The CITP Interconnection Measurement Project provides first-of-its-kind, direct measurement of interconnection capacity and utilization with a tool implemented by Internet service providers (ISPs). The visualizations linked below provide various views into the data.

Understanding the Data

Across all interconnects in the dataset, capacity is roughly half utilized during peak periods. During the eighteen-month time period we have studied to date, capacity has grown 74.3 percent cumulatively. Usage is also increasing, albeit at a more variable rate. A number of factors such as higher usage of Internet video over the winter holidays, likely influence the usage trends over time.

Another way to view interconnect utilization is as a proportion of the capacity that runs at higher utilization levels. In aggregate, a small proportion of total capacity runs at 95% utilization or higher during peak periods: in recent months, less than 5% of available capacity reaches this level of peak utilization (note data outage for December 2016 caused omission of certain measures, as discussed in more detail here). Examining the distribution of peak utilization reveals that, for each month, the proportion of capacity running at near 100 percent is often nearly equal to the proportion of capacity running at near-zero peak utilization. Note that high utilization levels do not necessarily correspond to negative quality of experience (QoE); the relationship between utilization and QoE is an exciting area for future work.

As shown in the above “Distribution of Aggregate Peak Interconnect Utilization by ISP,” the aggregate peak utilization for any individual participating ISP never exceeds 75 percent and the vast majority of participants have an aggregate peak utilization below 50 percent. To preserve confidentiality and protect proprietary information pertaining to participating ISPs and their interconnect partners, we do not provide a detailed breakdown of the data by ISP. See Views by Interconnect to see how utilization is distributed across available capacity.