2 ms·
Yeah, there are always going to be pricing shenanigans because people are not willing to pay the actual costs. The actual costs of an ISP are the $1MM one time
by jrockway 7y ago
Yeah, there are always going to be pricing shenanigans because people are not willing to pay the actual costs.
The actual costs of an ISP are the $1MM one time cost to dig up the streets in your neighborhood, and an ongoing cost of electricity for their datacenter, Juniper licenses, peering, customer support, etc. The problems the ISPs have is that no customer is going to pay that one time cost, so they have to figure out some way to make you pay it without that $1MM bill the first month.
I have worked for two ISPs and the internal debate is always around figuring out how to do this. At the last place I worked, our install costs were relatively low; the city already built tunnels under the streets for fiber, so it is financially viable for someone to just pay the ISP to run the fiber, and then pay the incremental costs. It is certainly not free (everyone wants their cut, including the building that you're already paying rent to), but is the kind of money that a medium-sized business or a condo co-op could find if they were so inclined. Individuals have no chance of finding the money, though, so something has to give.
(Where I'm going with this, is that paying per bit does seem reasonable to me. It's what AWS does; if you want to lower your bill, send fewer bits. But it's still a bit of a hack, not reflective of the actual costs. The first bit you send over your line costs a million dollars, but the rest are practically free. You're just paying to keep some routers powered on, and for them to be upgraded when other customers want to send more bits.)
Being able to sell your browsing history is one way to cover the infrastructure costs, but it's unfortunately not a workable idea because as we've seen, companies like Google and Cloudflare can just obscure that data from the ISP.
I hate to be all pro-government, but can you imagine how inefficient it would be if UPS, Fedex, and the USPS all had to build nationwide road networks in order to be able to deliver packages to your house? Instead, the government built the roads, and it made the economy stronger. Being able to drive to work or have someone send you a package from across the country is quite valuable. I don't see any reason why the last-mile Internet shouldn't be built that way; the city can get traffic to a peering point, and from there, you can pick what ISP you want to connect you to the actual Internet. But unfortunately, we half-assed that, instead paying private companies to built the last mile, and now they want to milk their nearly-free infrastructure. The idea of Internet to the home is a mature one, and the reality of a mature business is that it becomes a commodity, and that is going to drive profits down. Comcast, Spectrum, etc. should all be finding new ways to make money, and the easy one, "spy on our customers", is not viable. Sorry, investors. Too bad for you.
- novok 7y agoIsn't that $1MM upfront cost one of the major reasons why bonds exist? You get your $1MM bond now, and you repay with the future revenues you get from your customers. You make the calculation and that debt payment, plus ongoing costs, plus profit is what you charge your customers. Eventually you 'pay off the mortgage' and then revenue really goes up, but with the rate of technology upgrades, you will always be 'paying your mortgage'.
- jrockway 7y agoYou still eat the risk, though. Maybe 5G is a thing and all your customers cancel next year. Now you have a 30 year payment on something that generates no revenue. That's where the fear is.
- novok 7y agoIsn't that business in a nutshell? You wouldn't have a much of a business if everyone went to 5G bond or not. Competition and doing risky things is part of business and partly why bankruptcy law exists. If your business implodes due to 5G, the bond is written off as part of bankruptcy and the lender eats the loss while taking over the business assets. Also seeing how low range 5G is, it basically looks like cell phone providers become wifi access point providers, which smells really close to what a wired ISP does anyway, which sounds like another ISP competitor.
- jrockway 7y agoMy thought was that the AT&Ts of the world could easily build their fiber networks for 5G backhaul. My former employer's thought was that the AT&Ts of the world would pay them to do that, though. I don't know who's right. The biggest problem I've seen at ISPs is the general unreliability of all the ISP-grade networking equipment. The routers are bad, the OLTs are bad, the VoIP phones are bad, the CPE is bad, the WiFi routers are bad, and the WiFi clients are bad. If I were an ISP, I'd be working on fixing all of that; everything that interferes with your customer wanting to send a packet and the host on the Internet receiving that... that is the area to optimize. But I don't think the funding exists. (It did when I was at Google Fiber, I worked on CPE. But Google lost interest, and at the smaller ISPs, it's all about buying as much crap as you can get off the shelf and letting your customers integrate it.)