y
Abstract:
Connectionless traffic originating from LANs is carried as Available Bit Rate (ABR) traffic in the ATM network and is subjected to ATM level flow control, either rate-based or credit-based. In addition, transport layer flow control (e.g. TCP window control) is applied at the end points. In this paper, we study the interaction between TCP and ATM flow controls. We use simulation to compare the "goodput " of different ABR flow control schemes and to evaluate the sensitivity to different parameter selections (e.g., TCP Maximum Segment Size (MSS), round trip delay, etc). The study uncovers an unfairness situation caused by use of non uniform MSS values. 1
Citations
| 397 | TCP Vegas: New techniques for congestion detection and avoidance – Brakmo, O’Malley, et al. - 1994 |
| 362 | Congestion avoidance and control – Jacobson - 1988 |
| 130 | The Rate-Based Flow Control Framework for the Available Bit Rate – Bonomi, Fendick - 1995 |
| 80 | Evaluation of TCP Vegas: Emulation and experiment – Ahn, Danzig, et al. - 1995 |
| 42 | ATM rate based congestion control using a Smith predictor: An EPRCA implementation – Mascolo, Cavendish, et al. - 1996 |
| 21 | Receiver-Oriented Adaptive Buffer Allocation in Credit-Based Flow Control for ATM Networks – Kung, Chang - 1995 |
| 18 | A Language and Optimizing Environment for Distributed Simulation – BAGRODIA, MAISIE - 1990 |
| 4 | Internetworking with TCP/IP, volume vol. I – Comer - 1991 |
| 2 | Comparing ATM credit-based and ratebased controls for TCP source – Gerla, Pazos, et al. - 1995 |

