. 

This paper is written and submitted by Sujana Priya V.

Leave a Reply

Your email address will not be published. Required fields are marked *