This is from my swiss cheese memory ... The NAPs where in full swing by early 1995. We put together the MLPA at AADS by May 1, 1995 ... The date is on the web page: http://www.aads.net/MLPA.html Sprint, MCA, and ANS where the backbone providers brought into the NAPs ... originally viewed as the New York NAP, Chicago NAP and San Francisco NAP with MAE East tacked on as a legacy. MAE-East and MAE-West in part fed off the Federal exchanges, FIX-East and FIX-West/NASA. By the time of the MLPA the big boys where erecting fences in an attempt to keep others out of the club. I have a feeling that if peering was looked at purely based on economics it would make no sense for a large provide not to peer with a smaller provider if the later had a national backbone and could provide diverse peer points. The development of the network providers around DC in Maryland and Northern Virginia are tied into the shift from the University and Government Internet into the private sector Internet and the growth of MAE-East. That chapter alone would probably make for a good book. The history of the Internet so far does not support the view that the exclusionary policies of Sprint, MCI and ANS did anything to keep out competition, improve the respective networks, or keep them at the top of the heap. Bill Manning was involved in a lot of the exchange point development. I remember getting a crash course in exchange points from him on a marker board at ISI in late '95. Howard C. Berkowitz wrote:
At some point, there started to be a business case for large
providers to interconnect with bilateral private links as well as at exchanges. When did such links first get used for commercial traffic? In the beginning, were they short-haul connections between cages in exchanges, or WAN links between major provider hubs? I'm referring here only to interprovider links, not to transit customers.
Mark Borchers added,
Well, if this is to be a comprehensive list, private interconnects predated the commercial Internet. At OARnet in Ohio, we set up peering with CICNet in a facility near the Ohio Supercomputer Center in order to avoid long round trips.
Clearly, there's a need to broaden my scope!
I do want differentiate between good solid technical examples, such as yours for OARnet-CICNet, between examples where the suits had an economic motivation. The latter, I would assume, came later.
While I recognize I won't be able to publish the details of many NDA-covered commercial private peerings, it is my hope to identify when this practice began. A side motivation for this particular point is to identify when commercial NDA considerations might have restricted the potential ability of routing registries to give reasonably accurate representations of topology and policy.
Yeah, yeah...if everyone DID put all their policies in an RR. Yet Another Issue.
-- Joseph T. Klein +1 414 915 7489 Senior Network Engineer jtk@titania.net Adelphia Business Solutions joseph.klein@adelphiacom.com "... the true value of the Internet is its connectedness ..." -- John W. Stewart III