Update on our work on Internet Exchange Points

2 Comments

It’s been a little more than a year since we launched our Internet Exchange Point (IXP) initiative at CIRA, and we’ve made significant progress in that time.

A couple of weeks ago, the community celebrated the launch of Canada’s newest Internet Exchange Point, the Manitoba Internet Exchange (MBIX), in Winnipeg. In April the Montreal Internet Exchange (QIX) was launched in Montreal. These IXPs, along with TorIX in Toronto, OttIX in Ottawa, and BCNet in Vancouver are part of an evolving Canadian Internet infrastructure that is higher performing, more secure, resilient, and affordable. And, there are some productive discussions among the Internet community in Calgary about establishing an IXP in that city.

Through our research (.PDF) and our work with communities both global and domestic, we’ve learned a few things about what makes an IXP successful, and what doesn’t. Most importantly, I think, is the fact that there is no ‘one size fits all model’. Rather, while there key ingredients common in successful IXPs, each one takes on a local flavour.

What I have found to be critical is what I call ‘good governance’ – being open to understanding the local Internet community’s needs and being able to evolve as that community changes. It’s about operating the IXP in a transparent and responsive manner. We have also found that most often the IXs that work are not-for-profits and operate for the benefit of the local Internet community. They are located in facilities that are open for any organizations to peer with, and have the capacity to grow to meet local needs. They are also open to input and support from the broader community.

Let me give you a couple of examples.

QIX, the ‘new’ IXP in Montreal, came out of an already existing one managed by the Réseau d’informations scientifiques du Québec (RISQ), an arm of the Quebec provincial government. It was not open for any organizations to peer with, and was managed by RISQ. Once the need for an open IXP became apparent in Montreal, RISQ worked with the local community to enhance and open QIX, and establish a new governance structure. RISQ still manages the day-to-day operations of QIX, but this not-for-profit is now governed by an independent board of directors.

In contrast, MBIX was started from nothing more than an idea and a committed group of volunteers. Everything – from the technical infrastructure to the governance structure – had to be built from scratch.  The result is an open, not-for-profit IXP conceived of and built by the local community and run by a group of volunteers.

These two IXPs had very different beginnings and their current governance and operations have differences as well. However, they do have those ‘key ingredients’ I mentioned above in common: they are open and responsive to their local community, they are not-for-profit and are both located in a facility that is accessible and that can allow for growth.

When we started this initiative, our interest in establishing IXPs in Canada was driven by their key benefits: to improve the performance of the Internet in Canada through improved security, speed of data and network resilience. While we were also aware that IXPs can reduce the chance that national Internet traffic will travel to the U.S. Since then, the topic of IXPs has gotten a lot of traction in Canada. In light of the revelations that the National Security Agency (NSA) was monitoring Internet traffic crossing the U.S. border, the topic of IXPs has garnered significant media attention.

That discussion continues on Canadian Internet Forum, a national discussion on Internet-related issues hosted by CIRA.

I want to make two points clear. First, Internet traffic from all nations around the world destined for, or transiting, the U.S. can be subject to surveillance activity there. However, due to our geography and the configuration of North American networks, a large proportion – some say up to 40 per cent of Canadian domestic traffic that is traffic originating and terminating in Canada – transits the U.S., and is therefore affected by the NSA’s activities. This makes our IXP initiative more important than ever.

Second, while IXPs can reduce the chance that Canadian Internet traffic will flow to the U.S., that risk can NOT be eliminated. That’s not the way the Internet works. The Internet operates on the premise that bits of data travel through the fastest and most available route, regardless of national borders. Building more IXPs in Canada will build capacity, speed and resiliency in this country, creating opportunities for Canadian data to remain in this country. There is no way, however, to prevent all data to remain entirely in this country.

We have made significant progress and have continued to advance our understanding about establishing successful IXPs. I believe it is a good time to pause, reflect and celebrate our achievements. As a nation, we do have a long way to go before we can put Canada on the map as a digital leader with a robust network of Internet exchange points across the country. In the meantime, please find out more about your local IXP, or if your community doesn’t have one yet, get involved – contact us to help establish one!

Share on Tumblr
  • Bill Reid

    Byron, do not forget yycix in Calgary. They started this year and like MBIX started by volunteers from scratch.

  • Jonathan Stewart

    Good, informative post, Byron. As part of the MBIX volunteers, we were really glad for the help of CIRA. The equipement, and contacts, have allowed us to launch with much more interest and success than if we had been just alone.