I phoned Sprint asking them to look into it - and I do have to say they were very responsive and conducted extensive testing - and the bottleneck turned out to be, surprise surprise, MAE-WEST.
Actually, I'd be more than willing to bet money that it's NOT actually MAE-WEST itself that is the problem. What has been happening is that the major NSP's have been redesigning thier networks to try and get a handle on the incredible growth rates. The "big boys" have gone to a private interconnection type of arrangement that allows the traffic to flow between them at multiple points with much less congestion. (Jeff and Jack from MCI and Hank or Kurt from Sprint could possibly elaborate upon this topology) The one outstanding problem that results from this is that the resources of thses companies are no longer directed at the meet points as an important (allow me some license here guys) part of thier network. So now all of the traffic originating from peers at the meet points moves happily across the meet point to the NSP's router, and then hops onto a DS-3 to travel from the meet point router to thier backbone. There are two achilles heels here. The port at the meet point (a switched FDDI port still can only handle 100Mbps simplex), and the link between the NSP's meet router and thier backbone. I think in most cases one will find that these are the locations of the problem, and it's not really a problem that can be fixed by adding more hardware or new technology at the meet points. It's more a policy and financial decision that must be made by the NSP's.
Since all the discussion on what backbone (tier 1) providers can do to improve interconnection seems to have done something - intra-provider traffic seems to be much better nowadays than it was last year - can we start doing the same with major exchange points such as MAE-WEST and MAE-EAST, etc?
Again, I think one of the guys I mentioned above could better express thier ideas here...
If the problem is that the routers/gigaswitches are overloaded, perhaps some sort of intelligent traffic analysis and division might be in order?
Now this definately has merit. Strong planning (a point that I'm weak in) by both the Exchange operators and the NSP's could prevent a large part of the congestion that shows up. But again, you must look at things from the other side. If I give peering to someone (say a regional ISP), I have to invest money into my network to handle the traffic load they send me. If I sell a circuit to that regional, then there is a revenue stream to maintain the network. Peering has become a financial decision in addition to a network engineering decision. This is really where some type of settlement method needs to be established if peering as we know it today is to remain a viable method of interconnecting the networks. Chris A. Icide Senior Engineer Nap.Net, L.L.C.
In other words, the big players don't like the "open" naps and are deliberately not installing sufficient bandwidth to them? -- (313) 741-4442 http://branch.com/ Jon Zeeff Branch Internet Services Inc. jon@branch.com *** WWW Hosting Services, WWW Site Development and the Branch Malls ***
On Mon, 21 Oct 1996, Jon Zeeff wrote:
In other words, the big players don't like the "open" naps and are deliberately not installing sufficient bandwidth to them?
No, the open NAP's are bad engineering and the big players are fixing the topology by routing around them. Visualize a typical garden spiderweb. All the main strands lead to the core of the web but there are many short transverse strands that provide alternate routes for the spider to travel. Now visualize the Eiffel tower. There are major support beams interlaced with many shorter braces. Now visualize a typical statewide highway grid. There are major highways converging on a city but there are always bypass roads around the city unless the city is a very small one. The Internet absolutely needs exchange points. And all but the smallest exchange points absolutely need bypasses. I think that right now we are in a phase where the need for bypasses (private exchanges between two parties) is quite clear but the deployment is lagging a bit because of this tremendous wave of growth that everybody is trying to keep up with. If you've read about the distribution of market towns in Ancient Mesopotamia then another thing that we can expect to see is a regular pattern of small exchanges everywhere with the occasional larger regional exchange and a handful of major national exchange points. I'm not sure yet whether the local ISP qualifies as the "small exchanges everywhere" or whether this role will be filled by metropolitan exchanges in every city. Time will tell I suppose. Michael Dillon - ISP & Internet Consulting Memra Software Inc. - Fax: +1-604-546-3049 http://www.memra.com - E-mail: michael@memra.com
In other words, the big players don't like the "open" naps and are deliberately not installing sufficient bandwidth to them?
No, the open NAP's are bad engineering and the big players are fixing the topology by routing around them.
But people keep claiming that some naps are well engineered and have excess capacity and that the problem is the size of the pipes leading to them. Adding more pipes is easy enough, leading to the conclusion that there are other factors at play.
Visualize a typical garden spiderweb. All the main strands lead to the
But a spiderweb is for trapping things. Visualize Federal Express and their "super nap" model for moving packages.
Hola,
In other words, the big players don't like the "open" naps and are deliberately not installing sufficient bandwidth to them?
No, the open NAP's are bad engineering and the big players are fixing the topology by routing around them.
No no no no no.
But people keep claiming that some naps are well engineered and have excess capacity and that the problem is the size of the pipes leading to them. Adding more pipes is easy enough, leading to the conclusion that there are other factors at play.
Yes, well, maybe, no, not really. (except about the more pipes, see [*]) The NAPs are well engineered (for the most part) and are rarely if ever the source of ANY packet loss, ANY latency, or ANY congestion. Period. (well, not really... the links into the IXPs are often full, but not because they "can't afford another link" or "want to squeeze someone" but rather because they put all the sympathy peers on that one link/ixp (hint: rhymes with Twit, sort of...) The ISSUE at stake here is "how well connected" a particular ISP/NSP is to all the other NSP/ISPs. Now examine at the (virtual/peering) connection between two providers, Bigger and Smaller. Bigger wants Smaller to pay. Smaller wants to pretend they're bigger, and not pay. Bigger tells Smaller to take off. Smaller then has to buy connectivity to Bigger, and generally gets a poor-er connection (more hops). Larger wants to peer with Bigger. Bigger wants to peer with Larger. Would they rather peer betwixt themsleves, or through a public exchange? Customer sits behind Smaller. Customer signed a contract w/ no QOS. Customer should give money/resources to IPPM. In all of the above three issues (granted they're created for the purpose...) the larger issues is economics and politics. Not technology or capacity. This whole thing smacks of child labor, minimum wage, and unsafe medicine. The public is too silly/stupid/naieve to demand quantifiable numbers tied into contracts. Elysium forbid that the government ever mandates it.... -alan [*] adding more pipes into an IXP is not a trivial issue, certainly it is possible to design a system that allows you some semblance of balance across the lines, but the linear growth of lines is not at all scalable, especially as the traffic grows in certain regions and the economies of scale for regional/private IXPs increase.....
In other words, the big players don't like the "open" naps and are deliberately not installing sufficient bandwidth to them?
No, the open NAP's are bad engineering and the big players are fixing the topology by routing around them.
But people keep claiming that some naps are well engineered and have excess capacity and that the problem is the size of the pipes leading to them. Adding more pipes is easy enough, leading to the conclusion that there are other factors at play.
In fact... :) To borrow from your FedEx model.. We can build bigger doors (MTU) but the conveyer belts are stuck at 100Mm/sec. We need faster conveyer belts! (waiting for HPPI-64 or G-Ether technologies w/ baited breath :) And then watch out 2-4Gbps backplanes... (that is aggregate right?) -- --bill
I point you to the notes of D.C. Nanog, particularily the Sprint discussion. http://www.academ.com/nanog/may1996/sprintlink.html#Notes I don't remember for certain the figures that Jim showed on the detailed map of Sprint's National backbone, but I do recall a 90%+ utilization figure on mae-east and mae-west links by Sprint. I'm sure many others are very near that as well. Don't mean to pick on Sprint, but they made that information public at NANOG D.C. I have no idea what the "intentions" are of the "big players". Don't even want to get into that, as the only people that know the real intentions, are those making the decisions not to upgrade their links. What is fact is: the IXPs are not over utilized.. the links of the carriers entering the IXPs are. ROb
In other words, the big players don't like the "open" naps and are deliberately not installing sufficient bandwidth to them?
-- (313) 741-4442 http://branch.com/ Jon Zeeff Branch Internet Services Inc. jon@branch.com *** WWW Hosting Services, WWW Site Development and the Branch Malls ***
participants (6)
-
alan＠mindvision.com -
bmanning＠ISI.EDU -
Chris A. Icide -
jon＠branch.net -
Michael Dillon -
Robert Bowman