Slashdot Mirror
In NZ, Sharing Ethernet With A Whole CIty
ryuko writes: "Normally LANs are used by a single organization at best, but Wellington's 13-square-mile LAN comprises many of the city's businesses. The city council garnered a UNESCO Digital Access Award in recognition of its achievement in installing the 1,000 Mbps network.
The full article is here on ZDNet. Drool ... gigabit internet ..."
that is pretty cool. lots of other juicy details in there as well.
"It is a greater offense to steal men's labor, than their clothes"
Is this approach scaleable? Wellington itself is really not a large city and being the capital, an extraordinary portion of the business is governmental. Both of these have to cause problems when trying to extend this system beyond the "Windy City". BTW, Wellington is much windier than Chicago, the other "WC".
So long and thanks for all the fish . . . !!!
More like gigabit intranet. Once you hit the bottleneck, you're moving at the same speed as the rest of us. =)
Yeah but the point is that many US organizations can afford, and do afford, a lot more than a T1. Wiring a city solves the last mile problem, but 9 times out of ten the organizations that a company wants to video or audio teleconference isn't conveniently in the same town.
The article seems to be full of errors. Firstly they say that it's 1,000 Mbps, but then call it Fast Ethernet (which is 100Mbps), and then state that it's 67x faster than a T1 (which would imply 100Mbps). Later in the article they say "With 100 Mbps of capacity, businesses can easily implement video conferencing and voice over IP (VoIP)." 1000, 100, Fast Ethernet, 67x T1...blah.
Or not????
Better to s**t on everyone else, eh?
If the raw paranoiac/Hobbesian profit motive isn't behind it, most folks areound here would never go for it. Damn the benefits. Who knows? Maybe in 2050, members of Congress will be saying, "If we vote for legislation X, then we might catch up with New Zealand's GDP."
Acording to the article Everyone is responcible for their own firewall. So with the firewall who really cares what software you are running. Also sience they seem to be running Linux as the main servers I dont think they have a big corperation pushing them to run a perscribed software.
If something is so important that you feel the need to post it on the internet... It probably isn't that important.
why hasn't anyone thought of this before? I could certainly see broadband catching on as a public utility type of thing, instead of a luxury thing. Much how telephones and then cable television did in the past. Not only would it allow for cheaper overall costs, but having a citywide intranet @ gig-e speeds would be amazingly useful for telecommuting/VPN, gaming with friends, or any other number of good stuff.
OK - I think the ZDnet editor should get him/herself a dictionary of computer and networking terms:
Normally LANs are used by a single organization at best true, but for a good reason. LANs that span multiple buildings are technically refferred to as WANs, regardless of the underlying technology.
And the 2.5k$ gigabit router? Not. A commodity PC cannot even reach maximum throughput on a single gigabit NIC, nevermind routing between them. The only way to do this would be to use a decent server-class M/B with 64bit/66MHz PCI bus - which would take the total system cost above 2.5k$. A more moderate PC could indeed be used for residential/small business gateways, but you would not get gigabit throughput.
Just my 2c worth...
-justin
As the article points out, the companies hooked up to the Wellington backbone are responsible for their own network.
It states clearly that "It's a normal LAN with client-owned routers at the edge. Clients implement their own firewall protection"
The costs will probably be very low... using opensource and all their overhead will be at a minimum. The costs a company makes is nothing more then they normally would have to pay for materials like a router and firewall.. it can be whatever they want..
What i am interested in is if this scheme would work in rural area's. What would be it's breakeven point....?
Incentives mean that whoever lays the cable gets a monopoly on it. Then Time-Warner/AOL will start prioritizing packets so competitors' sites will perform poorly. After that, they'll realize that streaming video is competition to cable, and they'll limit video streams to a maximum of 10 minutes because "[They] didn't spend $56 billion laying cables just to have the blood sucked out of [them].
And there you have it. The internet will just become an enhancement to cable TV.
What we need is deregulation of the Cable internet access, like there is with DSL.
Those three letters were magic. At the time, computing was dominated by big expensive mainframes, and IBM had no less than 90% of that market. They were, in other words, the Microsoft of the 60s and 70s. To survive, your product had to be compatible with the IBM PC at every level. IBM itself took a long time to see this, and came out with non-compatible systems like the PCjr and the PS/2. Which is why the "IBM-compatible" market isn't dominated by IBM.
The one way Microsoft helped out was by providing a crappy operating system -- actually more like a glorified program loader. Since MS-DOS did such a lousy job of insulating applications from the hardware, apps had to incorporate a lot of hardware-specific functionality. Which forced IBM's competitors to emulate the PC at a very low level.
Everybody engineering to the same specs created opportunities for commodity manufacturers -- and created the "generic" computer. Which still has basic design features that totally suck -- like that big heat-generating internal power supply.
Perhaps if Microsoft had hired somebody who knew Jack Shit about re-entrant code or how to write a scheduler, we'd all still be using proprietary architectures. Kind of ironic.
QoS: No worries: Many IT departments say that prioritizing packets is vital if you want to run applications and send important files over the Internet. Because of Citylink's sheer speed and capacity, De Wit says adding quality of service (QoS) features isn't necessary. "QoS is a problem for others because they only have so much space in the pipe," he says. "We can fit all the traffic we want onto our Ethernet, so why do we need to worry about prioritizing?" Also, because of the generous capacity, DeWit says data collisions, which are often a concern on LANs, aren't such an issue with Citylink.
Seriously. QoS is a waste of time if you just have enough capacity.
sulli
RTFJ.