I recently participated on a session on “Building research capacity in internet measurements, cyber policy and digital rights” at the #FIFAfrica 2017 conference on Internet Freedom. My role was to give a technical perspective on Internet measurements for cyber policy development from a Regional Internet Registry’s (RIR) point of view. The other panelists were Eric Frenkil from BudgIT Nigeria, Allison Gilwald, Research ICT Africa and James Marchant, Small Media. Internet measurements are often sought to be a strictly technical subject and there are very few references in literature where Internet measurements have actually helped to formulate internet policies per se. The aim of this panel was to bring together network specialists and Internet policy researchers to identify the gaps in research capacity in the area of digital rights and cyber policy.
“Internet Freedom in Africa: Where are We Now?” by @PollicyOrg https://t.co/2OMktSzi8R
— Francis Hook (@FrankHook) October 13, 2017
What does one mean when they say that the internet should be measured and what value does internet measurements bring? Internet measurements actually help to provide insight about different aspects of the internet, about its performance, its robustness and can help to understand how resilient networks are based on the data collected. For e.g. a country with only one incumbent ISP bears multiple risks, it may act as a single point of failure (cable breaks, network outage, etc) or having only one entry and one exit point in a country can make it easy for mass scale surveillance. So basically, Internet measurement research must not solely look at one aspect of the Internet (technical or otherwise) but use different techniques to have a 360 degree understanding of the “state of the Internet”. As such research in this field should be multi-faceted ranging from Internet policy (accessibility, affordability and usage) to quality of service (latency, throughput and performance) and security.
Internet measurement also encompases the study of censorship and surveillance systems. Censorship consists of mainly the blocking of “illegal” content, which is different from surveillance, whose role is to track users sending doubtful traffic. Surveillance systems are usually Deep-Packet-Inspection (DPI) based. An Internet measurement tool used to detect censorship is OONI (Open Observatory of Network Interference) and works by the placement of probes in different eyeball networks. However, as with any other censorship measurement mechanism, hosting an OONI probe and running measurements are not without risks. Surveillance, on the other hand, is actually more difficult to detect as the web content or service is not blocked, but rather traffic is eavesdropped, such that private user information can be revealed. However, there are some subtleties (for e.g. HTTP header manipulation) that can help detect whether a data packet has transited through a middlebox (for e.g. a DPI module or a transparent proxy).
As we are all rapidly moving towards a converged means of communication over IP (Internet Protocol), cyber policy and governance are (rapidly) breaking away from the silos we have inherited from the telecommunications and broadcasting regulations over the years. Africa is currently experiencing this drastic change as Internet penetration is globally improving - which is a good indication of growth - as an improved Internet access helps to bring new opportunities, break barriers and bridge digital divides. Yet, not everyone is playing on level fields, for e.g. the prohibitive prices of mobile data in South Africa is preventing equal access to the Internet (#DataMustFall campaign), unequal performance or government wiretapping or censoring (legally or otherwise) their citizens are preventing Internet users from exercising their full rights as described in the The African Declaration on Internet Rights and Freedoms. The question is therefore: “Internet development - yes - but at what costs”? “Cost” does not only refer to affordability of access, but also accessibility, performance, security and the protection of citizen’s digital rights such privacy, freedom of speech and access to information and even more recently the “right to be forgotten”. All those constitute important “health metrics” of the Internet and should constantly be measured. There is definitely a need for new expertise for research in digital rights and ICT access on the continent. Internet measurements can provide useful insights which is normally difficult to have without longitudinal measurement, how do you evaluate policy impact for e.g. national peering policy.
