Files
Download Full Text (341 KB)
Abstract
Bandwidth across the internet is constricted by monetary factors and hardware development, so researchers are left to improve the efficiency of data transmitted in order to improve internet speeds. The transmission control protocol (TCP) is the primary transport protocol on the modern internet, ensuring reliable delivery of the majority of data transmitted. I evaluated the TCP header fields for efficient use of data overhead so as to determine current waste and to suggest possible areas for revision. I examined original specifications for TCP mechanisms, comparing them to modern implementations as determined by updated standards and modern practices in the networking community, and considered the data overhead that header fields related to these mechanisms entail. Consideration of implementations included evaluation of frequency of use as well as necessity of use. Current inefficiency in the data overhead of TCP should be addressed, because many of the header fields are either clearly wasteful or would be more efficient alternatively implemented. Certain core features of the TCP header cannot be revised without drastic alterations to the protocol, such as the sequencing and acknowledgement numbers. Other features, such as the header checksum, are inherent to the integrity of the header. However, most fields are arguably inefficient, as they either are not a continuing necessity for TCP’s function or they may be more efficiently implemented as TCP options. Further, several fields are no longer widely used, and are effectively totally wasteful. The proposed improvements to TCP overhead could result in a reduction of up to several bytes per segment transmitted. Admittedly, the savings of individual segments are only on the order of several bytes, which is a small percentage of most segments including payload. However, this small savings has the potential to result in savings orders of magnitude greater across the general internet. This potential suggests a need for further research into the viability of TCP header revision, followed by implementation of proposals. Results and methodologies used to reach these conclusions are additionally applicable to a variety of ongoing research (e.g., header compression, acknowledgement frequency reduction).
Publication Date
2015
Subject Major(s)
Computer Science
Keywords
Networking, TCP, Protocols, Transport Protocol, Internet, Header, Overhead, Efficiency
Disciplines
Computational Engineering | Computer Sciences | OS and Networks | Other Computer Sciences
Current Academic Year
Freshman
Faculty Advisor/Mentor
Faye Prichard
Rights
© The Author(s)
Included in
Computational Engineering Commons, OS and Networks Commons, Other Computer Sciences Commons