Push vs. Pull: Implications of Protocol Design on Controlling Unwanted Traffic
by Zhenhai Duan, Kartik Gopalan, Yingfei Dong
Usenix,
2005-07-07
Language:
English
Note: SRUTI'05 (Steps to Reducing Unwanted Traffic on the Internet Workshop), July 7 2005, Cambridge, MA.
Abstract
In this paper we argue that the difficulties in controlling unwanted Internet traffic, such as email SPAM, stem from the fact that many Internet applications are fundamentally sender-driven and distinctly lack receiver control over traffic delivery. However, since only receivers know what they want to receive, receiver-driven approaches may often have clear advantages in restraining unwanted traffic. In this paper, we re-examine the implications of the two common traffic delivery models: sender-push and receiver-pull. In the sender-push model, a sender can deliver traffic at will to a receiver, who can only passively accept the traffic, such as in the SMTP-based email delivery system. In contrast, in the receiver-pull model, receivers can regulate if and when they wish to retrieve data, such as the HTTP-based web access system. We argue that the problem of unwanted Internet traffic can be mitigated to a great extent if the receiver-pull model is employed by Internet applications, whenever appropriate. Using three popular applications - email, mobile text messages, and asynchronous voice messages - as examples, we demonstrate that asynchronous communication protocols can be easily designed using the receiver-pull communication model to suppress unwanted Internet traffic.
