Re: Major Asia-Pacific Public Internet Exchange (IX) Points
----
Hello again. I thought that I would pass some additional Internet growth ideas along, since we're talking about NGI and I2 here. The above topic concerning public Internet Exchange (IX) points in the Asia-Pacific region was brought to my attention, most recently, on the North American Network Operator Group list, NANOG.
A poster on that forum was asking where such centers existed, for his route planning purposes. Reply after reply contained opinions, pointers as to where to look, and techniques to use to circumvent the apparent lack of any substantial IX locations, apparently, other than some vaguely defined IXs in Japan, some large peering points in Hong Kong that didn't exactly fit the bill either, and some smaller ones in Australia which are being considered for international IX and peering purposes.
Odd, I thought, that these operators wouldn't know. I'm not referring to the lower percentile of the proverbial 6,500 ISPs in the USA in this case... although they too are in the dark on this subject. Rather, I am referring to some of the major players who were involved in this discussion, for the most part. It occurred to me that there should be an entire list of these locations in ready lookup form somewhere on the 'net, but there is not.
In reality, the fact is that international equivalents to what we have come to regard as the National Network Access Points (the original NSF NAPs) here in the States, as well as the other special purpose exchange points which have joined the lot since the NAPs have not been able to handle the load (CIX, PAIX, MAEs, FIX, etc.) - although most of these have now actually been assimilated into routes just like NAPs - simply don't exist in those regions on the same scale and functionality as they do here in the USA. It was the same way in Europe, until very recently, as well.
Yet, we see Digital Island, and now others such as Exodus, expanding their global backbone reaches and private peering relationships throughout that region, and other regions of the world, for their own proprietary [read: closed network] e-commerce-supporting purposes.
And in the true spirit of up and coming competitive sectors, IMO, we shall see more operators (both backbones and ASPs alike) move to this isolated network modality (from a network topology and admission policy perspective) as time goes by, moving farther away from the "open" Internet architecture, in favor of more deterministic response times, security, multimedia-enabled, etc., and other e-commerce mandates.
As if to highlight this space, Mark Andreesen has now announced that he will be starting another "back end infrastructure" kind of venture, in order to assist new e-business web companies. The point here being that it will not be a business entirely focused on John Q. Public's browser concerns, as was his mission in the past. Instead, and symbolically to a great degree, and consistent with this thesis, it will focus on the concerns of businesses who will in all likelihood be riding over [very large] private routes, and specialized server farms, if I'm reading his intentions correctly.
Given these observations, then, it occurs to me that the "growth" of the Internet in the form that we have been accustomed to it up until now may have reached a crossroad of sorts, a juncture which may spell out quite a different set of expectations going forward than we would have expected just six or 12 months ago.
The Internet, for all intents and purposes, is in some ways now going private. AOL, itself, is in some ways a private, closed network. Of course, there are always the possibilities of backdoors and gateways to allow cross overs between private and public venues. But for those customers whom I am referring to in this post, the ones who will be serviced by these newer and exlusively e-comm players, they will actually be utilizing private backbone networks, for the greater parts of their reaches and server needs.
Proportionately, these growing private parts of the 'net (which can be viewed as exclusionary backbones, although some of then are actually riding over IETF-defined autonomous systems, or ASs) may begin to make up the greater part of all new abundant bandwidth routes, as operators seek more profitable e-commerce markets as opposed to the potentially lower profit www routes - which are often more problematical both from technological, and administrative, points of view, as well.
Has anyone else here contemplated what these changes I've listed above portend, going forward? Curious and looking forward to some input on this topic.
Regards, Frank Coluccio |
