Abstract

Wireless and satellite networks often have non-negligible packet corruption rates that can significantly degrade TCP performance. This is due to TCP’s assumption that every packet loss is an indication of network congestion (causing TCP to reduce the transmission rate). This problem has received much attention in the literature. In this paper, we take a broad look at the problem of enhancing TCP performance under corruption losses, and include a discussion of the key issues. The main contributions of this paper are: (i) a confirmation of previous studies that show the reduction of TCP performance in the face of corruption loss, and in addition a plausible upper bound achievable with perfect knowledge of the cause of loss, (ii) a classification of the potential mitigation space, and (iii) the introduction of a promising new mitigation that employs rich cumulative information from intermediate nodes in a path to form a better congestion response.

We first illustrate the performance implications of corruption-based loss for a variety of networks via simulation. In addition, we show a rough upper bound on the performance gains a TCP could get if it could perfectly determine the cause of each segment loss––independent of any specific mechanism for TCP to learn the root cause of packet loss. Next, we provide a taxonomy of potential practical classes of mitigations that TCP end-points and intermediate network elements can cooperatively use to decrease the performance impact of corruption-based loss. Finally, we briefly consider a potential mitigation, called cumulative explicit transport error notification (CETEN), which covers a portion of the solution space previously unexplored. CETEN is shown to be a promising mitigation strategy, but a strategy with numerous formidable practical hurdles still to overcome.

Keywords: Explicit transport error notification (ETEN); Explicit loss notification (ELN); Explicit congestion notification (ECN); Wireless and satellite networks; TCP performance; Congestion; Corruption; Bit errors; Channel fades

Article Outline

1. Introduction

2. Can ETEN help?

2.1. Oracle notifications

2.2. Single flow simulations

2.3. Competing traffic

2.4. Discussion

3. A taxonomy of corruption notification and response mechanisms

3.1. Loss discrimination

3.1.1. Implicit loss discrimination

3.1.2. Explicit loss discrimination

3.2. ETEN node behavior

3.2.1. Notifier behavior

3.2.2. Sender behavior

3.3. Control loop

3.3.1. Feedback

3.3.1.1. Deployment challenges for FEC schemes with TCP/IP

3.3.2. Locus of ETEN

3.3.3. Granularity

3.3.4. In-band vs. out-of-band signaling

3.3.5. Direction of notification

3.3.6. Determinism

4. Cumulative ETEN

4.1. Determining the packet corruption rate

4.2. Computing the total loss rate

4.3. Alternate congestion responses

4.3.1. Probabilistic: CETEN_P

4.3.2. Adaptive adjustment: CETEN_A

4.4. CETEN simulations

4.5. Discussion

5. Summary

Acknowledgements

References

Vitae