TCP Congestion Control with a Misbehaving Receiver
Reviewer: Jie Zhou
TCP was originally designed for a cooperative environment, and its evolution through the years has built on this base. However, with the growth of Internet,
the situation is changed. Now, Internet community includes a huge population of users with different interests. It is not surprising that some users may
try to utilize the vulnerabilities within TCP to get benefit. What should we do to prevent such misbehavior and protect the common interest?
This paper studies the vulnerabilities that misbehaving receiver can exploit, and presents solutions to prevent these misbehaviors.
TCP contains several vulnerabilities that an unscrupulous receiver can exploit to obtain improved service at the expense of other network clients or
to implement a denial-of-service attack.
The design of TCP can be modified, without changing the nature of the congestion control function, to eliminate these vulnerabilities.
I give the paper a rate 4 (significant contribution), because it suggests a simple but effective machanism between untrusted parties to prevent receiver
misbehavior in Internet. The author gives clear analysis of TCP vulnerabilities and makes a modified TCP implementation which can get extremely high performance
at the expence of its competiters. This convincingly shows the problems inside TCP. Moreover, the solution that the paper presents is a general solution which
works not only for TCP but also for other sender-based congestion control scheme.
The author does not talk about how to deploy their solutions on Internet. Because the proposed Cumulative Nonce requires the particapation of all clients,
it should be deployed all at one time. This is expected to cause difficulty.
Techniques are usually developed on some assumptions. Those assumptions may become invalid as time passing by. In such case, we should modify the techniques
according to the new situation.