Method for improving TCP performance over wireless links

A TCP-based congestion management protocol for a wireless data connection, comprising:

• monitoring changes in the length of a transmission queue in a data connection;
• designating packet loss as being due to congestion if said packet loss is preceded by at least two consecutive intervals of increasing queue length;
• and
• designating packet loss as random loss if said packet loss is not preceded by at least two consecutive intervals of increasing queue length.

A method for differentiating congestion-related packet loss versus random packet loss in a wireless data connection, comprising:

• monitoring changes in the length of a transmission queue in a wireless data connection over an interval substantially equal to the amount of time it takes to transmit a window of data packets and receive acknowledgements corresponding to all data packets transmitted in the window;
• designating packet loss as being due to congestion if said packet loss is preceded by an increase in the queue length;
• and
• designating packet loss as random loss if said packet loss is not preceded by an increase in the queue length.

A TCP-based congestion management protocol for a wireless data connection, comprising:

• monitoring changes in the length of a transmission queue in a data connection over an interval substantially equal to the amount of time it takes to transmit a window of data packets and receive acknowledgements corresponding to all data packets transmitted in the window;
• designating packet loss as being due to congestion if said packet loss is preceded by at least two consecutive intervals of increasing queue length;
• and
• designating packet loss as random loss if said packet loss is not preceded by at least two consecutive intervals of increasing queue length.

A method for improving TCP performance over a wireless connection, comprising:

• detecting the initial stages of congestion in the connection, and identifying the direction of the congestion;
• determining whether congestion is developing in the forward or reverse path of the connection;
• isolating the forward throughput from events such as congestion that may occur on the reverse path;
• determining congestion by calculating the relative delay that one packet experiences with respect to another as it traverses the network;
• using said relative delay to estimate the number of packets residing in a bottleneck queue;
• keeping the number of packets in the bottleneck queue at a minimum level by adjusting the TCP source's congestion window;
• reducing the congestion window if the bottleneck queue length increases;
• and
• increasing the congestion window when the source detects additional bandwidth availability in the connection.