Open Source Working Group Minutes RIPE 74

11 May 2017, 11:00-12:30 

WG Chairs: Martin Winter, Ondrej Filip 

Scribe: Fergal Cunningham 

Status: Draft

A. Administrative Matters


Working Group Chairs

Ondrej welcomed everyone and noted there were no agenda changes ahead of the meeting. The minutes from the last meeting were approved. Martin noted at the next RIPE Meeting there would be an election for a working group chair and asked people to start thinking about this.

B. Kea - a Modern DHCP Server

Tomek Mrugalski, ISC

The presentation is available at:
https://ripe74.ripe.net/presentations/140-kea-ripe74-final.pdf

João Damas, APNIC, preferred options 4 and 3 on how to provide funding for Kea because he thought Tomek would not get traction unless there was a big installed base. He talked about difficulties with getting funding. He asked about maybe integrating it into other things that already exist so it gets more visibility.

Christian Petrasch, DENIC, asked Tomek if he had a forecast for implementation in tools such as copper install server or being able to install virtual machines and it works with DHCP automatically.

Tomek said there was no support for this yet.

Pavel Foremski, Polish Academy of Sciences, thanked Tomek and asked if he considered a different markup language and Christian said everyone wanted to use JSON but people could probably use their own language if they wanted.

Randy Bush, IETF Meeting Network Operations Centre Servers Division, said they were looking at moving from IFC DHCP to Kea, and at their scale he wouldn't expect to make much money from conversion tools because it looks too simple.

Christian said the configuration is the important issue because if that's good then it's OK, but this is not always the case.

C. Turris Omnia and its Crowdfunding Campaign


Ondrej Filip, cz.nic

The presentation is available at:
https://ripe74.ripe.net/presentations/142-Turris-20170511-OF-RIPE-Omnia.pdf

Fredy Kuenzler, init7, said it was a great product and has started reselling in Switzerland but he said the price needs to be half the current price to be able to sell.

Niall O'Reilly said it would be useful to be able to pay in euros.

D. RPKI Tools


Andreas Reuter, FU Berlin

RPKI MIRO and RTRlib are the two open source tools they developed.

The presentation is available at:
https://ripe74.ripe.net/presentations/124-ripe74-open-source.pdf

Tim Bruinzeels, RIPE NCC, said they will work on RPKI Validator this summer and wanted to talk offline about they could use that next version.

Andreas said he also wanted to talk to the RIPE NCC so he was happy to hear this.

E. FRRouting

Martin Winter, NetDEF, and Attilla de Groot, Cumulus Networks

The presentation is available at:
https://ripe74.ripe.net/presentations/134-FRR-RIPE74.pdf

Yan Filyurin, Bloomberg, said he sees this in datacentres a lot and asked if they were going to use it for data routing.

Martin replied that this already happens somewhat and he asked people to try it out and identify bugs so he can fix it.

Bengt Gorden, Resilans, expressed gratitude for open source routing and he says this has been quite a leap forward, especially with the use of Quagga.

Blake Willis, Zayo, thanked Martin for doing so much work on this over a long period of time and asked to talk about sponsorship.

Francois Devienne, BORDER 6, had a comment on using Quagga, saying if you reduce the number of routes installed in the hardware and you actually focus only on the top prefixes that make most of the traffic and put different code on top, you can use a white box switch for BGP edge routers, and this was a use he sees more and more with customers.

Martin acknowledged that some people had concerns about Quagga and said if you are running on white box switches a lot of them have very limited forwarding space for doing hardware forwarding but that's outside of Quagga or free range routing. He added that there are lots of vendors working on getting boxes with larger tables in them.

F. Open Source Lightning Updates

Gert Döring, speaking as an open VPN developer, said there would be a release today and anyone using it should upgrade. He said the problem is that a packet can be sent that is malformed which the internal checks catch and the server will then exit, which is secure error handling, but a somewhat stupid thing on external data. He said the patch basically turns this into an error message and drops the packet – there is no remote code execution, there is no loss of data, but your server can be made to stop.

F.1) Go binding for RIPE Atlas


Jerry Lundström, DNS OARC

The presentation is available at:
https://ripe74.ripe.net/presentations/78-RIPE74-OpenSourceWG-GoRIPEAtlas.pdf

Martin concluded the session by asking people to consider if they wanted to become a working group chair and said he hoped to see everyone in Dubai at RIPE 75.