Aussie Network Engineers Form Members-Only ISP 66
schliz writes "A group of Australian network engineers is planning to launch a not-for-profit internet service provider that will provide access to the nation's high-speed NBN fibre network for like-minded people. The cooperative, dubbed 'No ISP,' has no staff or add-on services to keep costs down. Members will be able to 'trade' excess download quota for a market-based price, depending on supply and demand."
I'd sign up in a second if I lived in .au (Score:4, Interesting)
All I want is reliable bandwidth and latency (what good is 50ms latency if it spikes to 1-2 seconds every so often? say good bye to skype and any online gaming) and ideally a static IP.
$110 per month for a terabyte plan on 100 Mbps down and 40 Mbps speeds over the fibre network
Is actually better than what I currently pay, I get 100 meg down, 5 up, 250 gig cap for $90 a month (Canadian duopoly, wheee). I hope they write up how they accomplish all this, might be time to start more of these co-ops. I also love the fact that with the trading scheme they encourage people to use the bandwidth, but intelligently. Right now since there's no real advantage of disadvantage to me when I run major downloads during prime time (and I notice that my speeds/latency are quite a bit worse during prime time), this co-op would result in me scripting most downloads it to run when bandwidth is "cheaper" (aka 3am). I suspect this is true for many other heavy users.
It's just not economically feasible (Score:3)
The project may be technically feasible, but economically I do not think it makes sense
I hope the Auzzies do a thorough feasibility study before launching it, or someone would end up paying a lot for nothing
I am not discouraging them from carrying it out, I'm just being practical
Re:It's just not economically feasible (Score:5, Interesting)
You must be kidding. Providing internet service is insanely high margin once the lines are in place. I seriously doubt they are planning to drop new fibre or any other high capital infrastructure investments. I'm sure they will be riding on mostly existing infrastructure, especially the last mile.
These are a bunch of network techs, so they do have some clue what the costs will be.
Not only that, but I'm sure for some of them, they will actually make money on the side from the service. Have a trunk link down, oh well, me and my buddy Jim will fix that right up at our normal contractor rate (without our boss taking a cut!). Not only that, but I am sure they will be buying a lot of second hand hardware that a corporation would never buy just because it is used and donating spares from their various work projects. Their infrastructure will be incredibly cheap. Not to mention, that they will be able to crowdsource a lot of the design and implementation. Usually, I consider crowdsourcing a way of making a crap situation crappier, but it works great when you are working with a group of experts in their field.
What's that? The boss wants to upgrade the ASA 5500's at the South East branch, awesome. Hey guys, I just got a new gateway cluster for the secondary trunk line.
Networking hardware is incredibly cheap if you take the time to shop around. Unfortunately, 9 out of 10 businesses just grab it at over inflated prices from CDW or whatever other rip-off artist they usually deal with.
There are two long term challenges which they will have to face that could come back to destroy everything in about 8-15 years. First is any kind of crowdsourcing turning into politics or committees. Second is letting too many people do their own thing with the network with no documentation or oversight and creating a horrific mess waiting to collapse. Strong, well considered leadership solves both of these issues very easily.
Taco, I respect you, but I think your idea of practical is not the same as the direction these guys are heading.
Re: (Score:2, Funny)
me and my mate Jim
fixedthatforyou.com.au
Re: (Score:1)
Except they're doing none of that and buy all services wholesale instead. They're not even leasing lines. They're just acting as a buying syndicate.
Since they're network engineers, they probably know that it's a cutthroat business. The profit margins are thin to negative. Networks are not made of routers, networks are made of buried fiber. Suppose you have 100 miles of fiber in the ground, which is nothing really, and that fiber has an average lifetime of 50 years (due to technological changes, construction
Re: (Score:2)
The vast majority of ISPs in Australia work on the VISP model, reselling various components of other ISPs' infrastructure. I work for one myself - we offer cloud, rackspace and WAN products tying in well with our traditional integration and support services.
The margins are very good as long as you know who to talk to. Buying transit or backhaul capacity off one supplier may mean you can get local tails at far below base cost, if they're willing to take the hit for the volume of services you can push through
Re: (Score:3, Insightful)
You must be kidding. Providing internet service is insanely high margin once the lines are in place. I seriously doubt they are planning to drop new fibre or any other high capital infrastructure investments. I'm sure they will be riding on mostly existing infrastructure, especially the last mile.
I think this is where a lot of European localities get it right: Layer 1 is owned by the government (municipal, province/state, etc.) or a neutral/non-profit third-party. The competition occurs at Layer 3 of the stack.
Re: (Score:1)
Providing internet service is insanely high margin once the lines are in place
Meaning, we're handing a lot of welfare^W"capital-gains" queens billions per year for nothing. They're providing a pittance of a service, far less for our money than what we get for taxes, and priding themselves on their "success".
Re: (Score:1)
Re: (Score:3)
It's perfectly feasible. It sounds like they're relying on the fact that at wholesale, you're paying for CIR not bits. So they buy x CIR, then distribute that among the members, and operate some sort of trading program to let the members buy bits from each other. Interesting concept, really.
Re: (Score:1)
Where do you live then, .com?
Re: (Score:3, Insightful)
Just wondering though, what would one really do with 100Mbps, analyze all LHC data ?
p0rn. Terabytes and Terabytes of pr0n.
Re: (Score:1)
>Just wondering though, what would one really do with 100Mbps, analyze all LHC data ?
>The slowest/cheapest internet over here is around 8Mbit, which is more than plenty for what
>I actually use...
Just try 100 mbits and you will see. Suddenly you start thinking that 5 minutes to download a linux dvd iso is horribly slow.
Re: (Score:3, Interesting)
I'm really curious about how they solved the last mile problem. I've thought about starting a non-profit ISP in Canada, but the real questions is how to avoid paying Bell their pound of flesh. Until that problem is solved and you've got ISP-agnostic fibre being laid, it means that you're charging what Bell is charging.
Re:I'd sign up in a second if I lived in .au (Score:5, Informative)
Because I can see you live under a rock: http://nbnco.com.au/ [nbnco.com.au]
Most people will have 1Gbps capable fibre in Australia. The rest get fixed wireless (LTE), & very remote areas get satellite, all for the same wholesale price.
It's actually cheaper to run FTTH these days with the polymer cables (what the NBN is using) than to run FTTN networks that rely on crumbling copper.
I've seen this one before (Score:5, Informative)
APANA [apana.org.au] part two?
No (Score:1)
What? (Score:2)
No ISP will not always compete successfully with ISP mainstays on price
An equivalent plan can be purchased for $10 less a month from iiNet, due mainly to the larger ISP’s existing scale and direct connection to the NBN.
That really defeats the purpose !!!
I get the peripheral benefits so sharing quota but most plans have shaping and I'm sure how well sharing would work ... doesn't seem all that compelling
Re:What? (Score:4, Informative)
Except you're actually getting what you pay for. They plan to buy a terabyte of bandwidth wholesale for each user, unlike traditional ISPs, which typically use traffic modeling to oversell their infrastructure. With the new model, if you're under your cap and I'm over my cap, I can buy the extra from you at cost, and keep getting full speed service instead of getting capped at 1MBps.
no staff? (Score:4, Interesting)
The cooperative, dubbed 'No ISP,' has no staff
I know not all countries have an established history of co-ops, but that's a completely disingenuous way of representing them.
Everyone may own an equal share. Some people may only contribute money (e.g. UK's Co-op model). Or it may be that only those who contribute work may be members, as in a partnership (e.g. John Lewis model). But as long as work needs to be done, there are staff.
Since these have been some of the most stable businesses in the UK through this recession, it's worth it for anyone who is intending to start up a new one to represent it properly.
Re:no staff? (Score:4, Interesting)
You give examples of several different co-op models yourself. Why couldn't there be another co-op model for this "No ISP"?
It should be perfectly possible for them to automate ~99% of daily operations and only occasionally require human intervention, so when they say "no staff", perhaps they mean that all work can be done in a few hours a week from home.
Re: (Score:2)
Except that if they're purchasing wholesale and operating no equipment themselves, then noone needs to be trained as a NOC administrator because there is no network operations.
Re: (Score:1)
Then there are people doing accountancy, marketing, etc.
Among the privatised, competitive (lol) energy and telecommunications retail providers of the UK, many companies are nothing more than resellers and bill-printers. It may require little work, but it doesn't require none.
Re: (Score:2)
Not necessarily. Staff have, by definition, some responsibilities within a business. If there's no ISP, and the co-op members are all highly competent, they can just use the network resources any way they like, with no actual responsibilities other than "be considerate".
Eg, if someone wants something done, they'll do it. If they feel like it. Or maybe not. Whatever.
Re: (Score:1)
OK, maybe it's just an issue of definition of "staff" and perhaps I'm making too big a deal of the "no staff" thing.
A co-op may pay its workers in produce/service rather than $. Those who contribute to a co-op to benefit from its production are certainly not "volunteers". I think the distinction is important because people hear words like "co-op" and think hippie commune, when in fact it's an equitable and often very stable business structure.
Re: (Score:1)
I don't think they get anything of monetary value in return for their work, so they really are unpaid in any sense of the word. There is a legal entity (the non-profit cooperative) which acts as the buyer and reseller of services, but any work that needs to be done to make that entity function is currently handled by people who just want to make it happen. You know, like when you order a bunch of items for yourself and your friends and then split the shipping evenly, even though you handled everything. Some
I had this idea ages ago (Score:2, Interesting)
No proof I did, and so what, I may have thought of the idea but these guys are actively pursuing it.
Support costs are HUGE burden for the ISP, anything to save them money and bring the costs down. and let's face it most of our DSL connections are reliable anyway.
My idea was slightly different though, I proposed that one of the peering exchanges (like WAIX or PIPE) allow DSL straight into a RAS on there network, and then you have a backdoor to your own equipment as an ISP. This will then allow you to reach y
Re: (Score:1)
How about removing the faux caps? (Score:4, Insightful)
I suspect what essentially amounts to price fixing is going on with the ISP's in Oz.
The cost to carry data is no where near the profit earned by instituting the caps and the charges for going over. Surely in a natural market the price would be lower, reflecting the cost and allowing for a profit, but not an obscene profit.
For a 1st world country having the caps they do is pretty pathetic. There is little excuse for it, and "people in the know" should do something about it.
Re: (Score:1)
There is little excuse for it, and "people in the know" should do something about it.
We're going to need a year or so to switch out the current "leadership" [tumblr.com]...
Also: +1 your sig.
Re: (Score:1)
Guess you missed how the Liberals' plan for the internet is "kill it all and let God sort it out".
Re: (Score:2)
Caps? Charges going over?
Some ISPs offer unlimited internet for $30/month providing you lease your landline from them as well at $30/month. And before anyone who doesn't know how this system works, would you rather spend $60/month on landline + unlimited internet from one company, or spend $20/month on your landline with one company and $60/month for 500GB datacap with another?
Yes the caps are bullshit, but at least they are going down. I still remember the bad old days of a 3GB cap (the worlds first cap I
Re:How about removing the faux caps? (Score:5, Informative)
>There is little excuse for it
Yeah there is, most traffic is to American sites and we have a limited international cable infrastructure which mostly relies on 2nd and 3rd parties in the link. On my last ISP I had connection problems on occasion because either SingTel or the US ISP at the landing in America would fuck with some settings. With the NBN I believe most of the international links and peering will be handled by NBNCo which should have more bargaining power then the small ISP's currently do (the big ISP's in Australia generally prefer fucking customers over).
Re: (Score:2)
The limited cable structure thing is bullshit. We have several high speed links to asia, which get data from the us without issue.
I hear that argument all the time and...it's just bollocks.
Re: (Score:2)
Umm no. All NBNCo do is provide a connection from you to your chosen RSP (reatil service provider), nothing more. Your RSP is the one that needs to provision the bandwidth from the PoI you're connected to, to the Internet.
Re:How about removing the faux caps? (Score:4, Funny)
For a 1st world country having the caps they do is pretty pathetic. There is little excuse for it, and "people in the know" should do something about it.
I live in Australia. My current plan is $50/month for 30Mb/1Mb cable with a 50GB cap. I'm not sure what other people do with their internet, but even watching unlimited porn I've only ever gone over this cap once in the many years I've had it, and this was when I downloaded a ton of movies I never watched while watching a ton of porn. I really don't get why the average Joe needs more than this?
Re: (Score:2)
Re: (Score:2)
Are you sure that their profit model isn't necessitated by having to lay thousands of miles of fiber just to reach about 15 million subscribers? If you're paying $100k/mile and you're trying to connect a city of 3 million (melbourne to perth?) about 2,000 miles away, you're paying 2 billion for a cable. If you're paying 5% interest on the cost of laying that cable you're paying 100 million a year in interest JUST ON THE CABLE. This doesn't include all of the fiber you need to lay for the last mile, equipmen
Re: (Score:2)
Sorry, I made an error. According to deutsche telecom the ISPs nowadays are only getting 100gbps per chanel which would equate to 6 terrabits per cable.
http://www.ispreview.co.uk/index.php/2012/03/deutsche-telekom-break-record-with-512gbps-data-speed-via-fibre-optic-cable.html [ispreview.co.uk]
Which cuts my example bandwidth in 1/4th and only give 1.5mbps per line to 4 million people. You can see why the costs of laying cable to Australia's cities with their low population and farflung cities would be challenging.
Re: (Score:2)
That argument might apply for more regional areas but it shouldn't hold for Sydney or Melbourne. Yet it does.
Re: (Score:2)
Also forgot to carry the 0. 200 million for that cable. I fail at maths. That would be the cost for 10 cables. Sorry.
Re: (Score:2)
Now, lets look at the real costs with your corrections. $200M for the cable (the cost/mile is almost entirely the digging and right-of-way administration, etc, laying 10 cables costs barely more than laying 1). That will be then $10M/year in interest. That means if you're serving 4M people at say $50/month, you pay for the cable in the FIRST MONTH. Or, you can deduct $5/month/customer to pay off the fiber in a year with interest.
That kinda blows the whole argument up.
At 6Tbps/cable, you're looking at 60Tbps
FDN (Score:3, Informative)
In France, the oldest provider still in activity (FDN = French Data Network) is an non-profit association (and now an association of association). FDN was created in 1992. Maybe they can learn a lot from this long experience in assiociative service providing and they can probably collaborate to have something like a cross-borders ISP.
http://www.fdn.fr/
http://fr.wikipedia.org/wiki/FDN
ISP Club for Men (Score:1)
Pipe dream? (Score:1)
There are already established players in the market, both the 'Premium' ISP iiNet and the 'Low cost' ISP Exetel both offer better priced plans.
"No ISP" (which really is an ISP because they will still have to handle billing, backhaul network & so on) are more expensive and without any of the extras or support of the other players.
As a tech who would be interested in helping my clients get on board with NBN (as I already do with ADSL), I would much rather resell an established player with full support tha
There is on in Boston, too (Score:1)
The Cambridge Bandwidth Consortium http://www.cambridge.bandwidth-consortium.us/ is similar and has a fairly long history.
Comment removed (Score:3)
Boston's had one for 15 years (Score:3)
yes, since the late 90s, there has been a non-profit ISP for 'people with clue' in the Cambridge/Boston area. Things went well for a while, when we were able to get T1 circuits at a discount, colocation, peering, various telco services. But as T1s die, much of the value proposition is dying as well. Doing more VPN stuff these days. But we have way more IP space than we need. We're just a dozen low key people.
This is an iteration of an common thing... (Score:1)
I worked for UK ISP Demon in the early 90s, had a 10Mb "baseband" connection to my bedroom direct from the NOC, and I was developing business, systems and network software for the corporate part of Demon. I gave various customers and friends at other ISPs a login on my home Linux box so that they would be able to do traceroutes and so on. Since .org domain is "kfs", my one requirement for anyone wanting a web directory was "come up with a name that uses the kfs initials" - e.g. the "Kite Fliers Site".
It bec
Re: (Score:2)