ETAP 2015

Recently I was invited to take part in a conference put together by the Institute of Electrical and Electronics Engineers, the “IEEE Experts in Technology and Policy Forum on Internet Governance, Cybersecurity, and Privacy.” IEEE is the principal organization of computer and communications hardware and software engineers; it’s probably best known for its Ethernet and Wi-Fi standards, IEEE 802.3 and 802.11, but these are produced by the IEEE Standards Association, a related but independent body.

IEEE is a global organization that has realized it’s in a unique position to influence the international policy debates over spectrum, Internet governance, privacy, and network security. So ETAP 2015 brought together experts in the two fields to discuss the best ways inject a dose of technical knowledge in the current policy debates. The conference featured a panel discussion about priorities followed by a series of concurrent breakout sessions on particular issues.

Panelists included DSL inventor John Cioffi, Nancy Cam-Miller of Cisco, Greg Shannon of Carnegie Mellon, ethicist Jared Bielby, and Steve Diamond of EMC. This was a non-attribution conference, so I can’t tell you who said what, just what was said. One panelist observed that the Internet’s de facto governance model — the multi-stakeholder system that permits individuals, NGOs, and commercial firms seats at the table alongside government — has raised expectations it can’t deliver. The Internet faces regulatory threats because some users are criminal, and national laws differ. The Dubai WCIT conference failed to deliver a consensus, so the current state of Internet governance is murky.

Another panelist observed that the Internet of Things raises issues that are currently unresolved to a new level of friction; to achieve some of its benefits, it’s going to be necessary to collect and distribute Big Data, but we can’t violate basic privacy rights in the process. Improving threat detection will be vital when the most common Internet user is a thing rather than a person, and a real anonymization procedure will be very, very important.

One (and only one) of the panelists was pessimistic, forecasting a scenario where the costs of using information and communications technologies would exceed their benefits. It’s not too hard to work out how that might happen in the post-Snowden context and in a world where we depend on technology to drive our cars, give us directions, and manage our infrastructure. Engineers don’t like to think about these scenarios, but we must.

Finally, two speakers returned the panel to the rosy side by observing that the major nuisance for Internet users today isn’t surveillance or security failures; it’s the simple failure of the Internet to provide speedy and reliable services all the time. We’re orders of magnitude more likely to land on a dead web server than we are to get hacked, and we don’t even know who’s to blame when pages don’t load or videos don’t stream.

One sage commenter observed that it’s taken us 45 years to arrive at the fun and speedy insecure Internet we have today, so we shouldn’t be too surprised if it takes us another 45 to get to a system that’s genuinely safe and secure.

The moderator then asked the panelists whether technologies effectively have politics embedded in them; this sounds like a loony question, but there certainly are a number of people who’ve called the Internet a “technology of liberation” and similar things.

This is most common between the honeymoon phase of demonstrating on the national square for the end of the current regime and the head bashing that signals the regime’s response. It’s pertinent in the Internet’s case because we know now, if we didn’t always know, that the Internet is just as good for surveillance as it is for free expression.

The most interesting observations on this question were unexpected. One panelist remarked that net neutrality opens the door to security threats because open systems are fundamentally attack vectors; but this same quality makes a business opportunity of security. Another panelist observed that the Internet reflects the circumstances of its birth as a research network for a trusted community; it’s certainly not there any more.

The moderator then observed that building infrastructure is boring and asked if the IoT is simply another infrastructure. Panelists observed that “one man’s infrastructure is another man’s innovation” so it needn’t be boring. Others underscored that point by highlighting the engineering challenges that have to be overcome to make the IoT live up to its promise; there are many and it may take a catastrophe to make us wake up to them.

My breakout session dealt with the issue of Internet fragmentation. This is a pet issue of mine because I believe the Internet’s biggest threat is stagnation, and the only way you get that is to insist on uniformity. The Internet is too big to change top-down, so we can only improve it incrementally and around the edges.

When we reported that we didn’t see fragmentation as a problem as long as it doesn’t affect cross-network data flows, we got the only pushback from the floor among all the reports: The ICANN representative was very animated in reminding us that ICANN sees the “unicity” of the Internet as its most important feature. So we reminded him that the Internet is a “network of networks” perfectly capable of operating according to principles of federalism just as the US does (in theory at least.) He wasn’t satisfied, but if your only critic is ICANN you’re doing pretty well.

ETAP 2015 was a remarkable day and a half. It demonstrated that engineers can discuss policy at a meaningful level, and with considerably more originality than we often hear in government circles. It also demonstrated, as do the Silicon Flatirons conferences in Colorado, that you often find the most intriguing policy notions outside the Beltway. Exposing DC to the engineer’s thought process is winner; and that’s why there’s a High Tech Forum.

Are engineers better equipped to have meaningful tech policy discussions than Beltway wonks? Sound off in the forum!