Slashdot Mirror
2.4 Gigabit Network Demoed
coaxial writes: "At SuperComputing '99, the fastest network in the world, 2.4 gigabits, was built between the University of Washington and Microsoft's Redmond campus thanks to the DARPA-sponsored National Transparent Optical Network (NTON), the university's Pacific/Northwest Gigapop, and Nortel.
You can read all about it from the NCSA now apart of The Alliance . " Cool, MP3's and DECSS'd DVD movies at the speed of the light.
It's called marketing! Besides, if I remember correctly, not one, not two but THREE tests recently have showed that NT scales better than Linux. And not only scales better, but scales a LOT better. Of course they would use that kind of information in their marketing - and rightly so. Here's hoping that Linux becomes a viable alternative sometime soon.
Hey
This isn't meant to be offensive to the Slashdot crew or Nathan in particular, and I understand that the comment added to the news posting was most likelt humorous in nature. But: Do you guys think that
> Cool, MP3's and DECSS'd DVD movies at the speed of the light.
is a good impression to give to non-hacker types reading Slashdot? (hacking, not cracking) Of course, lots of nerdy techies like myself
But many of the industry people who do read Slashdot and try to form an opinion of Open Source (Unreal and Q3 developers, for instance, both of whom have posted here previously), will form the opinion that we're just a bunch of immature geeks who'll steal their software, if necessary.
Again, I know the comment was probably meant to be funny - but guys, we aren't alone here anymore. 'Suits' and industry people read Slashdot, and their opinion counts if we want to bring proper DVD support to Linux, for instance. Or if we want full Linux support for games, etc.
We're in a phase right now where everything that happens to Linux really, really matters. Let's not throw it away by acting childishly, and fitting into the 'zealot geek' stereotype many would like to categorise us as.
Sorry, Nathan - this really isn't a flame against you.
Alex
"Your telnet is talking to itself. Welcome to the wacky world of TCP/IP."
Microsoft/AOL created the internet therefore they are king of it. If you don't like it then get your god damned head out of your ass, and kneel to King Microsoft. They have done everything for you and you spit in their face. Microsoft has sacrificed greatly and this is the thanks they get? Some loser whining about Microsoft? Quit whining you baby, Microsoft is the reason you're here.
It was almost a year ago that frontier globalcenter turned up the first OC-48 Packet Over Sonet link between SF and LA, so this is less than impressive. That link was turned up into a production network, using Cisco GSR routers (which will take OC-48 POS cards with no sweat). I don't see where these people get off saying this is the first application.
...
OC-192 POS, on the other hand, would be something nice and new
good. fast. cheap. (pick any two, you can't have all three)
Good for yo, Canadians! Anyway, I dont get it. It seems that MessyDos has a *far* better PR department than my employer has (suprise :-). In the Netherlands KPN has a fully redundant 2.4 Gb network between 4 cities for internet connections already. See dutch kpn site
-- for undocumented cisco commands, take a peek @ dotu
"On a related note, does this technology work over
the standard glass fiber that we're all used to"? Heh, you people w/ nifty handware suck, i wish *I* was used to fiber... oh well, maybe once i graduate.
-dilinger
Theoretical? Nope! This network already exist across Canada.
Well, I believe there have been several 100 gigabit fibers running the entire width of the country for quite some time. Maybe if Microsoft does it it's supposed to be an innovation but Microsoft is really only one company. There are many other companies building networks much faster than Microsoft's.
If you are "lucky", the economy will generate something much more efficient than the "free market" defined by copyright law and patents. There will be gobs of cash available for those who know how to direct focus groups and shift paradigms: don't worry, you'll be able to watch more great, soul-searching films like "Titanic", "Jurassic Park", and "Lion King" and listen to great, ground-breaking music like N-Sync, BackStreet Boys, Menudo, and Brittany Spears.
But where will this gobs of cache come from? Currently these things get money from distribution of the product, basically on a pay-per-use basis, which is fair: the people who want the service pay for it. However, in the new economy, who will pay for this? It costs literally millions of dollars to film movies and record music albums. If nobody can make money off it, who will invest in it? The only solution I can see is government subsidies; however, this is highly communistic, because we all pay for it, regardless of it we want it. What is wrong with the current system, and how will the new system improves things? If you are a typical slashdotter, you enjoy movies such as The Matrix, Star Wars, Star Trek... All of which are extremely expensive to produce, but all of which did extremely well in the theaters and on video. It is simply not true to propose that these works can be produced without the money that was put into them. Nor would they have been produced if there wasn't a possibility of making millions and millions of dollars off of it. If funding for making movies is cut off -- who pays for development of movies?
Er, you're assuming that $2 is the maximum profit-generating price point for a movie. The current price for a DVD is $20-$25, and I'm sure the companies know what they are doing in pricing these, assuming they are half-way rational. I also think you are underestimating the costs of electronic distribution. The cost of delivering say 1,000,000 copies of Titanic (which would be, what, about 1 GB in size?) electronically would be massive, and certainly not economically feasible under current bandwidth limitations. In 10 years, when everybody has 2.4 GB connections to their homes, things may be different, but now most people have just 56k.
Some of you may remember a while back that incident in Ohio where a local util cut 4 OC192 circuts. The agregate bandwidth from those circuts was about 40Gbps. OC circuts are point to point circuts, perhaps what they have there is a shared network. I dont know where they get the idea that 2.4Gbps is the fastest network. Hmmm. If you really want to check out high speed , look at Project Oxygen(http://www.projectoxygen.com/). It is a global undersea fiber net currently in the first stage of construction. it has a min bandwidth on each segment of 2.56Tbps. That's right TERA bps. The first active segments should be on line by the end of next year.
But where will this gobs of cache come from?
I wish I knew the answer to this. If I did, I'm sure I would be a very rich man in the coming years. This I do know: the system as it is, while imperfect, has worked fairly well for the past century or so. However, it is clear that it cannot sustain itself given the progress of information technologies. A different approach must (and will) be taken, because our current one simply isn't effiecient any more.
Even if it was, I expect technology to give rise to a much more competitive method. This alone is cause to expect the current system to collapse, with or without the help of those "anti-patent" crackpots.
What will it look like? I have no idea. The only shift im seeing currently is one to a service based information economy, where people are paid to write specific code to accomplish a specific task for a (sold-per-unit) material good.
Is this, combined with the non-profit GPL enough? Probably not. It doesn't cover big budget movies, or music. But I'm positive other models will prevail.
Personally, while I enjoyed The Matrix, Star Wars, and other big budget attractions immensely, I would not shed a tear if they were the last of their kind. Especially if it meant total freedom from information exchange restrictions.
I realize others disagree, but I'm sure a useful solution will arise from the muck.
check out www.avici.com can you say 560 oc192's
The multiplexors are available from serveral vendors. For example Nortel or Fujitsu. For example, let us say you bought yourself some serious bandwidth. Your provider (be it a CLEC or god forbid, a LEC) would place what is called a add drop multiplexor at the point of service termination which would then take the signal to a common DEMARC such as a DSX 3 rear cross connect panel which you would then terminate your equipment into using coax cable...just for example. Anybody know what is the highest level a router or CSU can recognize? STS1? Or is that not a problem?
That's a pretty misleading statement. The M40 uses ASICs to route packets directly, so the OS kernel isn't involved most of the time. Also, the M40 doesnt's seem to run a standard BSD kernel.
More generally, let's please get over this notion that every OS needs to be best at everything. Making one OS do everything is Bill Gates's obsession and is, in a large part, responsible for the bloat and complexity of that operating system. If Win2k wants to become a backbone router with a Win32 GUI, an incompatible CORBA clone at its heart, and a built-in relational database, good for it, and good luck to the programmers at Microsoft.
I'm pretty happy with Linux network performance the way it is. I'm sure Linux will keep pace with the performance of standard networking hardware and software.
Linux/Intel the clear leader?
You surely meant linux/PPC the clear leader if you were talking about pure performance per watt?
Slainte Torc-- NSY - SY OOT - Doric signs on local shop doors.
If this is the speed of light, does this mean we'll be breaking that in the upcoming years (months)? :)
hear hear! However, back to the point. I do believe network throughput is critical to the future success of windows as a server. I know several companies that are now using network appliances for the actual file storage and front-ending them with NT, simply because we all know NT on intel hardware can't really even push 100mb and doesn't come close to gigabit.
It's somewhat disengenious to say that "Tests have shown that NT scales better, a lot better...". Assuming that you're referring to the Mindcraft studies, what the tests showed was that in terms of scaling by adding network cards, NT scales better -- which we knew already, because Linux 2.2 still has a single threaded TCP stack, which means that adding a card does nothing to the machine (only one card can be used at a time), whereas NT has a multi-threaded TCP stack and can therefore use all 4 NICs at once. In terms of scaling other things, Linux generally wins; Linux can also generally do more on smaller hardware, which means less scalability is needed in practice. Also, BTW, 2.3 has already fixed the single threaded TCP stack problem.
Much of what you say is true. Artist do have to be fed somehow, and the current system DOES work to create creativity and innovation, at least to some extent. There is no doubt a lot of good music drowning in the one hundred times more plentiful crappy music (though you are wrong when you speak of the margins being used for unprofitable artists, music companies only sign artist for profit, though sometimes the gamble turns out wrong (and trust me, you will get dropped)).
If things were all equal, I think I might vote for keeping the current system. But things are not all equal. Copyrights are at odds with the information society, and some of the things that the industries dependant on them are doing are clearly dangerous to liberty on the Internet and our society.
On a free Internet, piracy is going to become easier and easier, and more people ARE going to do it. There is no solution to this bar wing clipping the very nature of the Net or trying to make the punishments high enough, and lock up enough kids, that people will keep away from it by fear.
As such, we have to start looking for options. When I talk about the endless Backstreet clones of the music industry it is not because I think that is reason enough to kill it: but because I think people need to be reminded that the current system is not perfect. Many people simply take IP for granted to the point where they believe it to be a right: it is not a right, it is a construction, and far from the ideal solution to its problem (fostering creation and innovation) at that.
Of course we will need to feed our artists. But there other ways to do this. I cannot tell the future, but I am convinced that as long as there is demand for their art in society they will be fed. Perhaps it will be sponsorship (if an artist endorses Pepsi, it is there interest that as many people be listening to him as possible), perhaps it will be along the lines of opensource (where art evolves independant of one single person needing to be paid), or perhaps artists will be more at the mercy of benevolant fans (if 500,000 people download your album, and 10% send you one dollar, you are making a decent living). Perhaps it will be all of the above. Or none. But whatever we do, we need to recognize that not stepping on our liberties in the process is the most important thing of all.
-
We cannot reason ourselves out of our basic irrationality. All we can do is learn the art of being irrational in a reasonable way.
What the hell is a gigabit application?
They were probably just downloading all of the
mp3's from the UW dorms at once.
sheesh
WTF would you want to put a Gigabit connection into ANY intel box? That's the kind of thing for routers to handle and split up.
BUT NOT for dynamic content, where NT falls flat on it's face.
Now name me a mainstream site that uses no dynamic content.
Perhaps someone can explain why 2.4 Gbps is such a feat here. As has been pointed out many times before many networks are faster. So why is this considered the fastest? How does this differ from, say, the Gigabit ethernet I have on my machine?
-- Moondog
Win2k could handle it, but it would spend 99.95% of its CPU time handling the IO, with that .05% of CPU time left, maybe you could use edlin...
BSD can already handle it, check out the Juniper Networks M40. Its FreeBSD based, and can already handle a terabit, pretty snazzy stuff.
To be honest, when testing Linux Gigabit stuff, it couldn't touch my BSD systems for throughput or PPS. I had 6 boxes (all exactly the same) plugged into a Cisco 3500 Gig-switch, Two with Linux, Two with FreeBSD, Two with OpenBSD. Doing a lil bit of throughput testing between the FreeBSD/OpenBSD boxes left Linux in the dust. Kinda dissapointing see how much CPU the Linux boxes were using too.
Not nearly as bizzare as mounting a Linux ext2fs file system as 'sync' rather than the default. Get bonnie to write out a 1GB file and you'll be waiting for *HOURS* on a U2W 10k 9GB drive.
I'm not sure what Nortel uses for CANet but it could be OPTera instead of OC-192. It's cheaper and faster, although it's less mature. It uses DWDM. Nortel demonstrated the fastest DWDM system already and OPTera will be a huge part of their portfolio.
Heh. You got Nortel to thank for that :) They were practically the only ones to offer OC-192 for two years I think. Now DWDM is the thing of the future, since it's getting stuff like OC-768 to work is incredibly expensive in comparison. It's most cost-effective to just have OC-48 at a whole bunch of frequencies.
may i reminde everyone Canada has the fastest network? this pewnie little thing holds no candle to CA*Net3. i refer you all to Case evidance #1 http://slashdot.org/articles/99/08/28/1823211.shtm l
Maybe the naysayers who are so prejudiced against Microsoft can now admit that sometimes Microsoft does something really cool. I suspect that if this was an "open source thang" it would be considered the bee's knees. Let's give credit where credit is due, and say nice job, Microsoft.
You argue that copyright law has nothing to do with the creation of art, and point out Linux as an example... but even the GPL'd code in Linux is copyrighted. I don't think the problem is with copyright in and of itself. It has more to do with how the copyright holder uses that copyright. Yes, the RIAA and MPAA use their copyrights to make vast sums of money and limit access to the material, but it doesn't have to be that way.
All the people attacking the concept of copyright would do well to back up and make sure they've got the right target in their sights.
----
I expect good books will still get written, great art will still get made, great software will still be engineered, with or without this IP crap.
If you are "lucky", the economy will generate something much more efficient than the "free market" defined by copyright law and patents. There will be gobs of cash available for those who know how to direct focus groups and shift paradigms: don't worry, you'll be able to watch more great, soul-searching films like "Titanic", "Jurassic Park", and "Lion King" and listen to great, ground-breaking music like N-Sync, BackStreet Boys, Menudo, and Brittany Spears.
Have no fear. There will be plenty of great ways to make your mind turn to tapioca and make millionaires and billionaires out of people willing to tell the public what they need, quality be damned.
Actually, we talked about it ;) They had a couple of technologies that I only briefly passed over that would have allowed that sort of thing, at least beowulfing the beowulfs (although careful application of net cards and ip's can do that too). It's really too bad that Quake servers are not designed to handle being beowulfed and use it effectively.
-- Who is the bigger fool? The fool or the fool who follows him? --
True, the propogation speed is faster in plain old copper than in fiber.
I have noticed that anytime someone posts a pro-Microsoft post like this that they almost always post as an Anonymous Coward. Is this because they're too afraid of slashdot "retribution"? Or is it perhaps because a certain percentage of them are Microsoft employees?
And I've also noticed that the tone of these pro-Microsoft posts are almost always along the lines of, "See, you Linux zealots. Microsoft does turn out some good thing *after all.* That ought to show you! Microsoft rulez!" Even despite some serious evidence that whatever is being commented on can be debunked by more than one other story.
In other news, my corporation now has a Microsoft employee in our building who is helping us with an evaluation of Windows NT. I met him and shoot his hand, but I didn't really know how to react. As a representative of my company, I was cordial. But I honestly had a lot of questions that I wanted to ask him. Lots of questions ranging from pointed to downright insulting.
I wonder what I'll say come Monday.
I don't make the rules. I just make fun of them.
Isn't CAnet3 supposed to be 80 Gb?...and it's up and running accross Canada?
NGI side of things, there's a faster (2.5Gbps) pipe between San Fransisco and LA built by our friends at MCI as part of their vBNS+ network.
I'm a loner Dottie, a Rebel.
The important things to note about this demo that make it interesting:
(A) it used TCP/IP over a high-speed WAN, not LAN. This means huge windows for TCP. A killer bandwidth delay product. These are called "long fat pipes" and are something most stacks aren't designed to handle well.
(B) it had a small number of applications fronting the data; my understanding is that each individual app was reponsible for a larger part of the bandwidth than had ever been acheived before over a WAN. Certainly with Intel hardware.
>If it is a bit faster, blame it on the configuration of the Linux box, and anyway the independent organization was actually bribed by MS. Everyone knows a real website operator would know all the obscure ways to optimise Linux, would have applied all the patches, etc.
>If you hear that it is vastly faster and more scalable, whine that nobody needs the speed anyway because most websites would never use it.
You know who you are.
Most OC-48s are just aggregating a bunch of smaller streams together into one fiber. The big deal here is that they were doing gigabit end-to-end between two individual TCP stacks. That's impressive.
In other words, it's not the size of the pipe that matters, it's what they're doing with it.
2.4G is OC48 and many ISPs have had OC48 for over a year including us. Is this Microsoft claiming something AGAIN that isnt true?
(though you are wrong when you speak of the margins being used for unprofitable artists, music companies only sign artist for profit, though sometimes the gamble turns out wrong (and trust me, you will get dropped)).
You are forgetting that all of the major record companies also support classical and jazz music (in addition to pop). Most of these records sell in very, very small quantities -- fewer than 10,000 copies per title. On an title-by-title basis, they lose money off of classical and jazz, except for the VERY occasional title which sells 100,000 copies (Gorecki Symphony #3, for example). Part of the importance of pop music is that it funds classical music and jazz for the major record companies.
As for the occasional gamble: MOST of the gambles turn out to be wrong. I forget the statistics, but they say that about one in ten new artists makes it big time. For every Pearl Jam and Nine Inch Nails, there are ten other artists (which receive equal investment) which turn out to be losses. YES, they are dropped, but that doesn't mean the record company gets their money back.
Basically, a record companies catalog looks like this (proportionally):
Most people just loook at the Pearl Jam and assume that the company is raking in money OFF OF THAT ARTIST. But it's not true: no record company consists soley of Pearl Jam calibre selling artists. These high profile artists make up perhaps about 5%-10% of the companies catalog. The other 90%-95% of projects fail to make money. Most of the money which is made from the pop superstars is used to fund the unprofitable projects.
People will say to the companies: cut costs. This means cutting the less selling genres (e.g. classical and jazz) which is detrimental to serious aesthetics (which is bad for consumers). Alternatively, it means cutting "gamble" artists: but this essentially means no new artists will get exposure (which is also bad for consumers).
The record industry is nearly as rosy as it looks. There were times in the not so distant past that the industry was not doing well at all, and was considering quitting. If piracy continues, we will no doubt see that again.
I am of the firm belief that the relatively high margins on CD's (for superstar artists) has more aggressively promoted new music than ever before, and I point to the absolutely thriving music scene of the 90's as evidence. In the 80's, vinyl was the primary format, which was more expensive to manufacture, but sold for less. The 80's was a relatively quiet period for music, and fewer new artists got the chance to record. There were fewer ambitious projects, many fewer independent labels, and certainly fewer new and exciting genres than in the 90's, which has been the biggest renaissance in music in a long, long time. I believe that electronic distribution will bring us back to the essentially more conservative 1980's, with a lower manufacturing cost, but also a lower profit margin. With a lower profit margin, the companies have less of a chance to recoup money lost on risky projects, and will undertake fewer risky projects, which means that us consumers will have to live with fewer music choices, and music which is more and more homogenized.
Check out the folks who created the west coast of DARPA's network and extended it for SC99 - GST Telecom.
The offical PR is at http://www.att.com/press/item/0,1 354,2228,00.html
Bill Stewart
New Fast-Compression-only CPR http://preview.tinyurl.com/dy575ks
No, you are totally wrong here. If the record companies were knowingly putting out music that they don't expect to profit off, I will go buy a stock in Sony and then sue the fuck out of them. Public companies do not, should not, and can not act benevolently.
The reason for the companies agressively funding new bands is just the high level competition: they are always scared shitless to pass up on the next thing. I love your example of using Christina Aguilera as the new artist, since she is the perfect example of the bad side. Everything but a young creative talent who got a chance, she is more the result of record company thinking "Hey, we ought to put out a star like Britney Spears since that worked so well."
As far as classical and jazz recordings go, they simply budget the very, very low. Producing cds is so cheap (and even studio technology nowadays), that minus promotion putting out album that only sells 10,000 copies can still be a good idea. Here in Sweden there is a thriving Swedish language music scene, yet only a little under 9 million people in the world are possibly interested in that. They generally sell as little or less than American classical cds, but it is still possible to make a profit off that.
Interestingly, many people consider Beck to be one of the most important artists of the ninetees. You know why? Because he had a hit with "Looser" before he signed a record deal, and then afterwards played the companies against one another (instead of falling for the spiced up rip off offers most new artists get, for example TLC who got so little they went bankrupt the year that "Crazy, Sexy, Cool" sold 10 million copies) and got complete control over his music from square one. NiN and Pearl Jam are, interestingly, other examples of bands that were lucky enough to get good deals (NiN are on their own label) that have allowed them a lot more freedom then most bands have. And your right about these bands not making the most money (Pearl Jam haven't had a monster hit since Ten, and NiN have never been real profitable) for the studios. Artist like N'sync and Britney Spears, who the companies can sell 10 million bland albums by, throw them a few scraps (cars and fucks and glittery parties), and then throw them back in the studio to record more music on the companies terms are.
We are moving into the information age. Making information, replicating information, and spreading information is not so remarkable any more. The companies that have done the latter two can not expect that they can keep a profitable monopoly over it any more (as someone at the base of this string said "they'll have to wake up and realize its all just bits"), and the people that have done the former have to understand that the dynamics of their role is about to change: greatly.
You cannot stop the world, no matter how much you like it.
-
We cannot reason ourselves out of our basic irrationality. All we can do is learn the art of being irrational in a reasonable way.
OC-192 isn't even implemented anywhere yet. OC-48 is still fairly new to production backbone networks.
If the record companies were knowingly putting out music that they don't expect to profit off, I will go buy a stock in Sony and then sue the fuck out of them. Public companies do not, should not, and can not act benevolently.
They do not KNOWINGLY put out music that won't sell per se. They put out LOTS of music in the hope that ONE will sell well. They understand that in order to develop one superstar artist, they need to bring up ten artists, one of which will succeed. They invest millions of dollars in each one, but only one brings back any profit at all. As far as the classical and jazz is concerned, it is a little bit less clear about why they do it, but they don't show any signs of stopping.
Risk is inherent in the music industry considerably more than any other industry. A record company has absolutely no idea if a product will succeed when it is produced. The market is essentially irrational.
I love your example of using Christina Aguilera as the new artist, since she is the perfect example of the bad side. Everything but a young creative talent who got a chance, she is more the result of record company thinking "Hey, we ought to put out a star like Britney Spears since that worked so well."
But, again, I can name about three dozen girl singers off the top of my head who came out in the last year and failed. Aguilera is the one who succeeded. The profit made of Aguilera is not (profit_made_from_aguilera - cost_of_producing_aguilera), but rather (profit_made_from_aguilera - cost_of_producing_a_dozen_girl_acts) -- and that it not a huge sum.
I am not interested in qrguing aesthetics of pop music, but if you don't like her, the industry similarly brough up Pearl Jam, Metallica, etc., etc. These artists would not have existed if the record industry did not have excess capital to spend on new/risky artists.
And your right about these bands not making the most money (Pearl Jam haven't had a monster hit since Ten, and NiN have never been real profitable) for the studios. Artist like N'sync and Britney Spears, who the companies can sell 10 million bland albums by, throw them a few scraps (cars and fucks and glittery parties), and then throw them back in the studio to record more music on the companies terms are.
Here you seem to be admitting the existence of the subsidizing phenomenon I described above. And your position seems even more extreme than my take on it.
Pearl Jam is a mega-successful band. They are one of the very best selling acts of the 90's, and "Ten" in particular, is one of the best selling albums of any genre, of all time. I am not convinced that they do not make much money (as you suggest above), but if you think they don't, that emphasizes my point even more.
Even if you don't like Backstreet Boys and N*Sync, and the like, you admit that they help the artists you are interested in. So, basically you admit, that without them, there couldn't be more serious artists. This is why it's important to support the system: because it works so well. Complex music such as most classical and jazz will never reach a mass audience; but that's OK, it still thrives under the current system.
I personally am not interested in either Backstreet Boys or Pearl Jam, but the music which sells few copies -- fewer than 100,000 copies. This is the music which is at most risk in the new information age. In the new market, I am afraid that any music which does not sell 1,000,000 copies will not exist. Our only choice will be Backstreet Boys and Pearl Jam, both of which, in my opinion, are watered down, simplistic, and not very musical.
Producing cds is so cheap (and even studio technology nowadays), that minus promotion putting out album that only sells 10,000 copies can still be a good idea.
What gives you the idea that producing CD's is cheap? A classical CD costs $250,000-$500,000 to produce. It sells 2,000-3,000 copies in its lifetime. The profit comes from the Pearl Jam and N*Sync albums.
Yes, you said that already. I understand all this: profits from sure selling artists are used to fund risky artists, the companies do this because most of their sure selling artists were risky artists at one time or another. The current system does work to produce new music, some of it even innovative, no doubt, only its effiency can be doubted (and we really have no reference to compare it too).
Now you understand the core question here: the current system is not sustainable in the information society. Anyone who advocates for the continued use of copyrights is going to have to answer to how you are going to police them in the connected future without turning the Internet into a police state. You can argue yourself silly about how well the current system works today in it will do you no good: instead you should be asking yourself how well it will work in 10 years when most people have broadband connectivity and will be able to download an album illegally in minutes...
-
We cannot reason ourselves out of our basic irrationality. All we can do is learn the art of being irrational in a reasonable way.
Waa! Why do they get the cool toys first? Slashdot.org or cdrom.com would be the logical first place to install this...
--
"Convictions are more dangerous enemies of truth than lies."
Fighting FUD with FUD? Grow up
Microsoft, the National Computational Science Alliance (Alliance), the University of Washington (UW) and Sony (in support of the ResearchTV consortium) demonstrated two working, real-time gigabit applications in their coordinated SC99 exhibits.
Heh... gigabit applications, sounds like Microsoft bloating to fill yet another available space.
But seriously, isn't OC-48 ~= 2.48Gb/s? Aren't some Internet backbone providers already using OC-48? Or at least having it deployed in test environments/networks?
but, saying things like "Cool, MP3's and DECSS'd DVD movies at the speed of the light." doesn't really help our credibility. I mean, the greatest fear of the MPAA is that bandwidth will get to the point where entire DVDs can easily be pirated from computer to computer...
:)
Ok, the comment was probably tongue in cheek, but still... do you *enjoy* taunting lawyers?
----
Oh, great. Here come more "scales better than (insert target OS here)" claims from the FUDmasters!
It also takes up space on hard drives faster than anything else on the planet!
"...America's great minds of today, teaching America's great minds of tomorrow. Poor bastards." -- A Beautiful Min
all I need now is a NIC that will 2.4 gigabits, not to mention some splicing tools and a very long cable. I'll be cheesed if it is not hooked up to the internet backbone.
Buying a Dell computer is equivalent to dropping the soap in a prison shower.
Does anyone have any information about what kind of routers and multiplexors can make use of this cutting edge technology? I'm guessing that it's not your standard Cisco router hooked up to an Adtran fiber MUX. On a related note, does this technology work over the standard glass fiber that we're all used to, or does it use a brand new fiber?
With 24 100MBs cards and 128 processors serving up 20,000,000 static pages a second to a little old lady who was having problems getting the mouse to work.
Here in Canada we already have the fastest real-life internet-like network, with 80gbps of bandwidth (and still getting faster), thousands of kilometers of fiber, 9 access points, AND MSFT didn't put a penny in it!
This demoed network is like a 14.4 access to internet compared to CA*Net3.
#DEFINE QUESTION (2b)||(!2b) -- William Shakespeare
I noticed that were several comments floated by some of the Microsoft mouthpieces stating that this demo proved the viability of Win2000 as an OS that can handle broadband networks/applications. Exactly what does Win2000's TCP/IP stack have anything to do with this demo? Are they trying to imply that any other OS' stack wouldn't fare well in a similar demo? It doesn't sound right to me but I don't know much about Win2000's stack.
Hates people who have stupid little sigs
The specs are there but I highly doubt there is an actual implementation of it.
Considering that the company I work for maintains a national 2.5Gbps POS ring, I don't see how 2.4Gbps is faster. Maybe 2.4>2.5 for sufficiently large values of 2.4, but I doubt that this is the case. While many areas do indeed have multi-gig circuits coming in, they are usually broken up into smaller circuits. In our case, we keep it at 2.5Gbps all the way around the ring. And, since we use packet over SONET, there's almost nothing lost to overhead. I really don't see how 2.4Gbps can claim to be the fastest.
-NOC Monkey (OOK!) Experience is what allows you to recognize a mistake the second time you make it.
The answer is 'Yes, and planning to move to OC-192'.
Grunt. Oink, oink.
Bah, that's nothing compared to my dual OC-712s. At 9227.52MBps, I can kick your ass anyday in the bandwidth arena.
Chris Hagar
"The price of freedom is eternal vigilance." - Thomas Jefferson
I hate reading articles like this when I am still connecting at 56k. I mean originally the ARPAnet backbone was 56k and that was how many years ago???
It seems that some people didn't bother to read the article, and some people didn't check their numbers, and some people are guilty of both. I also note that the article is pretty poorly written, making it difficult to discern exactly what is so interesting about it.
I find the following facts in the article:
I only find the last bullet to be in the least bit news-worthy. The quote that chilled me was this one though:
"enabling gigabit networking capabilities on what will eventually be tens of millions of desktops is the first step in unleashing developers worldwide to create the next generation of applications, architectures and content."
unleashing is the correct word. just think of the denial of service implications of tens of millions of script kiddies with gigabit networking capabilities...
so I'll just make a test :)
:)
hope nobody cares
--
Weasel
"There are no shortcuts to any place worth going."
"Be regular and orderly in your life, so that you may be violent and original in your work." -Flaubert
Wait till you see what's down the road. 1.6 Terabit networks are already being built using cutting edge DWDM technology.
Like woodworking? Build your own picture frames.
Going to risk the karma to say its actually impressive that Windows 2K could handle this. My somewhat educated but unfounded guess would be current stable Linux kernels couldn't do it, but the *BSDs have a pretty good chance.
80gbps? so whats that mean? 8 nortel networks OC-192 bays maxed out at every gigapop? The article said there was 11, 11 x 8 = 88. What do they cost $500k each?
What I am getting at is I don't think this adds up.
Don't you need a regen every little bit (300 miles or so)? Is that where these gigapops are?
The article mentioned that it was all optical, so what do the use to regen the signal over all of Canada? I thought the only way to regen that signal over long distances was optical/elec/optical in an OC-192 bay?
This just doesn't seem realistic to have 8 bays at every gigapop, maybe I am wrong...If so I am sure someone will tell me.
I know I wanted to emigrate when I read about it. Heck, Canada's much further south as well. Practically tropical climate for me :)
- eivind
I'm just wondering, how do you plug this into a PC?
Are there like OC-48 cards for PC's?
Or would they use gigabyte ethernet? ATM?
Now, I REALLY wish people would try and make one even faster than this one so that this one mould then be "old" therefore "obsolete" aka "hell of a lot cheaper." First I gots ta get me one of those hard drive refrigerators with 500 gigs of space!
If you think you know what the hell is really going on you're probably full of shit.
If you think you know what the hell is really going on you're probably full of shit.
jdube is who I am.
Also many other companies are soing those fiber amplifiers for example Ericsson which is quite big optical component manufacturer.
OK, so it probably wouldnt have worked, but "imagine playing quake on that thing"
-Yarn - Rio Karma: Excellent
I don't follow your logic here. MicroSquish buys something expensive that they have no use for, so that means they "rock"?
BFD. I'm much more impressed by the guys who actually invent technology. Lucent rocks, HP rocks, IBM (yeah, IBM) rocks, Motorola rocks, and thousands of other companies rock.
Micro~1 is nothing but a pack of marketing dweebs.
-jcr
The only title of honor that a tyrant can grant is "Enemy of the State."
SLight bit of misinformation here...The fastest net in the world was at supercomputing, but it was not the oc-48 between MS and SC99. It was the LAN, which was multiple OC-192's, each wavelength multiplexed with 10 wavelengths. They could have gone higher, but as one of my coworkers stated, "they didn't need to and were feeling lazy". It was fricking insane (I was there). BTW, this wasn't some theoretical net, this was actually implemented complete with routers. As my coworker, who helped implement it said, "It's the hottest network on earth."
;)
In addition, many other things that were at SC99 will be of interest to the Slashdotters. One was the incredible number of Beowulfs. The real world computing partnership had a ~30 node one, SGI had a 32 node one, VALinux had a 16 node one, Argonne, LBNL, and LNL all had VaLinux clusters as well. Dell had parts of one to go to PNL. Lots of clusters. There was a Cray T3, many Onyx 2's and several other large systems.
One of the neatest things I saw was a Sun e450 w/ 2 gigs of ram and 4 processors. It powers 26 Sunrays with netscape, StarOffice and Smart cards. It ran extrememly fast, as it took all 26 going hog wild before I noticed a slowdown.
Another neat thing was this thing that all I can remember is the software, called DomeGL. Basically, they take a wide angle lens that can project without focal loss across a 180 degree hemisphere, project it inside a hemispherical dome in a darkened room and it gives a strong illusion of 3d, even without shutter glasses.
Lets see...lots of stuff for the big computers (ASCI Red, Blue Pacific, etc), a robot name sprocket controlled rather directly that spoke ina rather saucy voice, IBM's new display that has 200 pixels per inch and displays at (cannot remmeber exact, but this is close) 2640x2048, which, I was told, is just inside the visual accuity range of a 20/20 vision adult, demos of everyones stuff, compaq's alpha clusters with their optimized gcc, Alta's clusters, and all sorts of other stuff. If you ever get a chance to help set up, I highly recommend it. I got the opportunity because my employer is an exhibitor, but they do have volunteers
As for some of the nontechnical stuff, IBM threw a party in the Oregon Museum of Science and Industry. Very nice, and lots of fun for geeks. On Wednesday night, VALinux threw a party at the Lucky Labrador, a brew pub. Good grub and free beer. SGI and SUn also had parties but they were the same night as IBM and I didn't go. I also did not get to go to any of the technical program, but I hear it was very good .
I was gonna write up an official report and send it in, but someone beat me to it
-- Who is the bigger fool? The fool or the fool who follows him? --
There is some insane bandwidth across Canada but can't remember what it is. I believe it is more of a theoretical "if we had the cash to buy everything" network. As you said it is rather pricey. Then again Nortel is one of the private sponsors
As far as amplification goes you can do it all optically. A specially doped fiber optic is used. A laser is spliced onto the fiber and the laser light "pumps" electrons in the fiber up to a higher energy state. When a passing signal photon collides with one of the higher state electrons a second identical photon is produced. The signal is amplified. This is the paraphrased version from my fiber optics notes. Either JDS Fitel and/or Lucent is doing this.