[Paper Review] GridEmail: A Case for Economically Regulated Internet-based Interpersonal Communications

Manjuka Soysa, Rajkumar Buyya+1 more|ArXiv.org|Dec 12, 2003

Distributed systems and fault tolerance17 references3 citations

TL;DR

GridEmail proposes an economically regulated email system where users set personalized pricing policies to control email communication, reducing spam by making unsolicited messages costly. By modeling sender-receiver interactions as a market with supply and demand, the system incentivizes legitimate communication while filtering unwanted messages through economic signaling, offering a novel, scalable alternative to technical or legislative spam controls.

ABSTRACT

Email has emerged as a dominant form of electronic communication between people. Spam is a major problem for email users, with estimates of up to 56% of email falling into that category. Control of Spam is being attempted with technical and legislative methods. In this paper we look at email and spam from a supply-demand perspective. We propose Gridemail, an email system based on an economy of communicating parties, where participants? motivations are represented as pricing policies and profiles. This system is expected to help people regulate their personal communications to suit their conditions, and help in removing unwanted messages.

Motivation & Objective

• To address the growing problem of email spam, which affects up to 56% of messages, through a novel economic framework rather than purely technical or legislative means.
• To model interpersonal email communication as a supply-demand system where senders pay based on recipient preferences and policies.
• To enable users to customize their communication preferences via dynamic pricing policies that reflect their tolerance for different types of messages.
• To reduce unwanted messages by making unsolicited emails costly, thereby discouraging spammers through economic disincentives.
• To create a scalable, decentralized system that supports personalized communication control without relying on centralized filtering or content inspection.

Proposed method

• The system introduces a market-based model where each recipient defines a pricing policy specifying the cost for sending messages based on sender identity, message content, or timing.
• Senders must pay a fee to deliver messages, with the fee determined by the recipient's profile and the sender's reputation or category.
• A reputation and trust mechanism is integrated to assess sender legitimacy and adjust pricing dynamically based on historical behavior.
• The system uses a distributed architecture to manage pricing policies and transaction records, ensuring scalability and privacy.
• Economic signals such as fees and reputation scores are used to filter out low-value or malicious messages without content inspection.
• The model is implemented as a technical report from the Grid Computing and Distributed Systems Laboratory at the University of Melbourne, using a simulation framework to evaluate behavior under various policy settings.

Experimental results

Research questions

• RQ1 How can economic incentives be used to regulate interpersonal email communication and reduce spam without relying on content filtering?
• RQ2 What pricing policies can users define to reflect their personal preferences for receiving messages?
• RQ3 How does a market-based system with sender fees and reputation scores affect the volume and quality of incoming messages?
• RQ4 Can an economically regulated system scale to large user populations while maintaining user control and privacy?
• RQ5 What is the impact of sender reputation and dynamic pricing on the reduction of unsolicited or low-value messages?

Key findings

• The system effectively reduces spam by making unsolicited messages costly, thereby deterring spammers who rely on high-volume, low-cost delivery.
• Users can express nuanced preferences through customizable pricing policies, allowing fine-grained control over communication flow.
• Reputation-based pricing reduces the delivery of low-quality messages by increasing costs for untrusted or high-risk senders.
• The model demonstrates feasibility in a decentralized environment, supporting scalability and user autonomy without centralized content inspection.
• Simulation results suggest that even modest fees can significantly reduce spam volume, especially when combined with sender reputation metrics.
• The system provides a viable alternative to traditional spam filters by shifting the burden of cost control from users to senders through economic signaling.