Improving the Measurement and Analysis of Gigabit Broadband Networks
Accurate measurements of broadband performance are important to help consumers, ISPs, edge providers, and regulators make informed decisions regarding the choice, design, and regulation of broadband services. Broadband has become an essential element of basic infrastructure and as offerings capable of gigabit data rates become more widely available, performance measurement and interpretation methodologies will pose challenges in managing end-user performance expectations and developing regulatory policies effectively.
A number of broadband measurement platforms are currently being used by regulators, special interest groups, academic researchers, and the general public to test the performance of gigabit broadband access services. Based on various assumptions about what should be tested, these tools employ different methodologies and test configurations. Understanding those differences is necessary in order to interpret test results appropriately.
As a backdrop, it is important to note that after the adoption of gigabit broadband service, the wired access link from a user’s home to the access provider is less likely to be the performance bottleneck. Recent test results have suggested that as speeds increase beyond 20 Mbps, in-home wireless bottlenecks dominate. Outside of the home, the performance bottlenecks for very high-speed links tend to arise from sources other than congestion on the access link itself.
Metrics such as latency, accuracy, and reliability are appropriate in some contexts, but speed test methodology becomes highly relevant when access link speeds increase to the point where performance bottlenecks are arising beyond that point in the network.
When applied to very high-speed broadband services, the current approaches for comparing and evaluating different service offerings, assessing the overall performance of broadband providers across multiple service tiers, and making inferences about end-to-end quality of service may result in misleading or incorrect conclusions. This can distort the overall picture of gigabit service quality, resulting in disincentives for providers to deploy higher speed offerings and prompting the misdirection of investments and policies targeted at addressing perceived performance problems.
The issue is not simply that diverse methodologies will produce varying results, but rather that they embody different sets of performance expectations. Understanding the methodologies and the associated performance expectations is a critical step for advancing performance testing and monitoring in the era of gigabit broadband.
The paper examines current approaches for measuring very high-speed broadband services, offers comments on the technical and policy challenges that arise in such an environment, and discusses potential strategies for addressing these challenges.
For more information see “Improving the Measurement and Analysis of Gigabit Broadband Networks”.