RIPE 70

Open Source Working Group Minutes - RIPE 70
Wednesday, 14 May 2015, 11:00 - 12:30
WG Chairs: Martin Winter, Ondrej Filip
Scribe: Alex Band

A. Admin matters

The working group chairs, Ondrej Filip and Martin Winters, welcomed the participants. The agenda was reviewed, there were no comments on the minutes of the WG session of RIPE 69.

There were no outstanding action items.

B. jFlowLib - Thomas King, DE CIX

The presentation is available here:
https://ripe70.ripe.net/presentations/59-e-TK-20150514-RIPE-Open-Source-WG-jFlowLib.pdf

C. OSNT - A Community-owned Platform for High-performance and Low-cost Network Testing - Gianni Antichi, University of Cambridge

The presentation is available here:
https://ripe70.ripe.net/presentations/68-ripe_14052015.pdf

Peter Hessler of OpenBSD asked if this requires special FPGA hardware or if it can work on a commodity Linux system. Gianni replied that the system is running on FPGA because this allows for high-precision time stamping, needed for high precision measurements. However, the code for the hardware is open source as well. If you have a different board you can use the same code and re-adapt the parts that require it. Peter also asked if they are focussing on performance, or also looking at compliance. Gianni said that this is something they are looking at too. Initially, they focussed on performance, but later realised that the system can do more. In the end, it's a community-based system, so we invite people to participate.

Stefan Plug from ECIX, Berlin asked if Gianni is implementing RFC2544 (network testing). Gianni said that they currently don't have this planned, but in the long-term, they envision an "app store" like experience on Github where developers can share implementations and code.

Aleksandr Saroyan asked if it's possible to connect multiple NetFPGA boards together and if so, what kind of bus or backplane will be used. Gianni said that essentially, there are a lot of FPGA boards available, ranging from one to four ports. When you have a board with multiple ports, you can connect them together. The plan is to port the framework to the new boards that have multiple ports, so that you can have more than 40Gb traffic generation. They will also have PCI-X Gen 3, allowing for the required bandwidth for this kind of testing.

D. Zonemaster - Do We Need Another DNS Testing Tool? - Patrik Wallström

The presentation is available here:
https://ripe70.ripe.net/presentations/136-Zonemaster-RIPE-Open-Source-wg-2015-05.pdf

There were no questions.

E. Building an Open Source IPv6 Configuration Architecture for OpenWrt - Steven Barth, OpenWrt

The presentation is available here:
https://ripe70.ripe.net/presentations/116-RIPE-2015.pdf

There were no questions.

F. Open Source Lightning Updates

F1. Lightning Talk: getdns API Implementation - Willem Toorop, NLnet Labs

The presentation is available here:
https://ripe70.ripe.net/presentations/135-getdns-oswg-lt-ripe70.pdf

There were no questions.

F2. Lightning Talk: Scapy, a Packet Manipulation Tool - Valadon Guillaume, ANSSI

The presentation is available here:
https://ripe70.ripe.net/presentations/142-Scapy.pdf

There were no questions.

F3. Lightning Talk: Cumulus Networks Automation Modules - Leslie Carr, Cumulus Networks

The presentation is available here:
https://ripe70.ripe.net/presentations/126-OS-wg-Leslie-modules.pdf

G. Closing Remarks

Peter Hessler warned that with regards to the presentation about custom memory management, he would like to remind people that exactly this was the cause of the Heartbleed vulnerability. It also means that monitoring tools don't pick it up and report it to the developer.