Asset Publisher
NETS: Large: Collaborative Research: User-Centric Network Measurement

The proposed project centers on measuring the Internet and concomitant errors and failures. Our central thrust is to push the measurement infrastructure as Jar to the edge as possible - via portals, browsers, and mobile platforms. Therefore the main computing facilities required in support of the project are back-end infrastructure for collecting and archiving large quantities of data, supporting tests and measurements, and analyzing said data.

Our collaborators at ICSI manage significant production infrastructure to support the existing Netalyzer effort and we expect to largely utilize and enhance this machinery. Dr. Beverly and his colleagues at NPS have non-trivial experience in deploying and maintaining such public infrastructure over time, having previously supported large-scale, on-going measurements of IP source spoofing, port filtering, ECN behavior, and email reliability. NPS and ICSI are connected via Internet2, and NPS maintains several powerful machines with terabytes of storage to perform detailed analysis over large datasets. We therefore expect the NPS and ICSI facility sharing to be constructive.

As is often the case in Internet measurement, unexpected results often present in non-compliant equipment, configuration corner-cases, incompatibilities, etc. NPS is well-equipped for various laboratory testing. The networks and mobile systems group at NPS maintains a laboratory containing more than 50 commercial routers, 30 switches, 20 consumer NATs, 20 wireless base stations (including GSM), and approximately 100 personal computers allowing us to investigate various scenarios in a tightly controlled testbed environment that can emulate aspects of real-world deployment. In conjunction with equipment available to ICSI, we anticipate having sufficient resources to successfully deploy the array of new measurements envisioned in the project, as well as validate and analyze the resulting data.
Computer Science
National Science Foundation
NSF
2016