Skip to main content

You're viewing an archived page. It is no longer being updated.

Connect Working Group Minutes RIPE 84

Wednesday, 18 May 12:00 - 13:00 (UTC+2)
Chaired By: Florence Lavroff, Remco van Mook, Will van Gulik
Scribe: Matt Parker
Status: Final

1. Opening & Housekeeping

Remco opened the session by welcoming the attendees and introducing his co-chairs, Will and Florence, as well as the scribe from the RIPE NCC.  He shared the agenda for the session, approved the minutes from the previous meeting and thanked the RIPE NCC for compiling them.  

There were no questions. 

2. 25Gig FTTH - One Year's Experience, Do's, Don't's and Other Pitfalls

Pascal Gloor, Init7

The presentation is available at:

Pascal shared how Init7 rolled out a 25Gig FTTH service in Switzerland.  He explained the infrastructure that they put in place, compared the technologies that they and their competitors employ, as well as sharing some of the challenges that they faced along the way. 

Gordon Gidofalvy, Ting Fiber, commented that the equipment presented is not temperature hardened as you would expect from an access switch.  He asked whether this limited their deployment or whether they have a lot of actively cooled locations.  Pascal replied that they are not cooled, but that they had not really had any temperature issues. 

Rinse Kloek, Kloek Internetdiensten, asked what type of CPE Init7’s customers were using?  Pascal replied that for 10Gig it is not too difficult as you can just take a MikroTik which is quite affordable for the customer, however for 25Gig it is more of a problem.  It can only really be solved using a PCI card in a server. 

Ben Cartwright-Cox,, noted that the peeringDB shows that Init7 has only 10Gig ports and questioned whether they were concerned that a customer could theoretically blow two ports in one go.  Pascal explained that it simply doesn’t happen and reiterated that operators should not be afraid of the bandwidth. 

There were no further questions. 

3. Euro-IX Update

Bijal Sanghani, Euro-IX and Leo Vegoda, Euro-IX/PeeringDB

The presentation is available at:

Bijal provided some background information on the European Internet Exchange Association (Euro-IX) including the events that they host, the services they provide and the tools/community projects that they are involved in.  Leo introduced the IXP Database project which provides automated information about IXPs, provided by the IXPs themselves.  He asked the WG participants to contact him directly if they were willing to help them by providing some feedback. 

There were no questions. 

4. Transceiver 100G/400G and Beyond

Florian Hibler, Arista Networks

The presentation is available at:

Florian talked about what is new in the world of 100G/400G transceivers.  He discussed the SN connector which connects 4 individual fibres to a single optic, allowing a single link to be pulled out at any time, mitigating ‘fat finger’ mistakes during maintenance.  He also discussed the SFP-DD and DSFP standards, concluding that they offer two approaches to achieving the same outcome and that, in most regions, only one of these variants is relevant.  Finally, he spoke about what is coming up after 400G, including the upgrade path to 800G.
There were no questions. 

5. IXP Neutrality Project

Michalis Oikonomakos (replaced by Massimiliano Stucci)

The presentation is available at:

Massimiliano introduced the IXP Neutrality Foundation, explaining that its goals are to shed some light on how IXPs can be more neutral.  He asked the WG to look at their website and to contact Michalis Oikonomakos directly if they would like any further information. 

There were no questions.

6. The Death of the Peering Manager

Remco van Mook, Asteroid

The presentation is available at:

Remco discussed the role of Peering Manager/Coordinator and how that role has evolved over the past years.  He noted that increasing volumes of traffic are moving to low-latency, encrypted protocols and that end users were now more concerned about the quality of the connection than the volume of traffic that it can handle.  He concluded that automation is the only scalable way forward and stressed that high-fidelity, large-scale network measurements are needed to facilitate this automation.

There were no questions.

Will thanked everybody for attending the working group session and reminded them to rate the talks.