LAN Turns 30, May Not See 40?

dratcw writes "The first commercial LAN was based on ARCnet technology and was installed some 30 years ago, according to a ComputerWorld article. Bob Metcalfe, one of the co-inventors of Ethernet, recalls the early battles between the different flavors of LAN and says some claims from the Token Ring backers such as IBM were lies. 'I know that sounds nasty, but for 10 years I had to put up with that crap from the IBM Token Ring people — you bet I'm bitter.' Besides dipping into networking nostalgia, the article also quotes an analyst who says the LAN may be nearing its demise and predicts that all machines will be individually connected to one huge WAN at gigabit speeds. Could the LAN actually be nearing the end of its lifecycle?"

Well, could it?

[Pharmboy]

Could the LAN actually be nearing the end of its lifecycle?

Yes. All computers in the future will be stand alone and the Interweb will be shut down.

Somewhat interesting article, stupid summary question.

As long as the need for a secure network exists...

[riseoftheindividual]

... the lan isn't going to disappear, at least not in 10 years. Can you imagine IBM, a defense corp, a huge pharma, etc... ditching their lans for wireless? yeah right, not any time soon.

LAN or WAN

[lthown]

doesn't matter what you want to call it, two computers connected to a local router/hub is a LOCAL area network.

going away?

[gEvil (beta)]

Could the LAN actually be nearing the end of its lifecycle?

Not as long as they let me control my own home network...

Re:As long as the need for a secure network exists

[HornWumpus]

It's not LAN vs wireless, it's LAN vs WAN.

Running a WAN without using LANs throughout is nonsense. IIRC a WAN is just bridged LANs by definition. Proposing that all the LANs will have one node is just silly.

Typical Bob Metcalfe of recent years. The man has lost it. Granted I haven't bothered reading anything he's written in a few years.

Silly prediction...

[MavEtJu]

The LAN as we knew it, the one ethernet cable going through all rooms and being looped on the wall with a small jumper, is already dead for a long time.

The LAN as we know it, one central switch with a lot of ethernet cables getting out to individual ports in rooms, has been here for ages.

What didn't go away was the local addressing methods for sending data to all hosts (broadcast) and interaction with higher level protocols (ARP for determining the IP address).

The LAN as we are going to know it, a bunch of intercepted central-and-not-so-central switches which put you in the right (V)LAN when you plug in your computer to a random port connected to it, is here also if your organisation requires it, but for smaller organisations this is not really necessary:

and predicts that all machines will be individually connected to one huge WAN at gigabit speeds

You need a gigabit WAN for that to work, not all smaller organisations have the need for this. But yes I have rolled it out for two customers.

Every doorway opens onto a freeway?

[dazedNconfuzed]

That reasoning amounts to expecting every doorway from every room to open onto a major automotive freeway.

LANs will survive indefinitely precisely because sometimes your data is just feet or yards away ... and because even Internet backbones can't handle the load of routing data for everyone's personal networked printers, storage servers, and media terminals.

Re:Well, could it?

[sm62704]

I agree, it's a stupid statement. Ethernet may be superceded by newer technologies, but there will always be uses for a local network.

Some networks, for example, should never be connected to the internet in any way.

Re:Well of course

[morgan_greywolf]

Don't trust any spec over 40.

Really? In that case:

I pronounce imminent the death of:
ASCII
UNIX
the mouse
the QWERTY keyboard
RS-232
SMTP

and lots of other completely useless technology.

Re:As long as the need for a secure network exists

[Anonymous Coward]

because I can assure you that my DVR will never have a public address

The DVR is one of the most logical computers to have a public address in the home. Think of the possibilities if every DVR acted as a bittorrent node.

Re:As long as the need for a secure network exists

[ronadams]

The comments on this article are making my eyeballs bleed.

1. RTFA
2. WAN != WLAN
3. Metcalfe didn't say a word about the idiotic question posed at the end of the article. Even RTFS would have told you that.
4. ???
5. Understanding!!!

Re:Reliability

[deanlandolt]

Insightful? Not Funny? Mods, really?!

That dubya in WAN does not stand for Wireless. It stands for "Wide [wikipedia.org]:, as in, as wide as the internets. That Belkin you speak of creates its own little LAN; a WLAN, if you will.

The future, Conan?

[Anonymous Coward]

Yes, Bob Metcalfe, all the way to the year 2000.

I'll make a bold statement- if the Ethernet switch could not have existed for some reason, Ethernet would be all but gone by now.

Ethernet is awesome, but without switches (or routers), all interfaces share a common bus. If 2 stations try to "talk" at the same time, a collision results, and they both have to try again. The more stations and traffic, the more collisions, the more retries, the more collisions, and it can saturate and become a third-world parliment fight.

Fiber networks are rings like Token ring. Token ring by definition has no collisions. One station is designated a master and sends out a token. You can only talk if you have the token, so token ring nets can use essentially 100% of the bandwidth. Extremely efficient and deterministic. However, if a station breaks, or the cable is pulled, the ring is BROKEN. Big nightmare.

IBM made very solid Token ring stuff- even had these great connector blocks with self-shorting connectors that kept the ring intact if you pulled out a station. But it was very pricey, so it went the way of the Betamax- technically better, but cost (much) too much more. Even with the old "thinnet", "T's", 50 ohm terminators, poorly crimped and flaky BNC connectors, Ethernet was cheaper and worth the many hassles. (Coming from a radio/analog electronics background, I had no problem with BNC splices, but I've seen some horribly crushed coax- some that still worked!)

When UTP (unshielded twisted pair) came out (1990ish), mass wiring got easier and telco people could install it reliably. But it could not, and still can't, be run more than 100 meters (in spec- of course there is good safety margin in the spec and you'll get longer runs to work fine.) I worked for a company where we literally strung a thinnet coax down the street to link 2 of our buildings. It was over 1000', as I recall, and with 2 lightning arrestors and 2 media converters, it worked perfectly. :) I don't know who ran the cable, but I was very impressed with the achievement.

When Ethernet switches came out, you could divvy up traffic much more easily than using routers, and when switches got really cheap, well, token ring is long gone.

Now with 10Gbit Ethernet (wow!!), switches, and trunking (parallel Ethernet paths), Ethernet bandwidth is keeping up nicely.

No wireless or WAN can come close to competing with the cost / bandwidth for local networks.

Fiber can be pricey, but is reasonable enough and great for interconnecting large campuses.

For me the bottom line is that if I relied on someone else's network, like a Comcast, Verizon, etc., and something broke, I have to rely on them to fix it. And I don't like that scenario. I want control of my realm, and I don't want a tech rep. from a Verizon telling me the problem is with my system when I know it's not.

So for direct control and management reasons, I want as much Ethernet and Fiber ring as possible, and only use WAN (VLAN, frame relay, leased line, etc.) where I can't string a cable down the street, or use a microwave link.

Re:WAN, SCHMAN

[snowraver1]

I think that the author is suggesting that each device will have it's own address (IPv6) and will be connected to the internet directly (possibly VIA shared modem, but with unique addresses). Sure you might only have one pipe coming into your house, but each device has a direct connection to the internet.

That being said, I completely disagree with the author. There is no way that companies want to put all thier servers (not to mention clients) directly on the Internet. Firewalls will always exist for security reasons, and thus so will LANs.

"Internet enabled"-everything

[DrYak]

Does my TiVo really need a direct connection to the internet, even a firewalled one?

Yes, of course ! How do you think that they'll enforce even more stupid forms of DRM (that will force RMS to counter writing even more complex versions of GPL) ?

And how do you think that de government will spy on you, using the RFID tag reader in your fridge and fine you if you don't buy the mandatory 10% corn-based products required by some law that some lobby pushed ?

In 10 years, even tinfoil hats will be network-enabled.

Re:NAT != Firewall.

[jeremyp]

NAT is here to stay and it's not an ugly hack. A company like, say, IBM does not want to have to go to ICANN every time it hooks another laptop to its internal network. Nor does the rest of the Internet need to know about IBM's internal network topology. NAT is actually a useful piece of technology to make TCP/IP networks manageable.

Re:Well, could it?

[Warbothong]

So the assumption is that LANs will go in favour of VPN type stuff over the net because net speeds will be fast enough? I think that's bogus, because people would be doing it now if that's the case. In 10 years time WAN connections will be very fast, yes, but I'm pretty sure LAN connections will be ultra mega fast. Couple this with moves towards thin client type applications being run from a business's server to its desktops and the unknowable crazy application ideas that will spring up thanks to ingenious startups and truly massive storage and transfer capacity and I'm quite certain that the killer apps being used in 10 years time will fail spectacularly if used over a mere GB/s connection and sysadmins will be laughing at these predictions like the 640K predictions we laugh at now.

Re:WAN, SCHMAN

[Dun Malg]

...I seriously question the authors assumption that LANs as we know them will cease to exist.

Indeed, this is often the problem with "visionaries". They have no real sense for the reality of the situation. It's like the quote supposedly from Steve Jobs at the private Segway unveiling: "Cities of the future will be built around this". This is a classic "visionary" statement. The same exact thing from a realistic (i.e. engineer's) point of view is: "Cities would have to be rebuilt before this thing would be particularly useful".

With regard to networks, it's basically inarguable that the many network-enabled devices in people's homes will be sharing a single pipe from an ISP. It is also essentially inarguable that (for the foreseeable future) Ethernet will remain the common hard-wire standard for network connections. Multiple Ethernet connections will require some sort of switching hub to manage the traffic into and out of the shared internet connection, as well as between the various devices. Wireless will likewise still require some sort of central access point. So where, exactly, does this "visionary" genius see the change happening? This is already what we have now, and there's no real reason to change it. Is it a veiled reference to IPv6? Is he simply saying that NAT is going to become superfluous and that somehow that means the same as "the LAN will disappear"? Is he really claiming that no one will firewall their home devices at their [cablemodem/DSL/FiOS] connection, and will choose to allow anyone on their subnet to come browse their shares? Seriously, the internet is a great tool for mass communication, but this ain't no hippy commune. Anyone with enough sense to come in out of the rain is going to want to separate their stuff from the rabble outside. And if so, how is that--- a set of IP addresses behind a firewall--- not basically a LAN?

Re:WAN, SCHMAN

[sumdumass]

One of the problems with placing firewalls directly on the devices instead of in a router or something somewhere is that defect in the devices aren't apparent until after they have been successfully exploited. More public Internet addresses means more problems in the end. Your actually doing yourself a favor by hiding hardware that doesn't need to be directly accessible from the internet in a subnet behind another device. There has been more then one virus that effected/infected the OS or services running on the OS that a simple router would have mitigated.

I don't expect problems like that to go away anytime within the next 10 years. I can see the effects and probabilities mitigated but not removed. A software firewall hasn't always been the best approach either. Sometimes it would crash the system, in situations like with symantec, the firewall itself could be exploited, and so on. Imagine if everyone did a flood attack or actually had a back door into your devices for years/months before it was noticed and patched.

Re:WAN, SCHMAN

[knorthern knight]

It does help to keep the bad guys guessing about your layout. Do you have 1 desktop, or a 100 servers? With NAT, they don't know. Makes a difference when deciding where to attack.