Cooperation Working Group Minutes RIPE 75

Wednesday 25 October 2017 9:00-10:30
WG co-Chairs: Johan (Julf) Helsingius and Achilleas Kemos (not present)
Scribe: Gergana Petrova

Julf welcomed attendees to the session. He reminded the room of the purpose of the Cooperation Working Group: to liaise with external entities such as the European Commission, governments, IEEE, ITU, IETF and work with and help organisations with which we share goals.

What's Happening at the IETF: Internet Standards and how to get involved - Salam Yamout, RIPE NCC Executive Board

The presentation is available at:
https://ripe75.ripe.net/archives/video/148/

Lee Howard, Retevia, WG co-chair at IRTF and member of IAB, added that if people are unable to attend IRTF meetings, they can also join the open mailing lists of the various Working Groups. Lee urged more operators to join the IETF to say what operators need. He added that vendors don’t always know the needs of operators, beyond their need to buy equipment.

Salam responded that she is disappointed not to see many people from the Middle East in the room today and also at IRTF meetings. The IETF policy programme and the IETF fellowship programme offered by Internet Society are helping bring more people, but Salam felt that the IETF is still missing points of view and allies from the Middle East.

Lee added that the Internet Society has a program for academics to get a free day pass or a discounted pass.

Salam replied that this might not be enough and that we need to think harder how to bring more people from the Middle East.

Lee agreed.

IEEE's Work on Ethics re the Internet and AI Karen McCabe, IEEE

The presentation is available at:
https://ripe75.ripe.net/archives/video/151

Jim Reed commented that since IEEE is membership-based organisation a lot of stuff is done behind closed doors. He asked how this is impacting the IEEE’s ability to engage with stakeholders from many different groups.

Karen responded that the IEEE does not have a member state membership and that anyone can join and participate.

Jim responded that, in his understanding, he can only access the IEEE documents if he has paid to back a member.

Karen responded that they are working to change this business rule. She added that this rule is a legacy from the time when the IEEE needed funding from the sale of standards in order to continue their work on creating standards. The midway currently employed is that people involved in the Working Groups have access to all documents the WG is working on. Karen summarised that IEEE’s end goal would be to make all standards available freely to everybody, but that they are still working to figure out other funding mechanisms.

Jim commented that there appeared to be people doing forum shopping and added he hopes to have a chat with Karen about this during the coffee break.

Chris Buckridge, RIPE NCC, mentioned that the RIPE NCC has been working more closely with the IEEE in recent years, for example by being part of the technical advisory group at the OECD. He mentioned it is good to have such presentations at RIPE meetings. Chris commented that the issue of trust has come up several times during the RIPE Meeting – during this presentation, but also during discussion on interconnection and IoT. He asked how the IEEE deals with the issue of trust, trusted devices, ability to trust, trusts in protocols and security.

Karen answered that trust is a driving force for the work in the technology policy space. If the standards the IEEE produces and the organisations that use those standards cannot be trusted, then the benefit of all this work will be lost. The issue of ethical considerations of technology is really important for the IEEE and they look at it from different levels: markets, industry and enterprise.

Naser Salam, RIPE NCC commented that the IEEE is working hard with institutions in the Middle East and have study groups for students. He asked whether the IEEE is dealing with the technical community and the telecom operators.

Karen answered that one of the foundations of the organisation is its distributed community, with hundreds of sections and chapters around the world. Every three years the IEEE hosts a sections congress, the last one being just a few months ago. This section leadership is paying attention to the technology, business and industry aspects. The IEEE has offices in India, China, Singapore and just opened an office in Europe, but the Middle East and South and Central America are still underrepresented. The IEEE’s board of directors is working on new regional strategies and Karen will bring this comment to them.

ICANN's Identifier Technology Health Indicators (ITHI) Alain Durand, ICANN

The presentation is available at:
https://ripe75.ripe.net/archives/video/152

Marco Hogewoning, RIPE NCC, asked if Alain has historic data.

Alain answered that the project’s data collection starts from 2017 and will collect data month by month or week by week, depending on the data and access to it.

Alex Semenyaka, RIPE NCC, asked if they’ve considered the opportunity to cooperate with passive DNS database operators concerning protocol parameters.

Alain replied that this is a great idea and that he would appreciate if Alex connects him with relevant people.

Peter Koch, DENIC, commented that ICANN has no role in deployment or promoting of TLSA. He asked Alain why he left out the EPP suit of protocols and their interesting diversity in special features.

Alain explained that they are going for the low hanging fruits. DNS operation is more related to ICANN’s work than parameters of telnet or FTP. The project is focusing on five problem areas, but Alain would be happy to create another one, focusing on EPP registration for example, and work with Peter to develop it further.

Peter commented that he is concerned that DNS abuse is mentioned under technology health indicators.

Alain admitted that the word “health” is a bit problematic. At the same time there seems to be a high concentration of abuse in some areas,  in some TLDs or some registrars, which is why tracking is important.

Peter agreed. In his view, however, car manufacturers are not responsible for criminals preferring to use a type of car with certain characteristics. Scholars in the governance field differentiate between macro and micro level. The names and numbers communities are concerned about the identifier itself, rather than actions happening several layers away from them. Mixing or not appreciating the boundaries between these two systems is dangerous.

Alain said that the identifier is the one that triggers a Botnet, so there is a link, a direct link, between the identifies and the type of abuse.

Peter said that they need to agree to disagree if the link is direct or not.

Patrick Falstrom, Netnod, asked how much GDPR is affecting ICANN and if they are working on a data processing agreement.

Alain answered that the idea for the project is to remove all personal identification date from the raw sources that they have.

Alain and Patrick agreed that they understand a data processing agreement to be an organisation’s way of handling and storing data.

Alain gave a specific example of a script working with DNS recursive operators, which makes sure that there is no private data exported to ICANN. The script will run in the service provider network and will only send ICANN the abstracted data, so they wouldn’t have any personal information to begin with.

Patrick asked if that means that even though ICANN provides the script and has an agreement with whoever is going to collect the data on ICANN’s behalf, ICANN is not a data processor.

Alain confirmed that this is his interpretation at the moment.

Patrick thanked Alain and added that this discussion helps those who are also dealing with GDPR.

Jim asked Alain to confirm if he means stripping off DNS query names, IP source addresses and so on. He added that what is considered personal information in one EU jurisdiction, might not be considered in another and vice versa. A long term solution could be for ICANN to enter into safe harbour arrangement to process and store personal identifying information under terms and conditions and added that the solution is not as simple as Alain suggests.

Alain agreed and said that it would depend on the source of data. In the example he gave, which looks at protocol parameters and DNS collected from recursive resolvers, only frequency of a protocol parameters will be extracted - no DNS names, no IP addresses. He agreed that in other cases where personal information might be collected they need to think about the GDPR.

Internet Services and Energy Demand - Mike Hazas, Lancaster University

The presentation is available at:
https://ripe75.ripe.net/archives/video/154

Alexander Isavnin, Russian Internet Protection Society, asked Mike if he knows how much energy was spent during the presentation.

Mike answered that he didn’t calculate the energy for this talk, but a similar talk in the past took 21 hours just to stream.

Jim Reed mentioned that in the UK the national grid monitors when some popular TV programs finish, in expectation that most people will rush to the kettle to make tea, so the grid prepares for the surge in demand. He asked if anybody has given any thought on monitoring patterns of Internet traffic usage and informing grids of potential peaks.

Mike answered that on the Internet things are less synchronised but warnings from YouTube, Netflix and similar services might be useful.

Brian Nisbet, HEAnet, said he doubts about urging individuals to change their behaviour: firstly, because he doubts a large amount of people would do it and secondly, because the majority of power usage is done by large corporations and content providers rather than by individuals. Brian also wondered if energy savings from networks make as much sense as energy savings from production of servers for example.

Mike agreed that a broader change happens by changing infrastructures. At the same time, he underlined the importance of raising the consciousness of people, which might in turn affect the big infrastructure providers. There is also a role for policy to play, governments can negotiate with large service providers.

Marco Hogewoning, RIPE NCC, pointed out that devices nowadays are much more energy efficient than before.

Mike replied that with the rise of capacity and efficiency came a rise in demand. Instead of hundreds of devices, we have billions; and instead of low quality, we have high quality streams.

Chris Buckridge, RIPE NCC, saw it problematic to put stickers on videos because the energy it takes to view it would largely depend on the device used, or the location of the user. He asked Mike if large operators have shared with him how their energy usage changes with different network deployments.

Mike responded that gathering evidence to make reasonable analysis is difficult, since data is hidden in black boxes and data centers and is protected by corporations. In the past, he has tried to find a sympathetic media or other organisations until there is some broader pressure for companies to do their own analysis. The quick rule of thumb is that in the UK the energy it takes to watch a 2-hour film on LED TV is equivalent to the energy it takes to stream it. It is double for a large TV and times ten for a small device.

Jim Reed agreed about the need for a multidisciplinary approach. He disagreed with Brian that a large amount of people won’t change their behaviour. Small economic incentives can make people change their behaviour, such as the charge for plastic bags in shops, which resulted in a 99% drop in their usage in Ireland.

Mike agreed and added that environmental ratings on services or videos might incentivise people to change their behaviour and even if it results in just a few percent of savings, that would already be significant. Qualitative data tells us what services mean in people’s lives - messaging friends in a group for example.

Dmitry Burkov, RIPE NCC Executive Board, questioned Mikes approach because it mixed energy and electroenergy. In the future, when cars become electric, Internet’s share of electroenergy usage would automatically drop. Dmitry proposed it would be more relevant and correct to use any energy, regardless of the source (electricity, gas etc).

Mike agreed.

Malcolm Hutty, LINX proposed instead of trying to persuade people to change their behaviour that it would make more sense to produce more energy efficient devices and waste less energy.

Mike responded that the market is good at achieving efficiency over time, but that efficiency does not keep up with demand.

Malcom added that lack of energy is not a problem in the universe. We shouldn’t aim to reduce energy over the benefits that people get from the consumption of that energy and turning it into useful work.

Alex Semenyaka, RIPE NCC, added that he would be curious to see the energy consumption of new technology, such as cryptocurrency.

Mike agreed that this would be very interesting.