Engineer internet peering

#LINX83 – traffic growth and regional expansion

I’m at LINX83. The eighty third quarterly meeting of the London Internet Exchange. The statistics associated with LINX continue to astound.

LINX has 492 members, 1191 connected ports 683 of which are 10GigE and 1 100GigE port with 2 orders in process. The LINX network carries 1.791 Tbps of peak traffic with 7.324 Tbs capacity (plenty of headroom there) to 60 member countries.

This is an enormous amount of traffic and capacity upon which I’m not going to dwell much more (unless as I’m sure it will, something interesting comes up during the LINX83 sessions).

I do however want to talk about regional peering efforts.

Regional Internet Exchanges are opening up all over the world, Indeed there are regional IXs opening in the UK – Leeds, Manchester and Edinburgh. Wherever there is a chance of a sufficient density of networks to pass enough traffic between each other to justify the setup costs.

The three big exchanges are AMS-IX (Amsterdam), DE-CIX (Frankfurt) and LINX.    Each of these is pursuing some sort of global expansion away from their home base with the USA being a specific target.

The mutual not for profit nature of the IX game is a particularly European one. Exchanges in the USA and Japan (for example) are set up as profit making businesses.

The idea of a not for profit IX that is specifically based in London, as is the case with LINX, seeking to expand into other areas of the world has been a difficult one to get my brain around. However I accept that we appoint people to run these exchanges and in LINX’s case the board consists of some very senior heavy hitters from the internet networking industry. I therefore have to buy in to the expansion strategy – these people are my friends.

If multiple large not for profit exchanges are seeking to open in the same countries as AMSIX, DECIX and LINX are in the USA there is a natural progress in my mind for the facilities offered by these exchanges to converge. This is not a competition. This is an activity run for the mutual benefit of the members.

Many network operators are members of all three of these large exchanges. It makes sense to me that there should be just one large mutual organisation serving the whole internet industry. Mutuality is in the DNA of the internet engineering community.

Discuss. I’m sure there will be different views out there.

Trefor Davies

By Trefor Davies

Liver of life, father of four, CTO of, writer, poet,

7 replies on “#LINX83 – traffic growth and regional expansion”

Hmmm it’s a fair point you make Tref but I’d be worried about the admin tedium that would be required in order to bring all three organisations together. You also have to ask, what would be practical advantage to ISPs? What would you get that doesn’t already exist?

I suspect there might be some cost savings but then not-for-profits by nature tend to be quite good at this and the admin required to bring all three together might also add some higher initial costs.

I also share Mike’s viewpoint that in the world of networking it doesn’t hurt to have choice and a bit of diversity. Mind you I’d need to take a closer look at the AMS-IX and DE-CIX design to see how diverse the existing diversity really is.

On the other hand it never hurts to explore a new possibility and I’d be interested to see what LINX itself thinks.


I actually think the three smaller organisations are easier to capture.

but the one positive item is that three different architectural directions so less likely of cascade failure.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.