Slashdot Mirror

Wrong. it is /14 not /21. /14 is 4 /16 or ~65k*4 => 256k IPs.

~$ host stanford.edu
stanford.edu has address 171.67.22.34
stanford.edu has address 171.67.20.36
stanford.edu has address 171.67.20.37
stanford.edu has address 171.67.22.26
stanford.edu has address 171.67.22.33
stanford.edu mail is handled by 20 mx2.stanford.edu.
stanford.edu mail is handled by 20 mx3.stanford.edu.
stanford.edu mail is handled by 40 mx4.stanford.edu.
stanford.edu mail is handled by 20 mx1.stanford.edu.

~$ whois 171.67.20.36

OrgName: Stanford University Network
OrgID: SUN-5
Address: Pine Hall, Room 115
City: Stanford
StateProv: CA
PostalCode: 94305-4122
Country: US

NetRange: 171.64.0.0 - 171.67.255.255
CIDR: 171.64.0.0/14
NetName: NETBLK-SUNET
NetHandle: NET-171-64-0-0-1
Parent: NET-171-0-0-0-0
NetType: Direct Assignment
NameServer: ARGUS.STANFORD.EDU
NameServer: AVALLONE.STANFORD.EDU
NameServer: ATALANTE.STANFORD.EDU
Comment:
RegDate: 1994-08-22
Updated: 2000-08-17

RTechHandle: JK535-ARIN
RTechName: Kohn, Jay
RTechPhone: +1-650-723-7515
RTechEmail: security@stanford.edu

# ARIN WHOIS database, last updated 2007-05-16 19:10
# Enter ? for additional hints on searching ARIN's WHOIS database.

I don't really know how things are in the US but maybe at least parts of your motor ways, sub ways, rail roads, airports, harbors, electrical grid, telephone grid, water plants, waste treatment / facilities, telephone network (maybe not and you used more satellites?) and such has been built by some sort of community organisation of whatever level.

Anyway I don't really see why a fiber network is different.

Even in the plan by the guy from SUNET (Swedish University Network) the idea was to just dig and lay the fiber just as the municipals do. Companies would then offer the services.

(In the apartment I have now I can Internet access by three different means, five by with wired and wireless telephone network.

Ethernet jack hooked up to Bredbandsbolaget = What I have now (and what has been around since beginning of year 2000 here, with 10/10 mbps for 200 SEK / month back then. But it was the first place in Sweden supposedly.)
Ethernet jack hooked up to Stadsnt Örebro/Kumla = The municipal fiber network with a bunch of providers, it cost 75 SEK = less than $10 / month to have the access to it and then you pick a provider and service level.
Cable jack.
(ADSL by copper.)
(3G/4G cellular.))

Here's what I can pick of from Stadsnt:
http://www.stadsnat.se/service...

And yeah. I can get that by the lake in the "summer/hobby house" area, and yeah that was in part made possible by money from the EU (I think it did cost about 15 000 SEK = $1800 or less to be hooked up there, but then you could likely chose from just the same service providers as above by road about 25 km from this city with around 100 000 people living it it by a lake in a region where not everyone live year around.

I want to reply on the linked threat from the original article.

It talks about Sweden but the fact is only some of the municipals have their own fiber networks even though 12-13 years ago a go from Sunet suggested that one would build fiber to everyone just like the electronic grid and that the price would be a reasonable 50 billion SEK.

Sadly ADSL and cable modems started showing up and I guess the government retards and old fucks was to weak and stupid to make it happen.

It was of course a very good and much better idea than anything else.

Instead they built fucking TV antennas for digital TV (and will be upgrading for digital radio) and are still stuck with the old telephone network. AND people have less competition and quality on their Internet connection and not the same amount of options everywhere.

In the linked article someone make the lame excuse that the US have so many people whereas Sweden only have the population of New York and hence it's not comparable.

But it's all about density of the population. To be fair though US have larger cities and hence in-between them maybe more open space.

On the other hand IN something like NY there's no reason you couldn't have what Sweden have in Stockholm for instance.

Also someone compared with California which have got four times the people in about the same amount of space but shouldn't that just mean that there's better possibilities of doing it in California? An even larger city or more densely populated area = less to dig.

In the case of Sweden those 50 billion would be 5 000 SEK / person = $600 but that's NOTHING!

Having a fiber network is a long lasting infrastructure piece and having it built everywhere and others compete for providing bandwidth for consumers likely lead to much better price for them. The nationwide network would lower prices on Internet connection and over time $600 is really cheap.

And as said it could be used for stuff like TV, radio (possibly), telephony and things they may not want to do now because it's not as obvious that everyone got an IP connection (government and municipal service, health-care, declaration of taxes, banking, ..)

How do you know which of the painters that are long dead were trying to capture an actual Zebra with the actual amount of stripes in the exact pattern and colors they saw?

Statistically, it doesn't matter. When people run polls, how do they know how many people are lying to them? That's why they use large samples, so the signal can rise above the noise.

I'll find Zebras with unicorn horns and monkeys riding them.

You'd be extremely unlikely to do so, and even if you did, as long as you sample enough zebra paintings, its noise would be swamped by the signal. Most zebra paintings would show a realistic, if not 100% accurate, size and number of stripes, especially if the artist could look out of his window and see a zebra.

The great majority of artists strive for emotion, not realism

The other inputs still don't matter in a large enough sample. Grass (on average) will be green. Clear daytime skies (on average) will be blue. Leaves on trees in autumn (on average) will be golden brown. Zebras (on average) will have black and white stripes and be horse-shaped.

These people did not all have photographs to go by.

They had real sunsets to go by (a photograph, particularly a digital one, wouldn't capture the full range of colour visible to the human eye, so in that sense a sunet is better). Even if they didn't paint them "live" - though it's not unreasonable to suspect that most did, at least in part - they would still have been influenced by true sunset colours around that time.

This changes color and hue drastically, especially when you are mixing your own paints.

My understanding of the paper is that they looked not only at red/green ratios per painting, but also red/green ratio changes within the painting. That will have removed some of the kind of uncertainty you're talking about.

On the one hand there are these scientists, presumably well educated in their field, who have spent days and weeks poring over data and making calculations, and have come up with what they freely admit is a tentative proposal. On the other hand there are the people who declare it to be strictly impossible after reading an article about the study. Why should I side with the latter over the former?

I fetched some iso to /dev/null just because some US college kid had tried and failed like three times before and got shitty speeds (from his school?) and was just about to finish.

He had 15 seconds less so I decided it would be cool to finish before him. Sadly I had to cancel and change the url because he was downloading the CD image rather then the DVD image so I finished later =P

It felt rather wasteful when I tried whatever a download for me from the US mirror he used would actually fail to find out whatever the trouble was at his end or not (considering that passed the Atlantic to here in Sweden just to also be dumped to /dev/null for just checking whatever it would finish or not. Sunet among others host the images (it was a openSUSE image.)

Akamai posts another widely skewed report, based on their own crap infrastructure, where they are subpar for some regions.

In the Nordic countries, Akamai is a brake on everything, no matter what time of day you have to download anything via their infrastructure. I currently have a 100Mbit/s symmetric connection, and I get HIGHER download rates via Akamai if I use a US proxy than if I try a straight download. Same thing with any update services or games etc that use Akamai, Nordic countries get the shaft there too. I have a feeling that they are also underdeveloped in the asian regions, which would skew the results too.

Some ballpark figures:

Downloading an ISO via Akamai: Peak out at 16Mbit/s and averaging 11.3Mbit/s going straight, peak out at 29.5Mbit/s and averaging 15.4Mbit/s proxying to the US.

Downloading an ISO via Limelight networks at Swedish prime time: Peak out at 97Mbit/s, average at 94Mbit/s.

Downloading an ISO from SUNET's FTP at swedish prime time: Peak out at 98Mbit/s, average at 96Mbit/s.

Some of my norwegian friends and colleagues are reporting similar experiences in how crap Akamai is for them, both privately and professionally.

You hit the nail on the head!

Anyone who actually grew up in SF (such as myself), or has lived here a significant amount of time knows how much of a pathetic fail MUNI is. It has always been like this, except that when I was in high school/college it cost less than a dollar.

I live in the Sunet district, and it takes me longer on the N-Judah to get downtown, than it does if I were to live in Concord, and take Bart. The new N-Judah exrpess alleviates the issue to a small degree, but it is simly a hack fix, and is not viable in the long term.

As a current student and network admin of a small fiefdom at Stanford, I can tell you that the story is partially incorrect; Google is currently installing their fiber in the "faculty ghetto," a large Stanford-owned neighborhood by the school's foothills. They are not providing fiber to students - all student housing, academic buildings, and the campus core have separate mouthwateringly fast internet, Internet2, and wireless (via the SUNet).

More importantly, though, Google is *not* installing fiber in Palo Alto. One of the things that likely helped Stanford's case when we were selected is that the school owns *all the land* and even, as far as I know, all the utility lines on our campus. When you buy a house at Stanford, you actually only buy the building – you only lease the land. Because of that, when Stanford says "we're gonna install fiber," it's probably not tied up in regulatory messes, multiple contracts, competitive bidding, or the like. It takes the school's approval process, which may or may not be slow, but that's the only one; we don't have to ask the county, the city, or AT&T if we can do something - something that definitely speeds our adoption. I'm kinda scared that those kinds of facts might hurt further development of Google fiber.

"I was also surprised not to see Finland in the top 10 - it seems like every time there is a discussion of broadband access and speeds someone brings up Finland as a shining example of good broadband availability in a relatively sparsely populated nation; apparently, at least from Akamai's view of the 'net, availability (or at least uptake) isn't nearly as extensive as some have suggested" Akamai do not serve the nordic countries very well. No local servers, and I suspect that they deliberately cap bandwidth use from us too. On my connection, where I can get 11.5MB/s download speed from at 19:00 swedish time when downloading the EVE client from CCP(via LLNW), grab stuff from Sunet, Funet or similar at the same speeds too, or get 8MB/s or so from some other sources, if I have to hit an Akamai server, I'm glad if I get 1.5MB/s

Here in 'Merika, we believe in this thing called the "free market" which means companies are "free" to screw us over.

I lol'd :D

Our electricity market is free here in Sweden to. And I've got the understandard that really lowered prices, at first atleast.

Anyway around 50% of our electricity is hydro power and 50% is nuclear, but back in the days there was an election about when we should stop using nuclear power, not if, there was just three variations of "no" to nuclear power. Anyway, so I think it was decided that we should stop using it in 2010.

Well, last year was 2010, and I think they have closed down two reactors pretty recently, but the rest of the ones who's been around lately is still on. Or well, should be. Because there haven't been any real maintenance and upgrades lately since they where supposed to be shut down. So around 3.5 of our 10 are kinda turned off.

Two of our biggest players is Vattenfall and E-on, Vattenfall is owned by the government and have both hydrogen and nuclear power. The funny thing though is that because of all this the nuclear plants run like shit now and did last year aswell, so Vattenfall couldn't produce as much electricity (and maybe not E-on either?) as they would had done if things worked as they should. The expected fallout of that would probably had been that consumers would had to pay more since electricity was scarce and Vattenfall would had huge issues since they didn't got any money from the electricity they didn't produced and also had to spend money on fixing things up. Right?

But nah, since they are one of the biggest players of course they lost on the nuclear power and production, but the scarcity raised prices and since they also produce hydro power they could produce more of it and actually earn more than they would had done if everything worked ..

And also since it's owned by the government I think it would had been OK / cool if they just decided to keep profits steady or kept them low by selling the electricity cheap and hold prices down on the whole market for the benefit of all the consumers. But obviously they didn't.

Generally electricity cost is higher tens - lower forties öre (1/100 SEK) / kwh in the winter. Last year Swedish Kraftnät decided that it was a good idea to do maintenance on the electric grid close to one of the rivers producing hydro power which would then had led to no productivity there while they where working, in freaking February. They decided the day before that they shouldn't do it because it was cold and so on but that was too late so production was turned off anyhow and prices sky rocketed to 5 SEK / kwh .. (from like 0.4 the year before.)

Great.

I guess this is what you get when you've got idiots as politicians, let idiots vote and no-one takes responsibility.

1) Decide to turn of nuclear power.
2) Turn some off.
3) Ignore maintenance and upgrades.
4) Don't build any alternatives (a few freaking wind plants but who gives a shit? They just cost money.)
5) Wait.
6) Profit! / SNAFU

Maybe they could had thought ahead like 5-10 years and come to the conclusion that "wait, we don't have the electricity production capacity available that we can turn of the nuclear reactors in 5-10 years time. Maybe we should be running them for some time longer?"

And what good is the election anyway then there was no "Yes, I like to have nuclear power!"

Free market was also allowed to fix the ISP situation, when ADSL and cable became common. Back in 2002 at Nordic University Computer Club Conference /NUCCC there was this guy from Swedish University Networks / SUNET who talked about the cost of building fiber to everyone just as the electricity network connects to "everyone" for around 50 billion SEK. That never happened because politicians are retarded and I guess it wasn't the cool thing to invest in seen from the general public and old farts. So inst

We have star all the way out to individual apartments. Channels and such can be switched on and off per endpoint if our current cable operator would use that(They bought up the last one, which used to do that. No need for decoder boxes and cards, you just called, ordered a channel subscription, 5 minutes later at most you'd have it available).

So, in our stairwell, there's a pretty beefy tube running that holds the individual cables for each apartment, going down to the central switch cabinet in the house. Each house is individually connected via fiber to the next step up, which is about 10km away from here. And yes, we're running DOCSIS 3, with 100Mb/s down, 10Mb/s up available for now, 50Mb/s will become available.

Even during peak times, I get 5-6MB/s download rates from Sunet with our 50/10 connection.

Where I live, we have cable in a star topology, rather than ring/loop, and just in these 4 houses, there are 220 apartments. Yet I can still hit 5-6 MB/s during peak hours, on a 50Mb/s down connection, from a decent FTP like say Sunet.

Well, as you saw from my result the test server was actually in another country, Denmark to be exact, and quite a few hops away from me (20ms ping time).

But, there's something to be said for your question, certainly. There are two parts to it IMHO, the first is as you say, getting to/from your ISP at all. And the second is getting to/from the US (where many of the interesting endpoints are). In the first case, when it comes to cheap broadband, there is certainly caveat emptor. I could get bandwidth cheaper, but with worse peering, and that's why I chose Bahnhof (Sweden's first ISP, with a heavy dose of anonymity etc. thrown in. They get IP in a way that e.g. Telia-Sonera doesn't; and they're not that bad). Of course, it doesn't hurt that TPB is Swedish and that many Swedes have a nice uplink as well as a nice downlink... :-) (As a matter of fact the strong asymmetry in the US hurts Bittorrent more in my experience than does the overall bandwidth. I upload significantly more to the US than I can get back typically.

When it comes to getting across the pond, we are mostly all in the same boat. There is very little you can do if you're not on a special network (such as the Swedish University Network; SUNET). That's to say, there's less you can do, but not all ISPs are exactly equal. As a small aside, somewhat amusingly, peering/network structure to/from the US (esp. in Scandinavia/Holland etc.) is typically better than to the rest of Europe at large. Getting to a server in Austria can bring tears to your eyes. This has mostly to do with the crap state of network infrastructure in continental Europe in general.

Of course ISPs don't advertise peering agreements/status, so one has to go to e.g. netnod to check that for oneself.

As it happens I'm both on fibre at home and on SUNET at "work", so if you want to arrange a test, I'm game. Just tell me what to up/download to/from where and how, and I'll post (or email) the results. (You can reach me at "middlename" (really my first name) at lastname dot cx if you want to go offline).

Doesn't work for me with Bredbandsbolaget and if they lost all the three Bs (Bredbandsbolaget, Bahnhof, Bredband2) we're doomed!" :D

Should had been hosted around the universities instead =P. I wonder if Sunet would had given a shit?

(Heh, damn people up north stealing all the bandwidth! http://stats.sunet.se/top10.html)

Ok, so how come us Swedes mostly started to go online in the mid-90s (using regular old dialup), then everyone but the college students who were on 10 Mbps SUNET connections switched to ADSL (at that time mostly g.dmt) or DOCSIS connections, then "everyone" switched to ADSL2+ and the DOCSIS networks began to disappear (except ComHem who kept upping the speed of their network to stay competitive) and the whole time fiber connections have become more and more common (ten years ago 10/10 Mbps was the "standard" for what you could get with a fiber connection, today it's 100/100 Mbps with some people having access to 1000/1000 Mbps)?

By your reasoning we should all still be stuck on dialup or first-gen (g.dmt) ADSL. Especially when you consider our low population density...

Well, I live in Sollentuna(North of central Stockholm), and I pay SEK 155 per month for a 100/100 line and phone. And yes, I peak around 100 Mbit if I download from good sources, say SUNET.

Not quite $11, but close enough.

In Stockholm, and maybe in other Swedish cities too, when there is work where they have to open up the street, they also lay down fiber. This fiber is then owned by the city, and anyone can "rent" bandwith on the fiber. Community owned, but since anyone can use it there is competition where it counts.

Much of the Internet backbone in Sweden is run by SUNET, the Swedish university computer network, or owned by TeliaSonera with strict rules to allow fair usage.

I've been there, several years ago when I lived in a dorm room my biggest bottleneck when downloading things using DC++ was the speed of my hard drives (and the speed of the hard drive in the machine sending the data). But back then I had a connection to SUNET through the university...

/Mikael

A bunch of years back I had a 10/10 Mbps connection to SUNET with a static IP address and pesky firewalls/NAT to "protect" me from myself, I would quite often download and upload around 1 MB/s, and at 1 MB/s that's 86400 MB/day. Admittedly I rarely transferred anything quite that big, but it was definitely possible had I wanted to.

/Mikael

That still seems quite limited, although if you're in Australia then I guess it's better than what most people are using...

OTOH, I'm swedish so ADSL2+ Annex M 24/3 Mbps with no transfer limits for USD50 per month is considered kind of expensive. I'm still waiting for my 100/100 ethernet jack though, had it in college through SUNET but I don't live in a student apartment anymore.

/Mikael