Ethernet Sets To Bridge The Last Mile

sacremon writes: "An article in EETimes reports on a recent meeting to finally bring Ethernet to the home user directly, rather than using broadband technology like DSL or Cable. At this point, they're only in the planning stages, and they don't expect to see implementation till sometime in 2003. Nonetheless, I would love to have a 100Mbps/full duplex line direct to the house. I can see the self help manual now -- 'OSPF and BGP for Dummies.'" Ethernet could bring good rates (for both data and dollars, if this article is correct), but I'm still looking forward to fiber running straight into the basement.

Comments

basement?

[grammar nazi]

I would rather have fiber running into my computer (switch, router), not into my basement!

Re:basement?

[Muad'Dave]

Why? For the purposes you want ot use it for, copper cables are just as adequate. Where you really win out with fiber is either long-distance, or extreme-bandwidth per strand, neither of which you will be using in your house.

Two words: "Copper Conducts." or "Lightning Strike." or "A Car-hits-pole-and-powerline-crosses-data-lines."

I would feel much better with a nice optoisolated fiber line into my house. I cringe every time we have a thunderstorm 'cause my ISDN router is hooked to a 2 mile long antenna asking for inductive (or worse, conductive) surges.

Re:basement?

[mindstrm]

Okay. I'll buy that. But opto-isolating copper is also easy.

I'm not saying that fiber doesn't have advantages.. just that, for today's purposes of the home user, it's not 'leaps and bounds' above other methods of broadband digital. THe media itself has great potential.. but will not be beneficial now.

Also, telecom lines have lightning arrestors all over.

It's been done

[Kalak]

This has already been done in Blacksburg, VA for the Blacksburg Electronic Village. As a result of this project which started a few years ago, a number of apartment complexes have gone to adding ethernet to units. $30/month for ethernet rocks! (Too bad I had to move out, but iit was great while it lasted for me.)

Re:It's been done

[rearden]

Not only has 100Mbs ethernet been done but two appartment complexes in Blacksburg do have fiber to the apartments. While the idea seemed good at the time the cost for the resedents has proved to be too high (cost of hubs/ switches/ NICs) so we quit doing the fiber. They complained that their friends just needed a $20 NIC (that most had from being in the dorms) and that they had to purchase more expensive fiber cards. In the end we quit wiring with fiber and went to copper.

Re:It's been done

[GrandCow]

I'was there last year, living in the Electronic Village. One of the problems with having so many ethernet lines running, with very few people that actually know what's going on, is the horrible service. The lines would randomly stop functioning, and access was very, very slow.

I'm all for ethernet right into the home, but just getting the lines out there isn't gonna solve the problem, we have to make sure that theres enough competent people running the whole thing to make sure that it actually serves a god purpose.

-C

Re:It's been done

[alhaz]

It's also been done in American Fork, UT. With a gigabit backbone as well. I used to work for the company that did it.

Re:It's been done

[Kalak]

The details of the implimentation of the original BEV project are still available.

Re:It's been done

[selectspec]

Yeah, BellSouth already has this here in some GA cities. This is what they use instead of copper DSL in some neighborhoods. They run fiber into the neighborhood, then drop regular old Ethernet from the fiber box into the guy's NID (network interface device is the box on the side of your house). Finally, BS requires PPP over Ethernet to encapsulate traffic from the customer and get it backhauled to the customer's network service provider (Digital Agent) via the same ATM connection the NSP uses for copper DSL ATM PVCs.

Re:It's been done

[Thucydides]

I have the 100Mb service mentioned in the previous post. It's called SwitchPoint (used to be called AirSwitch) and it's currently available in Springville and American Fork, UT.

It's basically a hacked version of Ethernet with whole town (most of it anyway) set up as giant switched LAN. Small pods connected to the backbone are located in each neighborhood. Each home using the service has a weatherproof CAT 5 Ethernet running from the pods to the house (usually underground). It's brought into the house just like a phone line and they install a wall jack inside your house. You get a static IP address and you're good to go. With the basic service you're limited to 500 MB per day, averaged over a seven period (ShoutCast streams will quickly get close to your quota if you're not careful). It was a pain to get them out install the whole thing, but dang it's fast!

Re:It's been done

[Xerithane]

A bit more pricey, but Ashland, OR has a fiber network allowing 100Mbit to almost the entire city for a starting price of a bit over $650 for the base package.

Re:It's been done

[MrBogus]

I'm curious if they are doing port filtering, or if your whole town is just one big NetBIOS broadcast storm.

fiber to the basement?

[mach-5]

Actually, the phone companies are working on getting fiber right up to the curb in front of your house. Problem is, they aren't gonna actually run it in to your house. I'm not quite sure why this is, but I can speculate that it has something to do with competing with DSL. My guess would be that the phone companies are going be ready to put fiber in your house, but are going to keep selling DSL until there is something else to compete with fiber.

Re:fiber to the basement?

[Evil Grinn]

My guess would be that the phone companies are going be ready to put fiber in your house, but are going to keep selling DSL until there is something else to compete with fiber.

My understanding was that if the phone lines in your neighborhood have been replaced by fiber, then you must use IFITL instead of DSL. And of course the phone company are the people who will provide IFITL.
---

Re:fiber to the basement?

[gabriel_aristos]

BellSouth is doing this in some new subdivisions. The crappy part about it is they cap you at about 1.5megabit/sec d/l and 128k u/l, and force you to use PPPoE.

Useless.

Re:fiber to the basement?

[autocracy]

They do this for the same reason that they don't run phone lines into your house. You're just so used to everybody having lines and the builders of people's houses doing that work at the same time as everything else. In fact, no utility does any work past their meter or switch box - not the water company, the phone company, the electrical company, or the cable company - at least not without you paying extra for it.

I can't be karma whoring - I've already hit 50!

Re:fiber to the basement?

[sjames]

In fact, no utility does any work past their meter or switch box - not the water company, the phone company, the electrical company, or the cable company - at least not without you paying extra for it.

At least in my area, the demarcation is at a small box mounted in the back of the house. They are responsable for everything up to the modular jack. The house wiring just plugs in to that. In the case of fiber, they are typically running it either to the head of the subdivision, or to smaller boxes at the curb. The run from curb to demarcation point is copper.

I suspect that a good part of this involves not wanting to undercut their $1000/month T1 business. That would go right out the window if they provided static IP and reliable DSL (without no server restrictions) at about the same cost to them, but $40 a month to customers. Currently, it looks like they are spending more money on DSL provision to make sure it doesn't become as good as a T1.

Re:fiber to the basement?

[unitron]

He was probably using that Simpson meter because he knows that he can trust it year after year, and if it's analog instead of digital you can spot transients and intermittents with it that won't show up in an understandable way on a digital.

Re:fiber to the basement?

[sjames]

In many cases, an actual T1 is not really necessary at all for a business, but allways on access is. In this sort of environment, the web server is typically in a colo though mail may well be a local machine. If they have experianced staff and a low traffic website, they might want the webserver to be local as well. In truth, they would be quite happy with 128Kb dedicated.

What they end up with is a frac T1. What they really need is DSL w/o the many restrictions and stupid PPPOE. They wouldn't mind paying $200/month for the lot (That's presuming $1/dedicated Kbps/month w/ 1 static IP + $72/month for the actual DSL). They end up paying around $1000/month. For a small office, that adds up fast.

Many people have noticed that the same line that "doesn't qualify" for DSL magically qualifies for more expensive ISDN.

Damn.

[Anonymous Coward]

I was looking forward to Arcnet to my house.

It's getting to be fairly common in California.

[jcr]

A number of apartment buildings around here (Santa Clara) are buying T1's and serving DHCP in each unit. It's becoming something that they feel they have to offer, just like Cable TV. -jcr

OSPF and BGP

[Phizzy]

Ok.. so I'd like to have ethernet or fibre into my house, but I _still_ wouldn't need to use OSPF or BGP. OSPF is a medium-to-large network routing protocol, which cannot be used on the internet, and is not suited for home use unless you have a _REALLY LARGE_ network at home, otherwise it would be a waste or resources.. static routing with a defualt route is much more efficient. And unless you plan on having more than one 100mb eth line or fibre line pulled into your house, and buying a router which can handle something around 200,000 routes (the current internet routing table x2, one view from each provider), the BGP isn't going to do you any good.. BGP is only helpful for a multi-homing situation.

SO.. do you research on your routing protocols.. you'd more likely need PPPoverEthernet for dummies, or maybe Routing for Dummies.

//Phizzy

Would you get 100mb?

[Sc00ter]

I don't know if you'd get a 100mb line into your house for so cheap. Just because they can do it doens't mean they will. Most cable modems do 10megs both ways, but most cable companies cap them at a much lower speed. I belive the DOCSIS standard will let them go up to 30megs both ways.

So just because the possibilty exsists don't expect to get it for dirt cheap right away. It's real easy to cap bandwith on a switch. And they will do it.

--

Re:Would you get 100mb?

[Arethan]

30Mbps is correct. DOCSIS does require that all compliant modems be capable to delivering this throughput for _downstream_. The upstream can vary from 320 Kbps to 10 Mbps. I used to work for a cable company in their broadband internet department. They capped their modems at 128kb upstream, and I have yet to find a cable provider that lets their modems run wide open. Better to cap the throuput in the firmware, and squeeze as many customers as you can into an OC192. At least, that's the apparent trend.

http://www.cabledatacomnews.com/cmic/cmic3b.html has a quick list of the DOCSIS specs.

Ethernet runs over fiber, too.

[hamjudo]

Once ethernet becomes a common last mile solution, fiber will surely be an option when copper can't go the distance and/or provide the bandwidth.

Ethernet over copper maxes out at 1 Gigabit. Ethernet over fiber is available at 1 and 10 gigabits today, 40 gigabits this year and 160 gigabits really soon now.

Some people may just get fiber for distance reasons.

Please explain.

[mindstrm]

Please explain what you mean by 'cannot be used on the internet'

Re:Please explain.

[Phizzy]

OSPF is a link-state protocol, rather than a distance vector protocol.. these are the two basic types of protocols. Basically, LS is like having a map of routes, from which you can determine the best path to a given destination, whereas distance vector is like having directions.. When Link State Databased get really large, they get unmanageable.. if you tried to run the internet on OSPF, every time a route changed, you would have to recalculate the LSD into the routing table, using the shortest-path algorithm, which is very processor-intensive.. and besides, since you have to keep a LS database as well a a routing table, you would need roughly 2x the memory.. BGP is a distance vector protocol (well.. some would say path vector, but we won't go there), which scales VERY well, so it is what is being used on the internet, but it is such a robust protocol, you need a large router to run it, and it only really matters if you have 2 different paths to take, which BGP can differentiate between (multi-homed), rather than just one connection.. (single-homed).. when you are single-homed, BGP is just a waste.

//Phizzy

Re:Please explain.

[_ganja_]

I think he means that due to the way OSPF builds neighbor relationships and the fact that LSAs get flooded throughout an area, it's not practical to use it in a very large scale network such as the Internet. So, it can be used on the Internet and of course is the IGP of choice for many ISPs but it just wouldn't scale to be the *main* routing protocol of the Internet, it wasn't ever designed for this anyway. Imagine 100,000 LSAs that needed to be refreshed every 30 mins :-)

I'm fairly sure that the poster of the article didn't *actually* mean we would actually all have to run OSPF or BGP to have ethernet to our homes but it was an attempt at a bit of a joke.

Re:Please explain.

[_ganja_]

Corrections:

BGP *is* a path vector routing protocol as it provides a vector to a path and does NOT use distance. RIP uses a distance in hop count, BGP uses the path in as_path. Path vector, period not "some might say".

Every time a route changed you would need to run the SPF algorithm? Nonsense. Its LSA not LSD, I think LSD is what you must be taking. Re-read the CCNA book again.

You need a large router to run BGP eh? I've had full dual views running on a 2611 (lowend router).

Phizzy, go back to the books and study harder, you are getting some things very mixed up, you are comparing BGP with OSPF directly. This is like comparing apples and oranges, OSPF or another IGP is generally required for running BGP. Running BGP when you are single homed is a waste? What happens when you are single homed with your own AS, how do you advertise routes to your up stream providor? This is very very common situation especially with companies that want flexability in moving ISPs.

Ganja the CCIE.

Re:Please explain.

[_ganja_]

BGP is commonly refered to as a path vector protocol, if you want to split hair about it, I'm going to call OSPF a distance vector protocol as well then, as Intra-area OSPF is distance vector.

OSPF on very large networks? Well, 2,000 routers is something I consider to be a large network and it runs fine. The key is a good addressing scheme, summarisation and correct use of stub areas, OSPF is very scalable.

2 full tables on a 2600 is really no problem, I'm not suggesting it's best practice but the router certainly doesn't melt. How processor intensive do you think bgp is? Try sh proc cpu on one of your other routers to get an idea. A 2600 with 64meg can handle two full views and this was last year, not 1994.

There is NO reason to have your own AS, unless you are multi-homed. PERIOD. I've taught the BCSN and that does indeed teach that, it's a good rule for that level of course but doesn't apply to the real world.
I think you are really showing how much you know here "You should have your ISP set up a static route and inject that into their BGP w/ a network statement": Where does that route orginate from? The ISPs AS in this case and a lot of customers don't want this, one, as they will no doubt be multihomed in the future and two, as I stated earlier it ties them in to an single ISP a great deal more than with their own AS.

Why (you muppet) would you want a full table in this case with only one connection? You fail to even understand that you would of course filter BGP routes from the ISP as you don't need them and static to the ISP, just because you have a peer doesn't mean you have to have the full routes. Jeezzz

ISIS, what are you talking about a real IGP? I work extensivly with ISIS currently and I find it very feature limited, how do you handle Multi-point interfaces? or DDR? and it means running an extra protocol besides IP. Good luck with the CCNP, I think you'll do OK but please remember there is a lot more to this stuff than that cert.

Solves nothing.

[Anonymous Coward]

Great. So now they're going to dig up concrete and run antiquated CAT-5 to my house.

This is a complete waste. The cabling itself is not causing the logistical nightmare of the last mile, digging up that concrete ($$) is what's bringing us down. (IIRC, CAT-5's nominal length doesn't even come close to a mile anyway ;))

That solution that is evolving with competition is the most obvious, and most profitable; utilizing existing infrastructure to get down the last 'mile.' -- be it over telephone, cable, or powerlines -- or a combination of the three.

Digging isn't an option, and unfortunately neither is wireless. I would dig it (no pun intended) if the power companies in the US began to put whatever connectivity into their substations and fed us 2mbs+ power-line networking.

This would also push JINI, etc. into my household devices - if your TV can get online through the same socket it gets its power, it may actually begin to have useful networking features (ie. program channel 82 to browse to so-and-so URL. Maybe the URL for my baby monitor, whatever.)


Think about it. If you want to talk in private about the commercial aspects of anything mentioned, let me know. ;)

Jason (jfisher AT feroxtech.com)

SOMEBODY SET UP US THE KATZ!!

Umm.. Ethernet at home is quite common in sweden..

[_GNU_]

My neighbourhood has a fiber backbone linking ~50 Cisco 1924 switches providing us with ethernet access.. 4 neighbourhoods here have 100Mbit and some, like my neighbourhood have 10Mbit.. 3 neighbourhoods even have fiber in their walls and you get a fiber nic in the package when signing up... wellwell..
All neighbourhoods are also interconnected in a fiber MAN... Nice for them dvdrips ;)

This is Borlänge, in Sweden...

Hope you get your ethernet in the states soon so we can bring that transatlantic bandwidth to it's knees ;)

// _GNU_

Why stop at the computer or basement?

[mizhi]

I won't be satisfied until I get a permanent, wireless nic implanted into the front of my skull. Just imagine tests with that... "Damnit, google is down and I can't remember the algorithm....

WTF?

[GodHead]

I'm still looking forward to fiber running straight into the basement

So you're telling us that Fast Ethernet is too slow for you? Good lord man, how much pr0n do you need?

G.H

Please, not on Copper!

[bill_mcgonigle]

To whomever may be doing this implemenation:
Please don't do this with copper. You're going to have to run new cable anyway. Cat 5 cable is very expensive too, and with the distance limitations involved you're going to have to spend more on repeaters than on wire if you go copper anywhere near 100Mb. Please consider fiber instead. You won't be locked into an obsolete standard, it'll be cheaper in the long run [no pun intended] and you'll be able to sell alot more services over it in the future.
Thank you,
the world

Re:Please, not on Copper!

[No-op]

cat5 is not all that expensive, when you're buying on that level- I pick up 3~4 mile spools of cat5e from graybar for about 140 bucks. I can't imagine how much cheaper it is to buy it from a general contracting standpoint, but it must be quite a bit lower.

and cat5(6/7/etc) are standard specs, so calling it obsolete is kind of silly, when you think about it.

that being said, running some fibre strands to neighbourhoods and then running cat5e or cat6 to a house is not unreasonable by any means- making customers buy fibre nics is quite pointless, as well as non-backwards/forwards compatible. fibre standards aren't as nicely categorized as copper, for ethernet anyway.

Already being done

[TheBrez]

Funny, why wait for 2002/2003? McLeodUSA's ATS project is doing this right now in Cedar Rapids, IA. Anywhere from 256K to 7192K upstream and downstream (half-duplex). Price on a 7Mb connection is comparable to a T1 from the ILECs. If you're in the area and interested in more info about it, contact sales. I'm just one of the techies who makes it work. :-)

Re:Already being done

[cyber-vandal]

Heh, don't mention Girl Scouts around here. ATS' bandwidth will be suddenly be flooded with a million 'resume.doc' files.

IPv6

[stomv]

With ethernet to the home, and Internet use spreading horizontally (more end users) and vertically (more businesses setting up LANs and other multiple IP-grabbing sub networks),

Should we expect an IEEE move to IPv6, and when?

I want more

[BroadbandBradley]

current cable modem technology based around the DOCSIS standard can give you up to 38Mbps downloads and 10 Mbps uploads. Cable companies currently cap that potential to about 1/10th of that. The Coax cable coming into your house has a 'bandwidth' of 750Mhz and each TV channel uses 6Mhz slots. Cable modems use 1 slot upstream and one slot downstream, meaning about 12Mhz out of the 750Mhz is used for your data connection. Digital TV boxes now can stuff about 10 TV channels into one 6Mhz slot. Obviously most of that 750Mhz is used for a broadcast medium.
Cable vendors can easily scale this approach to take away from 'broadcast' and move to 'download TV' where you could open up a guide and see what's on for the day and then pull a TIVO to download the show only if it's requested. in theory, this would allow them to dedicate more of that 750Mhz into Data-like connections and to provide that 38Mbps to anyone who wants it using the equipment currently installed in your home(if you already have a cable modem). think about 750Mhz divided into 6Mhz slots gives you 125 slots at a potential of 38Mbps per slot you come up with some 4,750Mbps downloading potential. Of course this approach would change the way people watch TV and fly in the face of traditional broadcast networks, but technology wise, the Cable providers are already there.
apllications? well I already stream my music in from the net at 128kbps, and downloading the latest Mozilla only takes a few minutes. Getting a copy of a new Linux distibution as ISO images (650Mb) still takes awhile.
still, I want More
currently I'd like more upstream to be able to do DV quality Video conferencing. I'd like more speed to be able to watch DVD quality video from the net like I stream my music today. DVD quality video can not yet be had with 100Mbps ethernet connection. I'd like to see them shoot for Gigabit ethernet to my house, I really need it.

Re:I want more

[No-op]

Consider that regardless of how fast your connection is on the cable network, you still have a limited pipe to the internet. I wouldn't dream of giving end users 38Mbit connections a piece- that would just pound on whatever connection I had, be it an OC-48 or OC-192, even.

It's the kind of thing where "if you build it, they will come"... someone develops an app like napster for movie trading on a large scale, (that actually works well!) with all these kiddies sitting on huge fat pipes and the whole internet turns to shit instantly. napster was bad enough. just deal with having connections NOW that most of the world can only dream of :P

Re:I want more

[sjames]

I wouldn't dream of giving end users 38Mbit connections a piece- that would just pound on whatever connection I had, be it an OC-48 or OC-192, even.

Use fair queueing on the upstream, and allow the full 38Mbit across town (within your own network). Offer an OPTIONAL squid proxy. Customers who use it will have a really nice web surfing experiance. The more paranoid will still do OK.

If the federal government here in the U.S. really wants to get the boom economy back, it might do well to pressure phone and cable companies to quit stalling technology in their endless effort to wring out the last penny, and just enjoy the substantial profits that are there for the taking now.

Re:I want more

[BroadbandBradley]

Badwidth is getting cheaper faster than Microprocessors. Most people will only download one thing at a time, a Movie perhaps. if it took 10 minutes to download a 2 hour movie, then the next hour and 50 minutes is sitting at Obps. the biggest challenge is getting networks to allow their content to be delivered in this manner. if all my cable TV content could be kept on a local proxy, there wouldn't be an impact on the net as a whole for at least the Network TV subscription services.
I constantly try to use as much bandwidth as possible on my connection, yes I am a web hog, because if there isn't demand, why scale. I want it to scale so I make all the demand I can generate. The internet as a whole will continue to expand and will grow to meet these needs as more broadband services become available like mp3.com and pixelon.com.

japan

[Alien54]

Someone mentioned in a comment in the past week just this sort of thing happening in japan. 100mb access to homes for $40/month

Between this and all of the other tech toys that never seem to make it to the rest of the planet, it sort of makes one jealous

IP v6 better get here soon then.

[Domini]

For permanent connection and fixed IP addresses one thing that is important is to pre-allocate IP addresses in such a way as to allow future expansion. Thus it will be important to allocate at least a class C (256 nodes) to every home.

So, I think it's going to be imperative for IP v6 to become more utilised soon.

Re:IP v6 better get here soon then.

[mrRaist-]

Class C to every home? What would the average home owner do with a /24? There are many reasons why they WON'T allocatate a /24 to every house.

1. ARIN does not allow public addresses to be allocated for workstations. I'm sure that ARIN considers a toaster, TV, home security system, PDA, whatever to be a workstation and therefore doesn't need a public IP address
2. If I had that kind of connectivity to my house, I would be certain to run some kind of firewall software. Even the most novice of user can setup Internet Connection sharing on their WinME computer. Most /.'ers will opt to some kind of Unix platform I'm sure, but none the less, a firewall should/would (I hope) be in place thus allowing private IPs to be used
3. How many computers do you have in your house? Between myself, my 2 roommates (all 3 of us are Telecom students) and our landlord, there are only 13 computers in total. Even if we had a killer LAN party and had 14 of our closest friends over, with 2 computers each, thats still only 41 computers.
4. 
   
   Now, don't get me wrong, I'd love a /24 to come bundled with my spankin' new 100mbps Inet connect, but I honestly can't see it happening. Most people could hardly justify a /29.
   
   My $0.02
   
   Brad

Internal competition?

[BlueUnderwear]

> I'm not quite sure why this is, but I can speculate that it has something to do with competing with DSL.

You think this is bad? Here in Luxembourg, the P&T (national telecom operator) doesn't roll out DSL in certain places for fear of competing with its (much more expensive) leased line offering. Kirchberg, which already has fiber-to-the-curb, never will get DSL, for fear that all the banks located there will drop their leased-line subscription and get DSL instead.

Asymmetrical

[Malc]

I bet the ISPs, especially the big ones, will still make it asymmetrical for residential users. My ISP gives me 1mbs\120kbs ADSL. Downloading at 100kB/s uses at least 20% of my upstream bandwidth. Any significant upstream activity quickly bites into my downstream throughput. It really sucks. However, my ISP obviously views it as an effective mechanism for stopping users running bandwidth-sucking servers. Unfortunately I don't have any other choices: the other ADSL re-sellers get the run-around from the telco, and they have transfer limits too; the cable company on the @Home franchise is just abysmal.

Re:Asymmetrical

[Greyfox]

You can get anything if you're willing to pay for it. Speakeasy was quite happy to hook me up with 768K both ways and the price is pretty reasonable considering I used to get 128K ISDN for the same cost.

IPV6 could wait if this was all done RIGHT!

[cr0sh]

The majority of homes - heck, I would say the majority of geek homes included - should only need a maximum of five addresses, if that. Your home network should be NAT'ed behind the firewall - after that, the network class could be damn near whatever you wanted. With the right firewall (read , a good one), you could have any addresses you wanted, or you could go the cheap route, and use the unroutable address ranges (10.x.x.x, there are two others, can't remember them off the top o' my head right now), for a NATural (in marketing-speak) firewall (heh, side note - have you noticed that is how they market the low cost firewall routers, such as the ones by Linksys? They call them natural firewalls - do they really think NAT means NATural?)...

I have a friend who lives in what I can only call a bachelor pad, who runs a cable modem with now firewalling at all, and each guy in the pad pays for their own IP. I keep trying to tell them how it would be cheaper (and better, since they run winders like mad) for them to NAT the place, but they won't do it - too hard to set up, I dunno.

The cable companies and DSL companies both have a marketing campaign to get the most bucks out of people by exploiting their lack of knowledge of networking. If they could get away with it (and I bet a lot of people are dumb enough to do it, if the telcos/cablecos could technically do it - actually, the cablecos can, they've been doing it with TVs all along) they would charge for a new line to each machine.

I hate fucking companies who prey on other's ignorance - then try to ram it down the throats of individuals who KNOW better.

Worldcom - Generation Duh!

Re:IPV6 could wait if this was all done RIGHT!

[Cato]

NAT breaks many applications (e.g. active FTP, NetMeeting, IPSec VPNs with IKE, many online games and so on). Where applications can get round NAT, they become more complex - e.g. Groove and other P2P apps must jump through hoops or rely on central servers. NAT does provide some security by default, but that's mainly by making it very hard for outside clients to talk to inside servers; a firewall can provide equivalent security quite easily.

If you ever want to be able to call in to a system at home (e.g. to tell your TiVo to record something), it will need a non-NATed configuration. IPv6 is the only way to do this without quickly running out of IP addresses - RSIP (Realm-Specific IP, a combination of tunnelling and IP address management) doesn't solve the 'call ing in' problem as far as I can see.

Before too long you'll want laptops to be able to roam between 3G networks when away from home, and then roam back to your Wi-Fi (802.11) wireless LAN at home. IPv6 enables much simpler IP mobility, i.e. your laptop keeps the same IP address (at least as far as TCP's concerned) no matter where you are. None of this is possible with NAT getting in the way - in fact, getting rid of NAT is one of the main reasons for IPv6.

For more information, see www.ipv6forum.com.

Re:IPV6 could wait if this was all done RIGHT!

[cr0sh]

I agree with you on that NAT can break applications, but I tend to think it isn't because of NAT, per se, but rather those applications aren't coded with the possibility of NAT in mind. Obviously many other commonly used applications can go through a NAT'ed system just fine - browsers, telnet clients, FTP, email, etc. The idea that we should get rid of NAT because some applications aren't coded properly just doesn't make sense in my book.

I also agree that a regular rule based firewall would protect just as well, provided that you had the IP's - such firewalls are available cheaply or freely for both Windows and Linux (and I suspect Macs and others as well), so that isn't an issue. However, the issue (or at least it should be an issue) with "consumers" (god I hate that term) is that the providers want to charge "per-PC" connected. $5.00 per IP may not sound like much, but 10 devices later it is a chunk of change (not possible? Let's see: Your two machines, your SO's machine, your kid's machine, the fridge, the two cars - audio, etc - the three TV's - BAM! There ya go!). Even if it was only $1.00 per IP, that would still be $10.00 extra dollars a month.

Calling in from outside? That is possible under NAT, just give the access to route a specific port to an internal address. I know it is possible, because I am planning on setting up a personal bookmark server behind my NAT firewall, that I want to admin. I looked into what the LRP could do, and what various distros based on it could do, and all of them allowed this. So, in theory, it should be easily possible to route to an internal box with a properly NAT'ed system.

Your last point is simply a matter of convenience. I can see having the transparency of a single IP no matter where you are - that would be a nice thing. Not that I will ever have a laptop with wireless capability (kinda outside of my price range, and I really don't have a use for it yet), but if I did, with a NAT'ed solution I could do a couple of things - I could manually change the IP (a slight pain), or I could set up some kind of system to detect what I was attempting to use, and have it automatically change the IP for me on the fly. Unless I am running some form of server on the laptop where others would need to know my IP, there shouldn't be any real problem.

I am not saying we shouldn't use IPv6, we should. But you better bet that when it is rolled out for all the plebes and their dogs to use, the providers will charge as much as the market can stand for each and every IP address, which is one good reason why NAT should be embraced.

Worldcom - Generation Duh!

Re:IPV6 could wait if this was all done RIGHT!

[cr0sh]

Essentially you set up the NAT'ed firewall system to listen for requests on a certain port. When it receives those requests, they are forwarded on down to a server on your internal network. The server processes the request, sends the response back to the NAT firewall, which translates it back, and sends it on to the user.

It shouldn't matter what web server you are running on that machine - but if your are running a NAT firewall system that you can't redirect ports, then that is a problem (I am sure some of the better NAT firewalls allow it - perhaps even TINY (for Windows) allows it - however, I am not certain that some of the "hardware" firewall/routers allow it (like the Linksys) or not).

You answered your own problem, I'm afraid - don't use "crappy" software (or hardware, whatever the case is). When I decided I needed a server for my bookmarks, I knew I wanted it behind the firewall. I knew that my firewall had to be able to forward the requests, so I looked and saw what my options were. I was running ZoneAlarm with a proxy server on a machine to act as my pseudo NAT system (I know it isn't, but what the hell else do you call it, other than a proxy server, I guess), but this wasn't the best solution, and it wouldn't work with what I wanted to do (plus there were a slew of other problems). I couldn't find info on whether the Linksys stuff would do what I wanted (hey, it may), so I decide to just go ahead and build my own router/firewall, run FreeSCO (based off of LRP, which will do what I want), and leave it at that.

Finally, if I didn't want a service visible to the outside world, I wouldn't hook it up to the net to begin with. Your last line is telling, though. NAT works well if you know what you want to do and how to do it. I am sure in many instances it simply doesn't play well, more than likely because the software doesn't want to communicate in a more standard manner (instead opting for something else for speed reasons, like FPSs). Maybe it isn't for you, but for my purposes I think it will work fine (and is cheaper in the long run).

Worldcom - Generation Duh!

Re:IPV6 could wait if this was all done RIGHT!

[Cato]

NAT breaks applications that are rash enough to embed IP addresses or port numbers (including FTP where you have to use the annoying passive FTP variant to get it to work) - this is common in slightly more advanced applications, and having to design around this is an unnecessary restriction (a bit like having to code within a 640K memory limit these days).

When IPv6 is deployed, a competitive market will grow up that delivers thousands of IPv6 addresses for a similar price to one IP address today. There is no extra cost to doing this, so any ISP that does not will be undercut by one who does. IPv4 address pricing is a direct result of the scarcity of address space. There are still hundreds to thousands of ISPs in most developed countries, so there is plenty of room for competition.

Your argument on IP mobility is similar to that for applications that don't code around NAT restrictions - there are always workarounds (e.g. reconnect all your TCP sessions) but they will cramp your style when moving between different Layer 2 technologies. If you don't want a wireless laptop, others will, and there are many other internet-connected devices such as PDAs and mobile phones that will also benefit.

The whole drive to P2P applications makes it much more desirable to have a static IP address. Again, not essential, but once you have enough IP addresses, application developers can concentrate on the important parts of their apps, rather than on beating NAT and the lack of static IP addresses.

Re:Dorm room?

[McSpew]

No, not like having a dorm room in college.

College dorm rooms are connected to large LANs which get their bandwidth the old-fashioned way (via routers connected to T1s and T3s which speak frame-relay or ATM).

What's being proposed here is using Ethernet signalling to carry your traffic from the phone company's CO to your house over plain old CAT3 copper, instead of using one of the DSL variants.

The concept of using Ethernet to carry signal down the last mile is not exactly new. Nortel Networks came up with a technology called Etherloop three years ago. Bob Metcalfe wrote about it extensively in his InfoWorld column back then. Nortel wound up spinning off a startup called Elastic Networks to develop and market the product.

Etherloop's biggest feature is that it automatically compensates for crosstalk in binder groups by treating them as Ethernet collision domains, although it doesn't actually incur collisions as normal half-duplex Ethernet does.

Re:Athentication?

[DrgnDancer]

Doesn't cable have the same basic issue though... You're still sitting on the same network loop as everyone else in the neighborhood. Actually, a good possible solution to this (considering the amount of bandwidth that would be avavilable) would be to have the subscriber's router and the local exchange router encrypt traffic between them. It would require a bit more power out of the local exchange router, but not so much as to be a huge issue. If each subscriber had their own encryption key, there could be no neighbor snooping. Anyone else see a problem with this scheme?

My issue with the article was this:

The flexibility of such an architecture is enormous. Over one connection, a user could conceivably run an Internet hookup at 6 Mbits/s, four concurrent telephone calls at 0.064 Mbit/s each and four concurrent videophone conversations at up to 768 kbits/s each

I personally have never gotten anything close to 10 Mbits a sec on a 10 Mbit/sec ethernet connection. At best they could reasonably hope for 5, and that would assume nearly ideal conditions.... While we are still talking about a vast improvement, I think they are being a little optimistic.

Surely the inhibitor is further upstream

[biglig2]

What's the point in having 10MBps to your ISP from home? The inhibitor is surely going to be what the ISP has to the internet.

Example: One of my ISP's is UUNET.

Across the Atlantic they have 5 x OC3c lines and 2 x OC48c lines which comes to... about 5.7 Gb/s

Where it might help I suppose issituations like you have now with cable companies, when you are accessing sites in the same ISP; or for sites that have been cached at the ISP level.

This is one reason why broadband companies are so keen on stuff like "watch movies from our media servers" because that's not choking their upstream feed.

Yeah.

[mindstrm]

That aside.. what does that have to do with using the protocols on the internet? It doesn't.

The original post said 'cannot be used on the internet' not 'cannot be used to run the whole internet'.

*yawn* Playing catchup again, kids?

[Snocone]

We've been able to get home Ethernet in Vancouver for AGES. You still creeping along at those glacial cable/DSL speeds? Dude, that is like so 2000.

Check out Novus High-Speed Internet.

Note that the prices are C$ arctic pesos and not real dollars to boot...

You can get away with less than 64+ MB of RAM

[jabbo]

Your upstream provider can offer to only propagate routes reflected within their netblocks, or default routes only, for BGP4.
If you don't have an upstream provider you won't need to worry about any of these things :-).

Otherwise nobody with a wimpy Cisco (eg. cheaper than a 7000 series) would be able to multihome.

The comment about OSPF is fair, though. What a fucking nightmare... I hate to admit it but the behavior of OSPF nudged me into running EIGRP. (the shame!)

RJ45 drops in my bedroom? Like it, but...

[connorbd]

It'll be interesting to see how this is pulled off. The big problem with it is that it might be even more limited than DSL unless you're redefining the Ethernet standard.

In 1995 I worked on an early broadband networking project at Boston College as a tester (Continental Cablevision, now AT&T Broadband). The main difference between the BC system as it was when it went online and the usual broadband thing was that there was no cable modem per se; the network/CATV signal came into the dorms over a fiber backbone and was split. The cable signal went to a series of coax taps, while the network signal fed to a large IBM box (essentially the "cable modem" and from there into a series of hubs.

This is a great system for a college or an apartment building; you just steal a closet in a hallway here and there, and then wire up the rooms as the opportunity arises. There are basically two problems once you get out of this setting, though...

The first is that Ethernet, like DSL, has a limited range. As I said, the BC network was (and is, I'm sure) built on a fiber-optic backbone. The Enet only comes into play when you enter a building; that's fine. However, there was an interesting layout problem; anybody here who's been to BC will know exactly where I'm talking about...

The Mods are a large patch of creaky, thirty-some-year-old prefab rowhouses that dominate about a quarter of BC's lower campus. Despite the running joke that "the Mods will be torn down by the time you're a senior", they are the most visible institution of campus life. Wiring them was... interesting. As I understood during my time on the testing crew, the Mods were split into two sections, each with its own feed from the backbone. In order to supply each block (with either two or four units per block), the cheapest way to do it was to steal one closet (referred to by students as "the keg closet"; obviously we're talking Party Zone here) per block as a network closet and lock it permanently.

This becomes a problem like so -- if you go into a residential neighborhood, where are you going to put the central network hub? Obviously you can't just rent out space in someone's basement. Building a "network shed" on each corner or telephone pole isn't especially practical either; for a dense neighborhood, can you imagine the thick bundle of Cat5 cable that you'd have to hang off the telephone poles or bury?

Now one could assume perhaps that these problems have certainly been worked on since 1995. But my thought on Enet@home is that you'd probably still be better off with a fiber or DSL drop coming in the front door and building your own network off of a router. I'm not saying it won't work, but I don't see it as being terribly practical...

/Brian

Re:RJ45 drops in my bedroom? Like it, but...

[connorbd]

And just barely practical there.

We have a big green box on a telephone pole too. But you can only have so many big green boxes, and retrofitting an existing neighborhood would be rather difficult. Also, it's a question of how old your neighborhood is -- in a dense area like Boston's inner suburbs it would be a lot harder to find space for something like that than it would in some three-year-old cul-de-sac community outside Phoenix. I stick by my original point: in apartment buildings and *maybe* new subdivisions it's workable but really wouldn't fly in already-developed areas.

/Brian

oh, and one more thing...

[connorbd]

I should disclaim this: this is based on my personal experience; IANANA (...network admin...)

/Brian

I want my fiber!

[fm6]

...but I'm still looking forward to fiber running straight into the basement.

For your megapixel VR version of Quake, I assume?

__

Re:What does ELF stand for?

[BlueUnderwear]

As far as Linux is concerned, it stands for Executable and Linking Format.

Re:Athentication?

[Tassach]

Doesn't cable have the same basic issue though... You're still sitting on the same network loop as everyone else in the neighborhood.

Yes and no. Yes, cable modem is a shared pipe. However, with a DOCSYS cable modem (virtually all, now), all the traffic between your modem and the head-end router are encrypted with DES (with DH key exchange IIRC). Also, there is also authentication done via the MAC address of the device(s) connected to the cable modem. The cable modem is actually a router, and is not supposed to pass any packets to the LAN other than ones directed to the ip(s) it serves. In some of the early rollouts, misconfigurations were common, allowing MS networking to see everyone on thier local loop in the Windows 'network neighborhood'. Most cable modem ISPs now filter the SMB ports

I personally have never gotten anything close to 10 Mbits a sec on a 10 Mbit/sec ethernet connection.

Of course you have. A large percentage of your total bandwidth is taken up by protocol overheads - frame headers, packet headers, control messages, which is why your maximum effective data transfer rate is going to be around 8 Mb/s on a 10 Mb/sec connection if you are using TCP/IP; probably a little higher if you were using UDP.

The only way to get transfer speeds approaching the maximum bandwith would be to write a minimalistic UDP-like data-transfer protocol which could read & write raw ethernet frames. The speed gain from such a protocol would be offset by the unreliability that would be introduced by stripping out all the error-checking and reliability elements present in TCP. It's impossible to use the full 10Mb/sec for your data, because the ethernet frames themselves have some overhead.

Re:Microsoft should love this

[Oscar26]

Well, what is the maximum amount of bandwidth required for Streaming video, a telephone call or 2, and a few web connections? I'm thinking of a family of 4, one watching a movie, 1-2 on the phone, and 2-3 surfing the web (the family members can be doing more than one item at a time.)

Someone before posted 768Kbps for a video, voice is 64kbps so that gives us 768 + 64 + 64 = 896Kbps, or We'll round up to 1Mbps. If you had a 10Mbps, it seems that connection would/could easily satisfy your needs, provided that the ISP could handle the traffic.

It also depends on the type of traffic that a web brousing session would sustain. It wouldn't be a constant or even exponential distribution, it probably would be more like a pareto distribution, must more bursty.

Who knows the bandwidth requirements of the future? I can only speculate. I'd say 10Mbps is a major improvement for all of us. Heck, a 768K DSL connection would be a major improvement for most everyone, saying cable modems have 5% of the online market, DSL has 2+%, and I guess a small amount is LANs and direct connections, but most people (80%? just a guess) still dial up the old fassion way.

BTW- I hate it when a analog modem dials up. Too much freakin sound. I'll be much happier in a quiet world.

Re:What does ELF stand for?

[cyberdonny]

Nope. According to this page, it was the name of the first rock group that Ronnie James Dio performed with back in the early 1970's.