This paper group is introduces 3 approaches dedicated to solve the problems that current TCP implementation has in high-speed networks. Papers and their outlines are listed below.

[1] "FAST TCP: motivation, architecture, algorithms, performance", INFOCOM 2004 paper

by Cheng Jin, David X. Wei and Steven H. Low.

• Flow and packet level problems of TCP Reno
• Benefits of delay based approach over loss based one 
• Strategy of FAST TCP in solving problems of TCP Reno
• Architecture and underlying algorithms
  • Notes on estimation and window control parts
  • Notes on packet level implementation and states of the system
• Mathematical proof of equilibrium and window control stability algorithm
• Detailed testing of FAST TCP in comparison with HSTCP, STCP and TCP Reno

[2]  "Modifying TCP's Congestion Control for High Speeds" paper

by S. Floyd, S. Ratnasamy, and S. Shenker.

Summary:  This paper shows how TCP response function can be tweaked to provide more aggressive window increase and more gentle decrease. It briefly goes over:

• Problems of link utilization TCP has at high speeds
• Explanation of drop rate influence on TCP performance
• Goals of HSTCP
• Change of TCP response function and its impact on overall performance and fairness

[3]  "Scalable TCP: Improving Performance in HighSpeed Wide Area Networks", Feb 2003 paper

by Tom Kelly

Summary:  This paper introduces approach very similar to the paper #2. Same goals solved with different response function. Paper's outline:

• Link utilization problems of TCP when working in high speed networks
• New response function and its properties
  • Scalability
  • Fairness
  • Window size
• Discussion of TCP connection's rate variation, convergence and stability
• Experiments with STCP vs TCP Reno and TCP with gigabit modifications

Presentation: slides