FairNet: A Measurement Framework for Traffic Discrimination Detection on the Internet

TL;DR

FairNet is a measurement framework designed to detect traffic discrimination on the Internet by comparing similar HTTPS streaming services and accounting for confounding factors, aiding enforcement of net neutrality laws.

Contribution

It introduces a novel end-to-end measurement framework with algorithms for detecting traffic discrimination, addressing challenges of confounding factors and legitimate traffic management.

Findings

FairNet accurately detects traffic discrimination in real network conditions.

The framework effectively distinguishes between natural network effects and deliberate discrimination.

Validation shows reliable detection of traffic discrimination using live Internet data.

Abstract

Network neutrality is related to the non-discriminatory treatment of packets on the Internet. Any deliberate discrimination of traffic of one application while favoring others violates the principle of neutrality. Many countries have enforced laws against such discrimination. To enforce such laws, one requires tools to detect any net neutrality violations. However, detecting such violations is challenging as it is hard to separate any degradation in quality due to natural network effects and selective degradation. Also, legitimate traffic management and deliberate discrimination methods can be technically the same, making it further challenging to distinguish them. We developed an end-to-end measurement framework named FairNet to detect discrimination of traffic. It compares the performance of similar services. Our focus is on HTTPS streaming services which constitute a predominant portion of the Internet traffic. The effect of confounding factors (congestion, traffic management policy, dynamic rate adaptation) is made `similar' on the test services to ensure a fair comparison. FairNet framework uses a ``replay server'' and user-client that exchanges correctly identifiable traffic streams over the Internet. The Server Name Indication (SNI) field in the TLS handshake, which goes in plaintext, ensures that the traffic from the replay server appears to network middle-boxes as that coming from its actual server. We validated that appropriate SNIs results in the correct classification of services using a commercial traffic shaper. FairNet uses two novel algorithms based on application-level throughput and connection status to detect traffic discrimination. We also validated the methodology's effectiveness by collecting network logs through mobile apps over the live Internet and analyzing them.