In this project, we intend to build and deliver a scientific and technical collaborative observatory to measure and improve the Internet service access as perceived by users. We will propose new original user-centered measurement methods, which will associate social sciences to better understand Internet usage and the quality of services and networks.
Our observatory can be defined as a vantage point, where :
- tools, models and algorithms/heuristics will be provided to collect data,
- acquired data will be analyzed, and shared appropriately with scientists, stakeholders and civil society, and
- new value-added services will be proposed to end-users.
Through this proposal, we want to create a real synergy among Inria teams (Spirals, Diana, Muse, Dionysos and Madynes, Inria Chile) within the Internet Measurement domain.
This project will also allow Inria to become a key reference in the digital field, not only for scientist researchers but also for policy makers, rulers, and, for citizens in general, by giving them a more accurate and reliable basis for decisions making at an individual scale (privacy strategies) or at a collective scale (legal norms). An originality of our proposal is to ensure that researchers from different disciplines (exact sciences, applied sciences and social sciences) will collaborate in the design of this observatory, in the dissemination of results for the research community, executives and public at large.
Over the last decade, the Internet has become a social infrastructure and thus a central component of communications, technology development and online business. Many essential activities have been deployed such as administration, financial system, voting mechanisms, education and medias. In particular, new usages — induced by mobile and Internet-of-Things (IoT) devices — have been recently emerging, thus enforcing societal dependency towards Internet. To cope with this diversity and recent exponential growth of Internet traffic, delivering Internet services has also increased in complexity. Intermediate players have been multiplying and each of them has been proposing solutions to enhance service access performance. Some operators have deployed content delivery systems to optimize delivery latency. Over-The-Top (OTT) content players, which provide multimedia services over the Internet, have proposed their own proxies to improve the quality of their services. Even web browsers such as Chrome have reduced web page retrieval by promoting progressively a new protocol called QUIC (Quick UDP Internet Connections).
The user local environment can also be fairly complex (e.g., NATs/firewalls, middleboxes, wireless LANs). In a simple web page visit for instance, a browser has to contact a DNS resolver and multiple other hosts, because different parts of a web page often come from servers in different locations and from HTTP proxies/caches. The download performance of a web page does not therefore strictly depend on the ISP, but also on several other factors and strategies implemented by the Content and Service Providers (CSP).
Given the best effort nature of the Internet and its unceasing growth both in size and heterogeneity, Internet stakeholders (researchers, service and content providers, civil society, regulators, etc.) need to better understand its characteristics, its limitations, and its dynamics, both globally and locally, to improve applications and protocols design, detect and correct anomalous behaviors. Therefore, a collaborative environment is required where observation tools and methods can be developed, and where data can be acquired, analyzed, and shared appropriately in order to get a better knowledge about the Internet as perceived by users. This knowledge of the Internet usage will first allow researchers in the field of networks, information technology, applied mathematics, social sciences to design theoretical models or new concepts and compare them with facts through the collected data. It will also provide public and private policy makers, and private citizens in general, with a more accurate and reliable basis for decisions making.
Measuring Internet usage will involve a new user-centered paradigm as users now are always connected and access different services through different networks and devices at different locations. In addition to measuring end-to-end Quality of user Experience (QoE), we have to diagnose the sources of degraded experience and understand how users deal with this degradation.