Transmission Control Protocol (TCP) is central to the workings of the Internet. TCP provides for the reliable, ordered, and verified delivery of data packets sent across Internet Protocol (IP) connections.
Network performance can be impacted due to network congestion, traffic load balancing, or other unpredictable network behavior. Amidst all this turmoil, IP packets can get lost, unnecessarily replicated, or arrive out of order. TCP detects these problems and takes the appropriate action. It rearranges out of sequence packets. It requests that lost data be retransmitted. It also regulates network traffic to reduce the occurrence of the other problems. But all this traffic-cop action can lead to problems.
TCP was developed long before anyone had dreamed of the World Wide Web or the modern Internet. TCP was devised at a time when networks consisted of a few computers in close proximity to each other. Because TCP assumed that the computers were all nearby, it used the time to acknowledge receipt of packets as a reliable indicator of network congestion. However, as networks grew larger and more complex, and communications move across the Internet from continent to continent, and as media changed from copper to wireless to satellite, the travel times increased. Thus, acknowledgements on a totally non-congested network can arrive much more slowly than TCP was designed to manage. As a result, performance suffers.
To improve TCP performance, Acceleration Systems minimizes the number of acknowledgements required to transfer data from our cloud servers to the destination by running proxy engines for many of the most “chatty” applications such as Exchange Server and CIFS file storage. In addition, our advanced compression and inline deduplication technologies reduce the total number of packets required to deliver information. These TCP optimizations combine to significantly improve performance.