Slashdot Mirror
Internet2 Gets a New Backbone
wrong_fuel writes "A few of you know that Internet2 and NLR (National Lambda Rail) have been in talks for some time regarding a merger of the two networks. Those talks have fallen apart and Internet2's contracts with Qwest communications had already been allowed to lapse. Internet2 has now reached an agreement with an unnamed carrier for its next generation backbone. The new network will likely be named later this year (the old one was referred to as "Abilene") and current member Universities will be migrated off of Abilene by September 2007."
I love those 5MB/s downloads from the open source software mirrors at other universities; even ones which are not too close to here (Pittsburgh) are really fast. I love you, I2.
Check out this page -one of the best examples from it:Also check out UMD's page: What kind of computing jobs are best paralellized with such network?
Anything easy enough for casual programmer to start working on?
Its not so much for computing jobs as use for researchers who require high bandwidth & low latency, or are conducting advanced network research
There are shills on slashdot. Apparently, I'm one of them.
I read a paper on the justification of high bandwidth systems recently. It outlined as one point, how society has always managed to fill the extra bandiwdth with data, reguardless of what that data may be, increasing the rate of dissemination of data amoung people all over the world. I can only imagine the same applies for scientists.
e r) being built by CERN, which is expected to produce data in quantities thousands of times greater then previous accelerator experiements. The need to disseminate this data to locations around the world is critical to its analysis.
0 /MRO_data.jpg
The article gave the example of the Large Hadron Collider (http://en.wikipedia.org/wiki/Large_Hadron_Collid
http://upload.wikimedia.org/wikipedia/commons/1/1
The Mars Reconnaissance Orbiter is expected to produce fairly large quantities of data also.
Along with these are that thousands upon thousands of experiments and measurmeents being taken every moment around the globe. All this data requires storage, transmission and compution. Weather simulations, aerodynamics, radiotelescope data, biochemical simulation, the list goes on.
Of course, if the sheer number of information producing tasks arn't enough, the definitive agument to why so much data is being generated is that with the increase of bandwidth and the power of computer, so too has the accuracy and speed of data collection increased. The micosecond is slow for todays chemical, physical and biological science.
Overall, its the number of experiments, the accuracy, resolution and speed of data generation, and the need for that data to be analysed around the globe that has created the mutual need, and provision of huge bandwidths such as those being investigated and used by I2.
For everyday folk like you and me, just go down to your accounting deparment and ask them how large their largest database is, you'll be suprised how unbelieveably data and bandwidth consuming financeal data has become since the revolution of the internet.
Japan has high population densities basically everywhere, so it's economically feasible to bring broadband everywhere. Nobody is very far from a local head-end installation (cable or telco), which is the limiting factor in bringing DSL and cable-Internet technologies to people in most places where it's not available now.
I'm willing to bet that the same situation is true in Sweden: those "remote villages" you're talking about aren't very big, and they're probably easier to wire for broadband than typical suburban-sprawl America. Although I'm sure the overall population density of Sweden is very low, I'm pretty confident that the density is distributed unevenly: small clusters of relatively high density (a village), separated by great distances. So again, you can bring the backbone, via microwave relays or fiber probably, out to the village's headend / telco building (the DSLAM), and then from there most of the subscribers are probably within cable modem or DSL range.
It's the same reason why I'm confident that Canada will achieve (if it hasn't already) greater broadband access than the U.S. to probably 80% of its population: a very large part of the population is concentrated in urban areas in a relatively small area of the country, contrary to what you'd expect if you just looked at an overall "persons per square mile" figure. Of course, that last 5-10% of people who don't live in the urban areas and are out in the Northern Territory or on farms in Saskatchewan are going to be a real bitch. In the U.S., we've already hit that limit: most people living in urban (and most suburban) areas have some type of broadband available. We're at that "last x percent" already, only in our case, x is very large due to the type of low density development that's common across much of the country.
The corporate-conspiracy stuff may play well, but there's very little truth behind it. If it were economically feasible to give every trailer and farmhouse in the boondocks of Pigs Knuckle, IA broadband, I'm sure all the providers would be falling over themselves to do it. But you can only cover so much area with broadband from a DSLAM, it's a pretty much fixed radius (I'm not sure exactly for cable but on DSL it's generally ~18000 line-feet); if you don't have people clustered together, that quickly becomes impractical. Heck, there are still places where cable TV is impractical, and it has a much larger radius from the head-end than broadband.
Wiring for broadband isn't a walk in the park. It's a pretty significant upgrade to systems that were only ever intended to carry frequencies up to a few thousand hertz, and whether you're a corporation or the government, at some point you have to do a cost/benefit analysis. It's not worth it to roll out $100,000 worth of infrastructure if it's only going to gain you 10 subscribers at forty bucks a month. Sure, you could subsidize the hell out of that development with tax money, but I think there are a whole lot of things that our taxes should be spent on (like, I don't know, teaching people to read) before we go throwing vast quantities of money at the problem, especially when the technology isn't mature. (And I think based on the lack of support for govt-subsidized Internet, this is pretty common.) We'd just barely have the whole country wired for 1MB cable and probably only be started paying off the trillions of dollars that it would cost, when people would be saying "one megabit?! Damn, man, you might as well be using 2400 baud. You can't do anything without [FTTN/FTTC/802.11n/$new_networking_technology]!" And we'd be off again.
I remember it wasn't that long ago when people were talking about getting universally available Internet access. Not free Internet, not high-speed Internet, just the AVAILABILITY of a local ISP to everyone in the country, without having to make a long-distance call. I'm pretty sure we made it there sometime during the Boom, but did you hear anyone talk about it? I didn't. Because by the time we actually found that goal, people
"Ladies and gentlemen, my killbot features Lotus Notes and a machine gun. It is the finest available."
Most point to point fiber connections have 2-10 ms latency. That's a slow LAN, but hellagood WAN latency, especially if you're coming from a DSL/T1 world. Generally an order of magnitude faster.
There are two types of people in the world: Those who crave closure
In case anyone was wondering...
The Meaning of Liff .sig under construction...