Automatic TCP

With the growth of high performance networking, a single host may have simultaneous connections that vary in band-width by as many as six orders of magnitude. We identify requirements for an automatically-tuning TCP to achieve maximum throughput across all connections simultaneously within the resource limits of the sender. Our auto-tuning TCP implementation makes use of several existing technologies and adds dynamically adjusting socket buuers to achieve maximum transfer rates on each connection without manual connguration. Our implementation involved slight modiications to a BSD-based socket interface and TCP stack. With these modiications, we achieved drastic improvements in performance over large bandwidth*delay paths compared to the default system connguration, and signiicant reductions in memory usage compared to hand-tuned connections, allowing servers to support at least twice as many simultaneous connections. 1 Introduction Paths in the Internet span more than 6 orders of magnitude in bandwidth. The congestion control algorithms RFC2001, Jac88] and large window extensions RFC1323] in TCP permit a host running a single TCP stack to support concurrent connections across the entire range of bandwidth. In principle, all application programs that use TCP should be able to enjoy the appropriate share of available band-width on any path without involving manual connguration by the application, user, or system administrator. While most would agree with such a simple statement, in many circumstances TCP connections require manual tuning to obtain respectable performance. For a given path, TCP requires at least one bandwidth-delay product of buuer space at each end of the connection. Because bandwidth-delay products in the Internet can span 4 orders of magnitude, it is impossible to conngure