Follow Slashdot stories on Twitter

 


Forgot your password?
Close
typodupeerror
×
Technology

Lucent Makes 10 Terabit Router 85

Posted by Hemos from the i-love-break-thrus dept.
maladroit wrote to us with an interesting press release from Lucent Technologies. The former Bell company has made a breakthru using microscopic mirrors in fiber that will allow data to be carried at upwards of ten times current speeds. Supposedly, a ten terabit router has already been created, and it will be rolled out to the public towards the end of next year.
This discussion has been archived. No new comments can be posted.

Lucent Makes 10 Terabit Router More

Lucent Makes 10 Terabit Router

Comments Filter:
  • Now that 60+ Mb DSL becomes much more feasable. Yea Baby!

    -Al-
  • is "wow!"

    That is some serious data transfering! (does that sound stupid to anyone else?)

    Seriously:
    Will any other hardware be able to keep up with this any time soon? Or is this just a ISP trunk-type device that users won't see for a very long time?

    Will this be availible to (for instance) my local ISP, giving better throughput?



  • If companies start incorporating this technology into their networks, will we see:

    1) an end to the "slashdot effect"
    2) download sites that have enough anonymous logins for everyone
    and 3) voice over IP and streaming video for trivial cost

    or simply more "First post" and "Wow, think of a Beowulf cluster of THOSE things!" messages to Slashdot. In fact, think of a Beowulf cluster using 10 terabit switched channels... *SMACK!*

  • If this means I'll be able to finally get decent speed on my cable modem (Austin is rather congested), then I'm all for it. On a side note: maybe those gigabit ethernet cards will finally have a purpose...
  • Offtopic, but this highlights the main reason I believe Lucent will dominate over Cisco.

    Lucent continually impresses me with their labs... breeding new technologies from within. Cisco, on the other hand, devotes very little to research. They do have an agreement with IBM for work on e-business applications but they are forced to buy the majority of their new technologies. While that is easy for them right now, with their $200 billion plus market cap, no debt and huge cash reserves, that can't last forever. Hell, the price thatthe paid for Cerent was simply outrageous! $7 billion for a company with revenues (and no profits) of approximately $20 million a year!?
    Not to forget them buying Montery Systems, Transmedia, StratumOne, MaxComm, Calista, KPMG Consulting... hell, there are over 20 in the past 6 months.

    In the long run, it just seems so much more cost efficient to invest that money into your own research and development, like Lucent does. But, I could be wrong.
  • Or perhaps you could learn to spell like a normal human-being. Why don't you grow up, you twit.

    Isn't there some way that Rob could implement some kind of script that would search for keywords (script kiddie nonsense and f1rst post bull) and filter them out?

    I would rather a moderator spend his/her points in getting worthwhile things moderated up.. than worthless trash like this moderated into oblivion.

    --


  • This is more like a very special patch panel than a router. It is NOT a breakthrough in optical computing, or a computing device at all. It uses tiny mirrors (and maybe some tiny smoke *grin*) to direct different wavelengths of light coming off one port to different ports. It apparently does NO routing or inspection, it just sends light from one fiber to another.

    Not to say this isn't interesting, but let's not get carried away and declare Lucent the ruler of high-speed optical routing yet!
  • You can already make free domestic (including AK and HI) PC to POTS (or PC to PC) calls at dialpad.com.

    Admittedly they don't have enough lines to meet demand yet, and there is a little latency effect when you do get an open line. But as far as costs, I expect it will be awhile until they start paying us place calls.

  • This is phenominal! With gigabit ethernet drivers coming out for FreeBSD and Linux, and terabit routers, corporate networks have the potential for becoming VERY VERY FAST!

    (Unlike the present-day situation, where most corporate networks remind me of a concussed ferrit in a molasses-filled pipe, with concrete boots. Though what a pipe is doing, wearing boots, I don't know. :)

    Personally, I would like to advocate that Lucent promote this technology to the world at large, by offering 4 such routers as prizes in a sweepstake, held on Slashdot. What does everyone else think?

  • I would rather a moderator spend his/her points in getting worthwhile things moderated up.. than worthless trash like this moderated into oblivion.

    Not to sanction this kind of behavior, but stuff like this is something moderators should ignore anyway. Like the guidelines [slashdot.org] say, use your points to moderate worthwhile posts up, don't worry about stupid things like this.

    -A.P.
    --


    "One World, one Web, one Program" - Microsoft promotional ad

  • The WaveStar LambdaRouter, as the product is known, operates at speeds that would, in effect, allow the printed contents of whole libraries to be transmitted in a single second.
    Almost any communications medium would allow this. Just change the size of the library. It's a meaningless phrase!

    And its capacity will grow even further over time.
    Does this mean it will upgrade itself automatically, for free?

    Argh!

  • Wow. (Score:1)

    by volsung ( 378 )
    Watching data transmission technology improve is like watching a tennis match. I keep oscillating from:
    Gee, what are we going to do with all that bandwidth?
    to:
    Gee, where are we going find a pipe big enough to send all of this data?

    I'm justed amazed that the data movers are coming up with new techniques as fast as the data producers can come up with stuff to send.

    Of course, I'm still waiting for advances like this to filter down to us plebians who just want symmetric 10 Mbit cable modem access. :)

  • Um... just 'cause you can *feed* a wire really fast, doesn't mean that the wire can handle it. For example, try putting your face under a hose at full blast and drinking. Good luck swallowing!! DSL has inherent problems with line length, conditioners (coils) on the line, splits, impedance, and a number of other problems. When I worked in the lab of a very very big DSL provider, we were lucky to get 7mb/s... 60 is ridiculous using copper at 12,000 feet, but good luck!

  • From the article:

    [corporate BS deleted] said it had created a router -- the equipment that directs traffic from one segment of a network to another -- capable of delivering 10 terabits, or trillion bits, of data per second, 10 times that of rival products.

    Now, it may not be apparent how Lucent looks at packets with mirrors in order to route them correctly, but if you take them at their literal word (a risky idea for a new product) then yes, they say twice in the release it is a router.

    The smoke is the most important part in determing the destination address when routing anyway.........

  • While the rates quoted are quite impressive, what I find more significant is the all-optical nature of this router. A router is not that far from a general-purpose computer. So what could this mean for an all-optical computer?

    Jack it directly into the phone trunk, and surf the datastream. While using the same computer to run the fastest single SETI-at-home on the planet.

    If Lucent can produce these things in serious volume, I predict we'll start seeing all-optical supercomputers living on terabit LANs with data warehouses nearby.


  • I would really like more information on the internal workings of this thing. I would like to know how this router will deal with input/output queueing and buffering. Currently, most routers have to buffer data while the logical connnections neccessary for routing to occur are negotiated within the routing table. I wonder how Lucent plans to buffer an optical signal without converting it to an electrical signal and storing it in RAM. I could forsee them employing a cut-through technique with virtual circuits between interfaces, but that would only really be practical in a switching environment rather than a routing environment. I what kind of an engine Lucent has come up with to switch packets without buffering them. Anyone know?

    Besides the impressive technology Lucent has managed to create here, think of the possibilities for the internet that this presents. The current backbones of the 'net dont have the capacity to deal with the huge amounts of data that will soon be flowing across them due to the high amount of Cable/DSL circuits that are being turned up all across the country, but if major Tier 1 ISPs start employing these Lucent routers and other all-fibre routers that I'm sure will follow soon, the backbone may just survive the next few years without imploding on itself. Cool.

    -Phizzy
    phizzy@psynet.net
  • It seems to be awsome technology, but, for now, it will not aleviate our usual complaints: latency and our local bandwidth.
    This will benefit the carriers, those companies like Qwest, etc. wich have massive fiber deployments and only the largests ISPs.
    From the relese, it mentions that it will route lambdas, or wavelengths, but not what interfaces it will have (oc3, oc12...oc192, fddi, etc) or wether it will do ip routing. It will probably only accept > oc48. The bottleneck will be moved more and more towards the edge, but the latency is not addressed (for a single ip packet); therefore,
    without QOS rules and such, high quality interactive video and voice for the normal user are still far away.

    cheers,
    Frank

    PS. did you guys hear about Nortel and Intel partnering on Open ip and IX ?
  • Disclosure: I work for Cisco, but believe me, I'm just the tiniest of schmucks in the corporate ladder over there. Translation: This is me talking, not them ;-)

    Lucent continually impresses me with their labs... breeding new technologies from within. Cisco, on the other hand, devotes very little to research.

    I wouldn't say this is flat out wrong, merely somewhat short-sighted. From what I've learned, Cisco's employee turnover rate is famously low--there probably is no other company in the high tech field that can claim such a successful acquisition department; indeed there was an article in Fortune about how skilled Cisco is in successfully absorbing new companies.

    This is rather critical, if you think about it. Low turnover means those individuals who have proven themselves productive in generating new technologies in a small company on the open market remain working with and for the company--thus, when Cisco purchases a highly skilled research organization, they actually get what makes the organization what it is--the people.

    You can't claim Cisco doesn't do any research when they've fully absorbed so many research heavy teams.

    That being said, there's a "baselining effect" that's at play here. Cisco spends immense amounts of money creating advances on R&D, and Lucent is not completely averse to buying companies(they recently purchased a major network services company, if I remember correctly). But you hear about Cisco buying companies, and you hear about Lucent inventing technologies, so that becomes the "baseline expectation" even though both organizations invest significantly in the other paradigm.

    Of course, it's probably fair to say Cisco does much more internal research than Lucent does external purchases. Lucent comes from Bell Labs, so you're looking at a corporate culture endowed with a pretty significant legacy of advancement--there's a very strong bias against "release early, release often"; rather they prefer to develop ad infinitum. Whether this is good or bad, I can't say :-)

    Yours Truly,

    Dan Kaminsky
  • Will any other hardware be able to keep up with this any time soon?

    It seems to me that if the new router is "10 times" faster than existing products and the data is split into "256 channels", then each channel is only delivering 3% of capacity we know we can handle now, and you should therefore be OK. You might need 30 times as many boxes, but they should be able to handle it.

    This is all-round good news. I wonder how much they will cost...

  • Allow me to clarify... (Score:3)

    by Brew Bird ( 59050 ) on Tuesday November 09, 1999 @06:02AM (#1549223)
    This is NOT a router for IP, it is a router for WAVES. Read DWDM please.
    The optical switch technology is used to switch diffrent lambdas through the switches.
    Primary use would be if you wanted to by a Wave on a carrier's fiber. (much cheaper than buying the fiber, because you don't have to light it up and maintain it yourself)
    So, the bottom line is, now I can buy a wave that crosses several pieces of fiber, and can be re-routed in case of a fiber cut!

  • Re:Wow. (Score:2)

    by Tet ( 2721 )
    us plebians who just want symmetric 10 Mbit cable modem access.

    That'd be nice, but right at this moment, I'd settle for BT sorting themselves out and giving me their pathetic 512Kb/s asymmetric line at their hugely inflated prices. They don't even seem to be able to get that right. Sigh. Until then, I'm stuck with ISDN, and even that is significantly better than the average UK household...

  • Depends. Is it high-order smoke, or low-order smoke? Also, is it aluminium-based, or copper?
  • Well, that's pretty cool.. but I'm alittle suspicious. I know that many switch manufacturers will create a 48 port do-it-all managed switch.. but the reality is the backplane of the switch can only handle 2-3gb/s of data.. which if you're running all your ports at 100, and under full load, will start dropping packets.

    I wonder what they mean by "terabyte" speeds. I would be more concerned about latency (How fast does it handle complex filtering rules, or NAT / load balancing?) and the maximum /effective/ bandwidth...



    --

  • Re:All I can Say... (Score:3)

    by Brew Bird ( 59050 ) on Tuesday November 09, 1999 @06:06AM (#1549227)
    This box doesn't actually route IP, it routes lambdas.
    Read up on DWDM for more details.
    There isn't a box out yet that can do OC-192 IP routing. Do the math on how long it takes to dig into an IP packet, make a forwarding decision, and execute it. It will be HARD, with current RAM read rates, to fill up an OC192.

    Of course, this doesn't prevent people from putting OC-12s and OC-48s on OC-192s and making press releases!

  • Look at http://www.switchcore.com they got the product that will rock your home network. Gigabit stuff
  • No, it IS a router, just not an IP router, which I believe the Press release would like you to believe.
    As the man says, it is a router for lambdas, and will be used by fiber carriers to wholesale 'dim-fiber' or 'waves' to non-facilites based carriers.

    In other words, this is NOTHING to get excited about, unless you own fiber.
  • Did you see the "in related news" bit, at the end? All those "oooh"s and "ahhh"s over PSInet buying some Lucent router, so that they could handle 10 gigabit pipes.

    Very impressive, maybe, if they'd made that announcement last week, or last month. But in the same day that Lucent starts shipping routers over 1000 times faster? It becomes less staggering and more "why did they buy the cheap stuff"?

  • This type of device is obviously a good thing(tm), however, I don't see it making a big difference in what we as the end users see for a long time.

    For one, just because you have a router that can switch packets at 10 Tbps, that doesn't mean you have the money to support the fiber lines you would need to even come close to those kinds of speed, much less pay for someone to route traffic over it to you. It is also likely that your local telco doesn't have the capacity at your CO just yet to support this either.

    Also, what happens when you start slapping on filters and access lists and start doing some layer 3 routing on top of this machine? Does performance degrade? Most likely. While being able to read packets at line rate is great, the real obstacle is to be able to build a full-fledged routing device that can route and filter at line rates. There is a huge difference.

    Don't get me wrong, this is a great breakthrough, but don't expect to see the end of the Slashdot effect just yet.

    -B
  • One thing to remember. Sure at first, that would be overkill, but as time goes on, people will always find a way to fill, and eventually overload the pipe, no matter how big it is.

  • According to the article, the new routers will no be generally available until December 2000, so they didn't really buy on "the same day that Lucent starts shipping" the new hardware.

    The point of the related news, presumably, was that Lucent seems to go from strength to strength. (Unlike Cisco, as a previous poster pointed out.)

  • Low turnover means those individuals who have proven themselves productive in generating new technologies in a small company on the open market remain working with and for the company--thus, when Cisco purchases a highly skilled research organization, they actually get what makes the organization what it is--the people.

    But is $7 billion a little steep price "for the people"?

    I think it is a fascinating environment. You basically have NorTel, Lucent and Cisco all trying to muscle in on each other's turf. Cisco being the youngest and most unique I would say... but there is most likely room enough for all three to be highly succesful in the future.

    You are correct that Lucent does buying as well, in fact they are trimming about 1,500 technician jobs from recent acquistitions (most are receiving full pension).

    I always like to look at things from an investment stand point. I see that Lucent trades at about 60 times their earnings (pretty expensive), while Cisco trades at 120 times earnings... that makes me nervous. What happens when we hit a dry spell and Cisco's cash reserves and market capitalization start to dwindle a bit. Can they prosper (I am sure they will survive) without the new acquisitions beefing up R & D?

    Aside, I do have friends that work for smaller companies... and essentially one of their main goals (as a company) is to be bought by Cisco.

    This is somewhat similar to my friends in Biotech companies that hope to be bought by the Pfizers, Merks and Eli Lilly's out there... kinda too bad they could not do well on their own (most fail).
  • This isn't a router, but its not a patch panel either... Since your doing layer2, it classifies as a Switch... not a router. Routers look at logical addressing via layer3, where as switchs, like this device look at the physical layer and switch out based on hardware addresses. And they say its all Optical... but I have a question... Since when did mirrors make OR and AND gates? If your going to route, you need to do some computations to figure out where to send the data... Layer2 all you need to do is look in a table and determine which port to send out on... much quicker when you ONLY care about directly attached devices and not devices not directly attached, which is required to do layer3 routing.
  • So would it make sense for Lucent to be rolling these out with a service agreement with a fiber optics company, say... Qwest?
  • In other words, this is NOTHING to get excited about, unless you own fiber.

    Sure it is worth getting excited about. I don't have fiber (in fact only a lowly 28,800 bps modem) but the internet backbone (as far as it exists) does, and freeing up the congestion there will help me get my data faster. (Especially when the UK wakes up and builds a decent comms infrastructure, but it is surprising how often my modem is running at less than full capacity when I'm downloading. Now if /. could have one of these connected to my ISP, and a faster server, I'd be very happy.)

  • Almost any communications medium would allow this. Just change the size of the library. It's a meaningless phrase!

    You missed the critical word: "...allow the printed contents of whole libraries to be transmitted in a single second."

    This thing will actually move the books!

  • THANK YOU for that post! 1)to run 10 Terabps you use DWDM- Dense Wave Division Multiplexing... so there are about 250 different colors of light in the fiber...from what I've read, the true benefit is that each wavelength is completely independent from the others so that when you rent certain wavelengths you can control QoS or optimize for throughput or latency or whatever you want without interfering with anyone else who is using the fiber. I'm not sure if it requires new fiber or if DWDM works well with current fiber. 2)It hits the market in 13 months (if that soon) so don't expect an instant impact. It is still significant, but remember the time value of Technology.
  • Yeah, I agree. I hate spanking hard earnt moderator (tu-wisted moderat-ah... Cooooo-me maaaaaark myyyy po--sts, slashdot surfing addict insane! - need to lay off the Prodigy, I think :) points on useless first posts.
    I've noticed recently with provocative issues (Roblimos' lovers guide, MS's legal troubles etc) that comments run to 375+ in number. Is there a constant calculated along the lines of

    Arbitrary constant = Total number of comments today / Total Moderator points available

    In these big discussions, I've found that the first 100 or so comments are moderated, the rest seem to be ignored. I've lost count of posts which I thought were worthy of a high score, only they have not been moderated up because they weren't posted in the first 100... So I think the proposal for M3 should include something like forcing nested or flat mode for moderators when browsing, and maybe prevent the average number of moderator points expended in any quartile (by timestamp) of comments exceed that of any other quartile by an arbitrary constant (50% maybe?).
    I'd also like to see two additional moderation classes - "Lame First Post" (you should get your moderation point back for this - but abuse of it should put you on a moderator black list) and "Offtopic But Interesting"...

    What do you guys think? How is the moderator system working for you?
  • Cool. Maybe libraries should use this instead of those little carts!
  • ...there are still a hell of alot of people using 33.6k modems, calling C range access numbers, praying for 3k/s.
  • But, it doesn't free up internet congestion. That is the point I was trying to make. All it does is make a fiber cut all that more of a disaster. (although it may be able to switch around one)
  • On thing to keep in mind is that your only as fast as your slowest piece of equipment?
  • 1) Again, that is not IP bandwidth, that is bandwidth overall. DWDM devices currently max out at about 16 lambdas per fiber. DWDM requires the 'good' fiber, as well as repeaters that are a lot closer together than most fiber carriers (like qwest and lvl3) have deployed.

    2) as for hitting the market in 13 months, please don't tell WCG that, they have it now...

    The bigger impact will be when we figure out how to
    a) parrallel route accross multiple OC48 trunks
    b) route IP at 10Gig :>

    Most people consider the answer to both of those questions to be some form of Tag Switching (MPLS or ATM)
  • Please read some of the other threads here, this is NOT an IP router....
    Don't confuse it with one.
  • Isn't this new router the same as the one developed by NexaBit (sp?) that we've been reading about for the last year or so? NexaBit was bought by Lucent a couple months ago for nearly a billion dollars. This may turn out to be a very good investment, but it is not cause to praise Lucent's ability to breed new technologies from within.

    Though I agree, Lucent is a very impressive company.

    Tim
  • WOW!!! with that, we can do some REALLY serious VR Mabey in a year or two, HTML browsers will become obsolete, making way for new virtual interfaces
  • "Makes"..."Created"...
    You'd think it's alive!

    calm down yer affection fer the router!

  • I think I see your point: it is not enough to be able to route tons of data, you also actually have to lay extra fiber to give you increased volumes. The limit on the Internet at the moment is not the capacity of the routers, but the total banwidth (bps, latency, etc. all in one). Right?

    Is this actually true? Are the routers on the internet limiting our access times, or is it only the "thin" wires? Any comments/thoughts or, even better, measurements, experiences, facts?

  • The limit is the routers, not the amount of available bandwith.
  • Three cheers for bandwidth! Yahoo(r)! Unfortunately, I think we'll see a lot of inconvenience with this new speed. At the same time we get our /. faster, we will see in other places the further proliferation of certain large companies' bandwidth hogging banner ads, dynamic whiz-bang streaming(r) Who-knows-what, and the hopeless drowning out of real(r) content by various other forms of internet garbage.

    Have you ever viewed the source on a page to see what the ratio was between actual human usable, relevant information, and something like tripleFlip.org's "click here to win a free inflatable tire" ad?

    Just me blowing steam. Some things can't be helped.

  • Thanks for that post! It is notes like that that makes /. worth the time and effort.

    For those of us that are a bit hazy on the finer points of Dense Wavelength Division Multiplexing (DWDM), there is a very decent tutorial by Lucent at http://www.webproforum.com/dwdm/topic0 1.html [webproforum.com].

    More list of resources can be found at http://www.atmdigest.com/WDMResources.htm [atmdigest.com] (DWDM specific), http://disa11.disa.atd.net/tutoria l/tutorial.htm [atd.net] (general ATM), and at your favourite search engines.

  • I actually just interviewed with Lucent. Allow me to explain how this fits with their product line:

    As mentioned several times previously, the main box featured in this article is a DWDM router, not an IP router. There is a big difference, because they do very different things. See previous posts for details.

    Lucent just recently finished their purchase of Nexabit, a private datacomm company. Nexabit currently is shipping the NX64000, a true 6.4 terabit IP router mentioned at the bottom of the article.

    Lucent is the only company I know of capable of routing OC-48 IP at line speed. They have imminent plans to route OC-192 IP at line speed, and I expect them to release it soon. Some can route gigabit ethernet, but routing OC-48 and up is where the men are separated from the boys.

    Lucent has one other competitor in the terabit routing space, and that is Juniper. Juniper is the only other company actually shipping a product. Avici, Zuma, NEO, Ironbridge, and Pluris all claim to have products, but aren't actually shipping anything capable of routing more than 1 terabit across their backplane/switch fabric.

    In any case, Lucent's box has the most single-chassis bandwidth of any competitor. Several of the aforementioned companies are trying to play PR games by stringing multiple boxes together, but you definitely can't achieve terabit speeds by daisy-chaining boxes.

    All of this information is true to the best of my knowledge, as I have personally talked with all companies I mentioned above. If I were an ISP, the box of choice would be a no-brainer.

  • Most people in the Internet world use router to mean something that switches packets at layer 3.

    Almost everything else is called a switch, so I think this should be called a DWDM switch, since it is working at layer 2 with DWDM (or maybe layer 1 depending on how you look at it). (DWDM = Dense Wavelength Division Multiplexing, or shoving several colours of light down a single fibre.)

    This sort of kit will be very useful though - before too long it should be possible to run DWDM to the customer premises, but even before that heady prospect DWDM should make it possible for xDSL and cable modems to run closer to full speed.

    All we need now is faster servers and 10-gigabit ethernet (which a standards group is working on I think).
  • This is an optical switch - it's best IMO to think of it as a layer 2 switch connecting a number of different routers with very fast 'logical wires' - each link will use a number of wavelengths, and the switch presumably converts stuff from one wavelength to another (though if it's all mirrors maybe they have to stay on the same wavelength).

    There is some other work using tunable laser transmitters and corresponding receivers, which lets you set up a LAN-like structure on DWDM - this seems more like a L2 switch, analogous to the way ATM virtual channels can be used to connect IP routers.

    The world will still need very fast IP routers, though - if you have too many IP routers connected via a DWDM cloud, you run into problems with the routing process in the router 'seeing' too many neighbours and thereby working inefficiently (particularly if an interface somewhere starts going up and down frequently - known as route flapping).

    The real issue for IP usage is how on earth do you build routers fast enough to keep up with all the wavelengths pouring out of a single fibre? Terabit routers will address this through massive parallelism, but they don't come cheap.

  • I would expect that this will have little effect on most end users. Except for one place, your local movie theater. I would expect these types of devices are what would be needed to deliver the kind of bandwidth needed to distribute completely digital movies, like the special showings of Phantom Menace this summer, to the thousands of theaters across the country.

    The only real problem with this scheme is how do you feed a pipe like this. ;)
  • *G* Oh, but they can..... :) :)

    *EVIL GRIN*

    ECN - Throttles the bandwidth of applications which are hogging the net. If that fails, it also throttles the computer until dead.

    RSVP - Allocation of bandwidth to -your- resources. You are absolutely guaranteed that amount. So, if you RSVP bandwidth between you and Slashdot, nobody else can grab that from you, by sending huge banners or real-time video.

    CBQ - Limits the bandwidth allocated to a given type of data stream, dropping excess packets as needed.

    And, my favourite:

    Return-To-Sender


  • How does this compare with the latest (or annonced version) of similar product by Cisco, Nortel, ...?

    I think Press Release often look like fantasy essay.
  • ...OK, so we run a 10 terabit backbone from NYC to SF, for instance. Then everyone starts selling their 100 gigabit routers super cheap, and every ISP on the planet buys two or three, and drops redundant links. The effect? Same as we have now. Except instead of trying to push a terabit down a gigabit pipe, we're pushing petabits down a terabit pipe. Same effect, just larger orders of magnitude. And remember, the number of people on the internet is doubling like twice a week now. I know, that's not accurate, but damn, sometimes it feels that way.

    60Mbit DSL to the doorstep isn't viable. Neither is 100FX to my bedroom from my ISP. In reality, even a T1 to every house really isnt viable. Yeah, I'd be a helluva lot happier with an OC-12 to my desk so I could download StarOffice in 14 seconds flat, but I'm happy with a quarter-meg DSL or a 128K ISDN. It's not too slow, I can download a linux kernel in 15 minutes as opposed to an hour. In fact, I'm all for discouraging faster internet access. Yes, you and I will be limited to the "paltry" 128K... but the backbone won't be crushed.

    If you have, for instance, one million DSL users running at 768K, that's 768 million kilobits of bandwidth. 768 gigabits. And we all know there's prolly many more than a million now.

    And then there's the college students. They do porn, MP3's, and warez like it's a requirement for graduation. For example, my uni, Drexel University, has a DS3. UPenn has 2. Temple has a 5 meg FDS3. TJU has a 16 meg FDS3. That's just the major colleges in -ONE- -CITY-, eating up 200 megabits. Then there's PSU, Pitt, and the other 49 states, etc etc....

    Then we have commercial entities like connexion.com, Microsoft, Sun, IBM, Oracle, and our favorite news for nerds site /., eating bandwidth like it's going out of style.

    10 terabit is good. But we need to stop letting technologies like this creep out towards the edge, because it clogs the core. Ah hell, I'm just ranting. But my UserFriendly was slow this morning, and now I know why. Some smartassed kid who came into some money bought himself a 10tbit line and started mirroring every porn site on earth.

    --jd
  • Nah. Give it 20 years, we'll all have them on our desktops.
  • Look bub, having Cisco certification doesn't mean you work for Cisco.

    No, sitting at my desk in Building C, 2nd Floor, in San Jose means I work for Cisco.

    You're probably some stupid contractor.

    No red badge on me. Anyway, I'd be happy to be a contractor; many of the smartest people in the industry contract.

    I've read some of your posts and to be quite honest, you're very stupid.

    Why thank you. If I've offended the likes of you, I'm doing something right.

    If you are indicative of the quality of people at Cisco, I'm calling my financial advisor and instructing him to dump my Cisco stock.

    I think that speaks more of your intelligence than of mine.

    Yours Truly,

    Someone who sees right through you

    Ciao, Jizmak.
  • This puppy will do oc192.
  • Lucent is the only company I know of capable of routing OC-48 IP at line speed. They have imminent plans to route OC-192 IP at line speed, and I expect them to release it soon. Some can route gigabit ethernet, but routing OC-48 and up is where the men are separated from the boys.
    Cisco's performance OC-48 card (not the old time-to-market card) for the GSR's does line rate ip quite beautifully.
    Lucent has one other competitor in the terabit routing space, and that is Juniper.
    Juniper's box is *not* a terabit router. It has 40 gbps of bandwidth. There is no such thing as a terabit router yet. Nexabit's OC192 is not real OC192. It's basically Nexabit's OC48 engine (which sucks) combined with the Lucent mux/optics package. Nothing special, and not too surprising. Just lets you run the 4*OC48 straight into the router rather than using an external optics package. The wait for true OC192c continues.
    Lucent just recently finished their purchase of Nexabit, a private datacomm company. Nexabit currently is shipping the NX64000, a true 6.4 terabit IP router mentioned at the bottom of the article.
    I don't know where you got the 6.4Tb number comes from. Let's see...we have 16 slots per box. Put an OC192 in each slot. That's 10 gigs per slot. Hrmm... that's 160 gigs. Even if you could put *10* OC192s in a slot (which you can't), that's only 1.6Tb/s. Yeah, it's a no-brainer all right. Just give me a Cisco 12012. At least they're honest and call their box a gigabit router.
  • I made one of these a little while ago, by accident. Spilling marine diesel into the ocean provides the same technology.
  • well we all can't be in the US of A.
    try to get something better than that in australia without
    spending some serious money.
  • Well, in the world of optical communications, the word wavelength router has been around for years, if not for decades, to mean a device that can direct light with different wavelength from a single fiber to designated spots. In that sense, it is a router. If one confines the definition of router to mean "data packet router" in electronic sense, maybe your argument stands... (I do agree that it is not an optical computing component!!) Obviously, sending information down an optical fiber by lining up information from 100 different people in series, as done today, is not the smartest way to do things. If you use 100 different colors instead, the information can be routed all-optically, without having to read the header (storing the data in a buffer) and deciding from that where it should go. Once you go to an all-optical system, the routing can be much faster than the present "data-packet" approach. Well, all-optical communication will be much different from the communication we know as of today, and it looks like there is a lot you have to learn!!! Welcome to the new world of optical communications!!!

Slashdot Top Deals

Pause for storage relocation.

Close