Making Last-Mile Ethernet A Reality 113
vannevar writes: "Is that erbium-doped fiber you're smoking, or are those bandwidth crack-heads in the Ethernet First Mile Study Group turning up GigE fiber to the garage? Of course, no good deed or innovation goes unpunished, but at least someone is busting knuckles, carpal tunnels, wallets, and reputations to make Gigabit Ethernet To The Home a reality." You may remember this earlier mention of the same concept, but rather than just ideas and proposals, here are pretty pictures and delivery speeds that might even make non-Californians want to relocate.
Re:In some places this has been going on for (Score:1)
Already being done in Cedar Rapids, Iowa (Score:2)
Re:Missed Naming Opportunity (Score:2)
Down that path lies madness. On the other hand, the road to hell is paved with melting snowballs.
Re:Yeah...ok...but where -- in the UK, now. (Score:2)
Re:This is nice and good, but... (Score:1)
About the only company trying to get broadband to everyone is Earthlink. Now that they offer DSL, Cable, Sattelite, and Wireless, you have to figure they can offer high speed internet to just about anyone.
Dastardly
Re:cool, but some questions... (Score:1)
You've obviously never heard of a fiber splice before, have you? Have you ever heard of a fusion splicer? You can actually make a pretty good mechanical splice, too. Yes, there's a dB loss, but there always is. It's not that big a deal.
More info: (Score:2)
Yum
I'm not sure if I understand (Score:4)
I can hardly blame them for being self-congratulatory in tone - they deserve it, surely - but some explanation of how they did it seems to be in order for those of us who would just love to do likewise.
As others have quite rightly said, the fact that their server survives a slashdotting is pretty impressive. I see they even have video! Now is that cheeky or what, even if the Linux system I have at work can't handle it
So tell me, how was this done? What's the history? Something like JWZ's DNA Lounge chronicle would seem to be in order, and I couldn't find it. Can some kind soul point me to that?
Thanks.
D
(who lives in Los Angeles and is stuck with iDSL
----
Re:Interesting thought (Score:2)
This can be done anywhere (Score:2)
Why doesn't this happen today for ethernet? Basic 100BaseTX can support a neighborhood easily. but herein lies the problem... When Mr Lawyer down the street get his computer Hacked he instantly sues everyone within view of his home because he was a moron and didn't have a firewall or properly secure his box. Voila, the neighborhood net is now dead because of one a-hole lawyer or other type of idiot. Solution? you have to be a corperation to do this.. Now you have to get neighbors to give you land to run your cables,install tech boxes (basically waterproof boxes with a switch inside) etc... now you are a company, the city wants a piece of your action, and regulate you..... to death..... why? because those morons you voted for dont know jack about computing or networking... so they want to call you either a phone company or cable tv company (you are nither but they could care less... they want moola from you)
Basically... you can do it, only if you overthrow your local government or own a huge plot of land and subdivide it having this infrastructure in before you sell the plots (then they cant do a damned thing to "regulate" you.)
Hard part - fine me one cable company that sells direct burial CAT-6E... it don't exist.
sorry if I am rambling, but they didnt do anything revolutionary, they just copied what many did in the 60's.
Fiber is silly when DSL and cable are capped (Score:1)
So what would fiber buy you? $$$$ price. 10x the theoretical bandwidth. And no change in actual performance.
Had this for 1.5 year (Score:2)
Real Estate (Score:3)
Wake me up when they do this in East Palo Alto.
Content providers (Score:4)
s as it is, with hits coming in at innundating rates, imagine what they'll have to do when the limiting factor on all data transactions becomes the bandwidth of their hard drives and memory in the servers? I mean, "last mile ethernet" may sound great, but who's going to upgrade the backbones to multiple OC-4098 circuits to handle the new traffic? And what do the providers do once they've spent their entire start up capital on their own gigabit connection (you know, something with more QoS than Ethernet and therefore a higher pricetag) just so people don't bitch about their service being slow?
It's a double-edged sword. Rapidly increasing the bandwidth at the fringes of the Internet instead of the core is going to cause some serious problems and side-effects.
-Chris
...More Powerful than Otto Preminger...
Re:Real Estate (Score:1)
There is still some "affordable" housing in the area (mobile home parks, etc.) but don't expect it to last long.
Funny (Score:1)
Just a thought
Why. (Score:2)
I found out why they think everyone needs gigabit ethernet to their garage, take a look at the image sizes on that homepage. If they just took the time to find out about this new format jpg that everyone's been talking about, I think they'd be just fine with dialup.
I already have this (Ashland, OR) (Score:1)
They also run digital cable-tv over it too.
Re:Content providers (Score:2)
Second, if everyone has a fast connection, that does not necessarily lead to a server getting hit more often. What it will mean is that a person can get in and out faster. This leads to shorter queues. It is possible that a faster connection can lead to a more efficient web farm, since there will be fewer request sitting in an output queue waiting on an ACK so that it can send more packets.
Of course, in the real world, a faster connection means that I can and will look at more things in a set time period. But if I served 1000 customers/hr all using modems, and then everyone switched to OC-48 overnight, I most likely will still only server 1000 customers/hr.
I don't care about this... (Score:2)
Worldcom [worldcom.com] - Generation Duh!
Re:Interesting thought (Score:2)
Oh, you mean they were clothed. Never mind.
cheers,
mike
Skeptical... (Score:5)
This guy does a naptser download to compare the relative speeds of DSL, Cable and GigE to the house. While I agree with the basic conclusions (that symetric is going to be better than asymetric, and that GigE will be faster), some of the things he says stretches credibilty, and for obvious reasons.
It's just *NOT* a good test to use Napster as a mechanism for determing the relative speed of a first mile infrastructure. Or for that matter, any internet connected service. There are WAY too many variables in between me and the end site that I'm connected to on the Internet to be able to say that the underlying first mile infras is the problem. In particular the remote site may have an over subscription problem. Or the available internet bandwidth (beyond the first mile) may not be sufficient. NONE of these type of problems indicate anything about the capabilities of the first mile infrastructure.
If you want good tests for the first mile, stick a server on the other end of the first mile and do bandwidth tests to that. Otherwise, it's just useless hype, and it doesn't really tell you anything. The conclusion that a DSL or Cable modem really doesn't offer any speed advantages over a regular modem is just plain wrong.
That page, with its gross inaccuracies, would make me skeptical, as a customer as to whether or not anything provided by this organization would be reliable.
$.02
--
This is nice and good, but... (Score:1)
Ah well, I'm stuck at work and bitter, so I'm just rambling...
Re:This is nice and good, but... (Score:1)
Re:In some places this has been going on for (Score:1)
I am linked with a 10Mb ethernet since last august.....
is this news?
bindo
Priceless (Score:1)
Terabyte disk array: $20,000
Largest collection of live goat porn the world has ever seen: Priceless.
Re:Palo Alto broadband proposal (Score:2)
It probably got tossed in the bit bucket the same time half the companies in Palo Alto lost their VC and went belly up [google.com]. Silicon Valley's tax base is a tad smaller than it was 2 years ago.
Hey great... (Score:5)
Now if only I had some fricken power to run my computer...
Re:Content providers (Score:1)
You hit the nail on the head in part here. You can keep adding and adding bandwidth, but your still going to have some of the same old problems you always did because of IP. IP packets get sent along and bounce through routers in a way known as "best effort". Meaning that some packets may get there before others, or may not. What is really needed is mass implementation of some of the new Qos protocols like MPLS. Then we can actaully take advantage of the bandwidth.
Re:Do we need it? (Score:1)
Re:Do we need it? (Score:1)
Absolutely. I couldn't agree more. The solution is needed, but I just don't see the widespread need anytime soon (next year or two), and maybe that's the kind of time they'd need to get it ready for "prime time".
There's a big problem, because all the backbone providers spent tons of money building up infrastructure for bandwidth that, in the end, couldn't be supplied to enough home users at a high-enough speed. We've got all these fat backbones waiting for either more users or faster connections to the home. Unforunately for many of these companies, it was a matter of bad timing and cost them dearly, financially.
Do we need it? (Score:3)
I'm no knocking it and honestly, I'd probably get it 'cause I'm a geek, but do I really need it? It looks like a technology waiting for a purpose.
# of people on Internet not exploding (Score:2)
All of which points to the conclusion that ethernet for your average user isn't going to hammer the infrastructure too badly. Some upgrades will obviously be required, but demand isn't going to just explode to eat up the new supply.
____________________
Re:Relocation (Score:1)
Just do what my cable company did... (Score:1)
Re:Just do what my cable company did... (Score:1)
Also, with the *shared bandwidth* it is possible to have VLANs on different channels, given enough bandwidth.
In any event, I have certainly not experienced this much bandwidth from any other cable or DSL service. If you can explain how my friends can get their cable co to allocate this kind of speed for them on a single pipe then I am sure they would appreciate it.
Re:Just do what my cable company did... (Score:1)
Before you try to insult someone, please learn how to spell. It makes you sound less ignorant.
Re:Do we need it? (Score:1)
Re:Useless without power (Score:1)
Re:Metro Network Services could be great. (Score:1)
Sigh. Our laws will never let this happen, will they. Liability for copyright infringement, city guvmint wanting a piece of the pie.
Is it possible a quiet co-op could build such a thing without attracting the attention of the whole mad legal world? Why not maintain our own file servers? Probably not possible. Too many possible lawsuits, for just about everything. Not to mention angry ISPs wanting those interlopers made an example.
Re:Do we need it? (Score:2)
I only mention this because of the obvious analogy to how the high-speed infrastructure is being built-up in the U.S. Slow, expensive, and every mile must be justified by shown profit. This is going to take forever.
Re:Palo Alto broadband proposal (Score:3)
Re:Skeptical... (Score:2)
Me too. For one, the use of Napster as a benchmark is laughable, as the referenced post mentioned.
As many slashdot users probably know, another problem with comparing Cable with DSL is that bandwidth is almost completely dependent on the provider. Granted, for most heavy-bandwidth and low-latency applications symmetric service is preferable (straight out web surfing and Napster leeching aside). So certainly look for a provider that provides symmetric service (I have found that these also tend to be more oriented towards people who like to *use* their bandwidth, none of the classic Verizon DSL uptime of about 15 days/month).
In my area, cable modems are just plain *faster* than DSL. None of this balogna about "if you hav a lot of people accessing in your neighborhood...". This is because we have AT&T Broadband for a cable provider and Verizon for a DSL provider. With my cable connection I get downloads comparable to a T1 and uploads of about half that (give or take). However, the uploads are fast enough that I do not have to worry about having an asymmetric connection, because 1/2 of An Awful Lot is still A Lot.
So okay, I believe that these folks got the speeds they posted, and obviously having Gigabit Ethernet is superior, but Your Mileage May Vary to the point where you get oppposite results.
Re:Relocation (Score:1)
'Guess we'll have to move Lady Liberty to the Other Coast.
Re:Interesting thought (Score:2)
Now, if they'd just drop the price on that 51" flat screen monitor...
Re:Missed Naming Opportunity (Score:1)
In light of the new SOAP protocol, I think calling it "Network of Optical Passive Ethernet" would have been even better. Then we could have fun saying that we can do SOAP over NOPE (via ROPE [soaprpc.com], too!).
Re:Content providers (Score:2)
Sounds like a good justification for P2P.
Re:better network diagram available? (Score:2)
Re:Metro Network Services could be great. (Score:1)
But, if you are downloading a linux ISO image at 1Gbps, and 10 of your neighbors are doing the same, the upstream pipe needs to be > 11Gbps. That's why I said 'the aggregate bandwidth'.
Even if the provider is running an OC-192 HUGE pipe to the Internet, eventually it will get saturated, and maybe you can only download the CD image at a paultry 100Mbps or so.
With the local environment more able to scale to providing everyone bandwidth approaching 1Gbps, there is a big opportunity metropolitan area applications and services.
Metro Network Services could be great. (Score:2)
Re:streaming HDTV file size (Score:2)
That sounds like a lot, but the proof is in the output.. 1080i HDTV looks incredible. It blows away DVD's.
Check out the Digital TV "crash course" on the PBS web site for a lot better info than I can provide: http://www.pbs.org/opb/crashcourse/ [pbs.org]
Missed Naming Opportunity (Score:1)
With the letters E, P, O and N they could have at least been OPEN (Optical Passive Ethernet Network), or more sinisterly PEON. This may have highlighted that an ethernet network is kind of a tautology.
Of course this minor whinging is down to the fact that my Locale's c-time (time for Californian Innovations to be on offer) is about 67 months. And that is only until BT screw you
Re:Missed Naming Opportunity (Score:1)
I knew I was on dangerous ground when I was thinking of kind of a tautology. And I take my hat off to you for that fine word which I admit to having to look up. Here is the definition from m-w.com, which is bordering very closely as an example of the word it defines:
the use of more words than those necessary to denote mere sense (as in the man he said)
A phrase I will have to remember next time I'm in an all day meeting with the PHBs. Although I think the meaning is more to do with redundancy than self contradiction and meaningless filler
Re:Content providers (Score:1)
Re:Interesting thought (Score:2)
I have to ssh to my company (from home), then hop to 2 other nodes (all with ssh), until finally I end up at my desk linux box. I then start netscrape (or even vmware...) and in a few minutes the X11 connection appears and I'm able to do those visual things that can't be done with emacs, tin and elm.
of course bandwidth and network latency are not always hand in hand. I'd usually trade some extra b/w for shorter latency; especially when doing interactive things (eg, keystrokes) over the net.
--
better network diagram available? (Score:2)
Doesn't really cut it. I can't make heads or tales out of that, especially when the text on the labels is 1 pixel tall. I would really like to see a good diagram or read some detailed specs.
upstream connections? (Score:2)
streaming HDTV file size (Score:2)
HDTV (1080i) = 1920*1080*16bitcolor*30fps = 9.4x10^8 bits/sec = 940 Mbit/sec uncompressed for the video alone. Ignoring audio, that's going to require 47:1 compression. Even DiVX won't be able to compress that well and still make the image look good, at least not on an HDTV. Still need to figure in the thruput required to transport the dolby digital or dts sound as well, I don't know how well they compress.
Sure, it's a tradeoff, but if I bought an HDTV or high-res projector for my PC, I'm not going to want to stream crap video through it, regardless of the cost.
high dollar servers (Score:2)
http://www.fiberhood.net/first/images/DSC00111.JP
18 years ago... (Score:2)
I wish modern computers were even 1/4 as reliable and stable as that old Apple IIe. If there was an easy way to open Word and Excel files as well as interface the thing to my color inkjet printer and use different fonts, I would probably continue to use it. Over the 7 years I heavily used that machine, the only problem I ever encoundered was the lousey spacebar, I had to reseat the little metal support about once a year.
A heardy thank you to the Woz for creating such a delightful series of machines.
Re:20Mbps? 19.2Mbps (Score:2)
I was lucky enough to spend some time working with digital video at SGI (Silicon Graphics) earlier this year. Quite an education. One thing that I had overlooked was that compressed video has to come from somewhere... and SGI had that somewhere on Octane2 workstations. I belive 1080i was between 124 MByte/sec and 248 MByte/sec depending on the bitdepth (2, 3, or 4 bytes per pixel -- the film industry loves 4 byte / 48 bit color). To even play back uncompressed 1080i video required three channels of fibrechannel disk arrays attached to the workstation. Overkill, perhaps, but the machine handled the huge uncompressed video like a modern PC can handle a 320x240@15fps AVI.
MPEG2 compresses digital video very well and depending on the settings used, 1080 can be done with not much more thruput (bandwidth) than 720 or even 480. But belive me, an uncompressed digital version of a film digitized to 1080p/24Hz is a sight to be seen, especially on a Sony HD studio monitor. Without compression, "HDTV" really shines but the ungodly amounts of disk space required don't make it worthwhile with current consumer technology. Uncompressed video is beautiful, compressed video (when done right) is still *very* good.
I don't know about the thruput of streaming digital video, but based on my limited knowledge of MPEG2 and the various flavors of HD digital video would lead me to belive that 1080i at 20Mbit/sec would certainly be obtainable. I'm not sure about the audio, though, especially if it's to be played on a high-quality 5 or 6 channel system.
As far as the term "HDTV", you won't hear it used much outside of marketing circles. It's sort of like saying "computer" rather than "1.4 GHz Athlon running Linux" or "833 MHz Alpha running Tru64". HDTV is a vague consumer term.
The folks in that Palo Alto neighborhood had better upgrade their TVs/monitors/projectors! Very cool stuff indeed.
Re:Real Estate (Score:1)
Palo Alto broadband proposal (Score:2)
Re:Real Estate (Score:2)
Re:In some places this has been going on for (Score:1)
In some places this has been going on for (Score:3)
Re:Missed Naming Opportunity (Score:1)
Re:Just do what my cable company did... (Score:2)
Re:In some places this has been going on for (Score:1)
Enigma
Re:Interesting thought (Score:1)
http://us.imdb.com/Title?0065076 [imdb.com]
Re:In some places this has been going on for (Score:1)
I checked out airswitch.com and they have changed their name to switchpoint, and there is absolutely no information on the service or for ordering. They seem to have transitioned into a "technology" company, only interested in licensing their system to someone who actually feels like implementing it. Lame. I wonder if their (former?)subscribers in Springville UT are still connected?
Lame. Lame. Lame.
Re:In some places this has been going on for (Score:3)
It was heaven, with fries, biggie sized.
NOPE (Score:2)
What good is gigabit ethernet... (Score:1)
Interesting thought (Score:4)
I suppose we will all eventually have these kind of connections, but by the time it reaches my little community, it will be slow relative to the applications out at the time.
Relocation (Score:3)
Re:Do we need it? (Score:1)
Big pipes make things possible. Really neat things. Things like live worldwide videoconferencing.
Things which aren't possible now because too many people you want to connect to are on 56K or less.
Admittedly, getting people DSL would solve much of this problem. But really: in many ways ethernet cable is easier to lay in than to retrofit old-style telco/cable networks with tcp/ip capabilities.
I think it would be fine to just give everyone a 10/100 drop in the wall. That'd be great. But even that would be way more than DSL bandwidth.
Re:Hey great... (Score:2)
Yeah, maybe you'll make Jimmy Carter smile.
might even make non-Californians want to relocate (Score:2)
Whoops, someone left the gate open again. Sorry, we're all full up, try another state.
Seriously, we've had high speed, we've had bandwidth, we've had promises, promises, promises. Problem is, you make the technology available (even something as humble as DSL (nowadays)) and you still have trouble finding any company willing to spend the billions it takes to dig up all those out-lawns or climb all those utility poles to run the stuff around. With the market the way it is right now, it's all just a dream.
Funny how most high speed is still being carried over this awful copper which was laid in the 50's and 60's. Probably has something to do with JDS Uniphase, Nortel and Lucent all biting the bullet...
-- .sig are belong to us!
All your
It's less of a stop-gap than DSL. Cool... (Score:3)
--CTH
---
Well (Score:2)
Re:Do we need it? (Score:1)
Let's go back a little more than a hundred years. Did we need cars when we had trains? Did we need planes when we had cars later on?
Life, as we know it, is just like puters. They all basically do the same thing. The only thing that we pay more for is speed and storage.
There will allways be a purpose for it. We may not know what it is at this time. But if you were to ask someone 20 years ago if there was a need for a fax machine that would have said, nope, mail works fine for me.
That is my
dave
Re:Already being done in Cedar Rapids, Iowa (Score:1)
Re:Content providers (Score:1)
MPLS has some great features for VPN's though that are nice. But with MPLS outside of your core, you are hosed. The next router has to be able to handle the MPLS tag.
An interesting side note would be if, the net was built with switches instead of routers, we would have no need for mpls! Layer 3 is a great place to handle packets. The problem that the net has is that there is no true connect. You have ATM cells to worry about, you have Frame PVC's to worry about. If everything was Ethernet instead of Telco forced standards then the net would scream!
Ok, I am off my soapbox.
dave
Re:Content providers (Score:1)
The reality, in my opinion, is that the Cisco/Lucent/Alcatel/Nortels of the world are hosed. They have based all of their equipment to handle the Telcos. Every piece of their equipment is built to do Ethernet to Sonet conversion. That costs money. The crap that you hear from them is that without Sonet you have no protection. Well, I guess they don't know squat about ethernet. Ethernet was built with collission in mind. If we drop a packet it will resend it! With Sonet, you, in theory never have to worry about that. If they built their equipment to do just pure ethernet then they would be fine. Well, guess what, they do and their called switches! To put pricing in perspective: Cisco has just priced out their OC-192 blade for their GSR line of routers. The routers start @ 250k, so these are the big boys. The blade is 250k in itself. OC-192 in essence is 10 gigabit. Hmm, last time I checked, if I was a ultra cheap bastard, I could get gigabit ports (fiber) for under $300. So give me ten and let them aggregate! There are also more issues in doing Sonet conversion to Ethernet. OC 48 is 2.5 gigabit. OC 12 is 622. So if they want a gig, they have to burn an OC 48. So they effectively lose about 60% of the utilization of the pipe!
The computer people of the world need to force the telcos of the world to work in something other than a 64k, ds0 channel. Ethernet has allways worked in a base 10. Why? Because we like easy math
dave
PS. This was typed as I started to fall asleep, so don't mod me down because of typos or lack of sleep
Re:This can be done anywhere (Score:1)
Re:This can be done anywhere - more links! (Score:1)
Re:Ethernet to the home (Score:1)
Re:Interesting thought (Score:1)
I suppose they shouldn't build it...
Changes very little (Score:1)
cool, but some questions... (Score:3)
How about stringing that very delicate fiber over long distances? What about when there are breaks? Fiber is almost impossible to put back together from two ends, as it has to be 1/4 wavelength+ or scattering will result in an increased error ratio.
I also don't know why they're dissing cable. Cable is awesome, at times much better than DSL. I don't know why he thinks a cable modem only gets around 2.5 kb/s in real life performance. I have a cable modem and can get up to 2.5 Mbps download and over 500 kbps upload. You should also note cable modems ARE capable of high upstream bandwidth. My modem, right now for example, has a maximum bitrate of over 2 MB/s, with a power level of 50 db. However, this is limited at the provider and through the modem via QoS.
The Napster test was equally stupid; everyone knows 28.8 users select 'cable' for whatever reason. These programs should really report the average real bandwidth instead of allowing user selections, which are for the most part pointless.
Also: for the person who was talking about the general slowness of the net and the fact that it won't matter how fast a connection you have -- you'll still only get a max of 500 kbps at even a very good site: I've got news. If, for example, we all had Gigabit connections the net would run MUCH faster. My neighboor also has a cable modem, and if I connect directly to his modem to send files, etc. I can get up to 2 MBps transfer speed. This is partly due to the fact that in modern cable setups, more and more routing is done on the neighborhood level ('micro-routing' and many slow routers -- MSR). If everyone gets a high speed connection like this, the Internet will run a lot faster. And as soon as providers realize more small routers are better than a huge few, things should improve.
Problerm (Score:2)
But nonetheless, I want access!! ;)
________________________________________________
640k (Score:2)
Sometimes, the need isn't there until the product/service allows for it. Downloading DVDs may seem like a pipe dream today, but it sure would save those long trips to the rental store! :)
Gold, shmold (Score:2)
Re:Interesting thought (Score:2)
Setting up a totally kick-ass Freenet node, one that services your entire time zone.
This is old news.. (Score:2)
There isn't much information available in english yet, though.
Analogies (Score:2)
It's never going to happen (Score:2)
Re:Content providers (Score:2)
Info asked for, info provided (Score:2)
http://www.alloptic.com/
Here are some tech docs.
http://www.alloptic.com/techcorner.htm
Redq6