In this session we're going to talk about internet exchange point history and internet exchange point design and before we go into the design details I'd like to take a little bit of a review of where the internet exchange points came from. This is not a complete history but it's intended to show where the internet exchange points came from and why they were considered to be so important back in the earlier days of the internet and why today we simply cannot live without them so we start back in the late 80s there was the National Science Foundation Network which was one major backbone it was very much the Internet as it was known back then it was funded by the United States National Science Foundation and it connected academic and research institutions together on this early network infrastructure it also connected a few private company networks under an acceptable use policy at various network access points across the United States and the acceptable use policy was very strict it said no commercial activity and we'll come back to this there were four network access points one in Chicago run by Mary tech one in New York run by Sprint one in San Francisco run by Pacific Bell and one in Vienna Virginia run by MF s you'll note that these four operators were all major telecom providers in the United States as the internet developed private companies needed to interconnect the networks to they could certainly access then SF net through these network access points but these network access points prohibited the use of commercial traffic across the interconnect so they could not use the NSFNET for example a commercial operator on the east coast of the United States wanting to talk to a commercial entity on the west coast of the United States could not cross the NSFNET they needed their own way of getting there this resulted in the first commercial ISPs being formed and of course it resulted in the first commercial internet exchanges being created as well the two best-known were Kix West which was on the west coast of the United States in the Bay Area and there was Mae East on the east coast of the United States in Virginia along with these the routing arbiter project was created to help with the coordination of routing exchange between providers traffic from ISP a needed to get to ISP be and the routing arbiter helped document the policy used to achieve this the routing arbiter is no more and it's long since been superseded by today's internet routing registries the IRR in fact the RA DB is well known to many operators today and it is the only remnant of this routing arbiter project the end of the NSFNET in 1995 meant the move towards full commercial internet private companies were selling the bandwidth in north america and europe and in parts of asia the network access points established leighton NSFNET life was some of these original exchange ponds the issue was the NAP operators were providing commercial Internet access as well and so the Sprint pacbell and Ameritech naps disappeared and were a place by the neutral or commercial internet exchange points the NFS hosted Mae East replaced the Vienna nap in more or less the same location ans won the contract to operate the late NSFNET were forced to join these exchange points as well a global distributed global internet exchange was proposed in the mid 90s but this never happened it was planned to be cakes west Mae East the new Swedish exchange point and a proposed interconnect in Paris the Swedish exchange point was formed in Stockholm late 1992 three major ISPs interconnected and the main aim there was for latency reduction and performance gains and of course to keep local traffic local links was formed in London in 1994 the 5 UK network operators at the time interconnected again for latency reduction performance gains and so that local traffic stays local the main feature of links at the time was in absence of any European telecoms deregulation was to reduce the huge cost of crossing the Atlantic Ocean to get to the u.s. part of the internet in fact links was proposed to be part of the DAX when the Paris option fell through but as I mentioned led cakes never happened in Asia HK IX was formed in Hong Kong in 1995 again a vibrant internet community several small operators again looking for latency performance and local traffic benefits rather than having to halt traffic all the way across the Pacific Ocean to interconnect in the United States and this model continued with several other European exchange parts been formed in 1996 1997 and so on today we know about m6d cakes very well known exchange parts in Europe and were joined by many other smaller ones in several European countries the exchange point was here to stay so what is it well an exchange point is a neutral location for network operators freely interconnect the networks to exchange traffic physically it is nothing more than an Ethernet switch located in this neutral location how most exchange points work today is that the exchange point host whoever that may be provides the switch and the Rackspace for that switch and the devices that accompany at the exchange part the network operators usually bring routers and interconnect them via the exchange point fabric it's a very very simple concept any place where providers meet to exchange traffic so what are they well an exchange point today runs at layer two and we use Ethernet Ethernet add a hundred gigabits per second at the largest exchange points ten gigabits one gigabits are very very common interconnect speeds one hundred megabits we're seeing less of as the one gigabit technology becomes so readily available on even the low-end managed switches there were several other technologies used for exchange points in the past including ATM frame relay SRP fiddy and SMDs I'll leave that as an exercise to you to look them up in your history books there were also layer three exchange parts and this is now historical as well it was router based and the best known example was Kix West it was a very early model cisco router that went through a few iterations but the router very quickly became overwhelmed by the sophisticated requirements of the rapidly growing Internet. There were several issues including manageability, scalability and general throughput and so the layer 3 exchange point is no more and is not an option for exchange points created today.

© Produced by Philip Smith and the Network Startup Resource Center, through the University of Oregon.

Attribution-NonCommercial 4.0 International (CC BY-NC 4.0)
This is a human-readable summary of (and not a substitute for) the license. Disclaimer. You are free to: Share — copy and redistribute the material in any medium or format Adapt — remix, transform, and build upon the material The licensor cannot revoke these freedoms as long as you follow the license terms. Under the following terms: Attribution — You must give appropriate credit, provide a link to the license, and indicate if changes were made. You may do so in any reasonable manner, but not in any way that suggests the licensor endorses you or your use. NonCommercial — You may not use the material for commercial purposes. No additional restrictions — You may not apply legal terms or technological measures that legally restrict others from doing anything the license permits.