FCC Chair Hints He’ll Open the Can of Worms
by Michael Cain
A couple of weeks ago I wrote a post about the Congressional Republicans’ network neutrality bill. Last week the chairman of the FCC wrote a column in Wired outlining what he will put in a proposed new rule for internet access service to be distributed to the Committee later this month. Since I believe that the internet is the fundamental change in communications that happened during my career in the field, and is likely to be the fundamental change in the field in my lifetime, I couldn’t resist writing a short critique of the outline. The short version of my conclusion on the Congressional bill was that it is good enough for now since it enforces network neutrality immediately and it doesn’t open the huge can of worms that full reclassification of internet access will. Short version of the chairman’s proposal: reclassification and an attempt to deal with some of the worms. Here are the main points, along with brief comments on each.
Strong network neutrality will be enforced. Chairman Wheeler’s proposal will ban paid prioritization and blocking or throttling of lawful content and services. This is absolutely the right thing to do, because internet innovation has always been done by people writing software that runs on devices connected to a content-neutral network. The correct answer to congestion in IP networks is always to provide more bandwidth, not to attempt negotiation of priority service across an unknown number of intermediate networks (some of which may not support priority anyway), and especially not to impose arbitrary traffic-shaping rules.
The new rules will apply to mobile as well as wired providers. A good choice, I think. Another strength of IP as a network protocol is that other than as a matter of purely local concern it’s independent of the underlying transport mechanism. We used to joke that it you didn’t mind the low speeds, it ought to be possible to implement IP carried on carpet static. If internet access is a communications service, then there should be a broad set of rules that apply to it that are independent of the underlying transport.
No last mile unbundling. This one is simply a matter of accepting reality. Mobile providers paid large sums of money for exclusive use of limited radio frequencies. Historical rules regarding the frequencies on which over-the-air and cable television must operate make upstream bandwidth in the cable companies’ fiber-coax networks a scarce commodity that is difficult to share. Google Fiber and municipalities will be less willing to build new distribution networks if they must make them available to other providers. Expect a hugeamount of difficulty in deciding what to do about legacy DSL over the phone company’s loops.
No rate regulation or tariffs. I’m torn about this one as an absolute. I would like to believe that market pressures will ensure that poor areas will get a reasonable minimum level of service at prices people can afford. The no-unbundling is going to limit the number of network providers, though, so I’m not entirely convinced that things will work out that way. One question raised by a no-tariff rule would be how to keep providers from changing service terms whenever they want: “I’m sorry, your plan has been discontinued, you’ll need the shiny new plan with different data caps and speeds and prices.”
Assuming that this is an accurate description of the rule the Chairman actually introduces, and that the Republican bill dies in either a Senate filibuster or by the President’s veto, what happens? First, I feel comfortable in predicting that net neutrality gets pushed at least two years into the future while the FCC goes through its process and multiple court cases are resolved. Possibly longer, since some of the plaintiffs have deep pockets and things will go all the way to the Supreme Court at least once. Second, I’ll predict that there will be at least one “gotcha” that comes out of the proceedings and court cases that no one is talking about. I’m speculating, but I would not be surprised if all of those clauses in current service contracts about “can’t operate servers” simply go away. As a communications service, I expect internet access to mean that I get a static public address good for as long as I continue to buy the service and that I can accept “incoming” connections as well as originate outgoing ones if I want to.
[About the image… The background is a partial map of the connectivity of the internet’s class C addresses, from Wikimedia Commons, courtesy of Matt Britt.]
First, one small comment before I get to my main point: Dynamic DNS is a thing now which is supported commonly by routers available for home. This allows you to associate a domain or subdomain with your local network and notify upstream DNS servers whenever your IP address is changed (by the upstream DHCP server, presumably). So while I agree that “no server” clauses will go away, I don’t think it follows that everyone gets a static IP address.
The larger comment is that my reading of the bill proposed by Senator Thune is that it is inadequate to address issues that happened last year. In particular, according to the testimony of Level 3 (see, for instance this and other entries from their blog), Verizon deliberately failed to upgrade their service for all customers in a region in order to put more pressure on Netflix to engage in paid peering. Verizon tried to pass this off as someone else’s problem, and indeed has been fairly successful in this regard. Of course, the “someone else” in this case was Level 3, and they aren’t happy about it.
In a marketplace where consumers 1) had many other options for broadband and 2) can correctly attribute bad internet service to their broadband provider, this would not require a regulator to intervene. I contend that neither (1) or (2) are true.
What Verizon did can’t really be construed as “prioritization”, it seems to me. I think common carrier status for broadband would address this kind of behavior. Therefore, I do not support Thune’s bill. It is inadequate to the situation. Common carrier status is pretty well understood, I have long wondered why broadband didn’t have common carrier status, though I would not describe myself as an expert on it.Report
Excellent comment.
What are dynamic DNS propagation times running these days? I might as well have had a static IP address for the last several years — I held the same dynamically-assigned one from Comcast until I changed out my cable modem recently.
Look, I also think internet access ought to be a regulated communications service. I thought so 20 years ago, but lost the internal argument so my company didn’t even try to make it so. Don’t know if Verizon’s behavior would be improper under the Thune bill; I think so, and complaints filed under the new regulatory authority the bill grants the FCC would at least make such behavior public; but you might be right. More importantly, though, the bill would put something in place within months. Reclassification is going to take years to iron out, since the courts are going to be heavily involved.Report
Once the world goes IPv6 (the day after controlled fusion is achieved), why not give everyone a static IP?Report
IIRC, IPv6 works out to allowing you to assign a static IP address to every separate square inch of the planet. It’s a sorta kinda big address space.Report
Lots more addresses than that — about 4.2e20 addresses per square inch of the Earth’s surface. Alternatively, every bacterium on the planet can be assigned 68 million addresses.
OTOH, as we move to the Internet of Things (which was a really big deal at this year’s CES), there are a lot of my things that I don’t want to have a public address. Or at least don’t want that address to be reachable from the public internet.Report
Hey, I was only off by twenty orders of magnitude. I knew the address space was ridiculous, I just forgot exactly HOW ridiculous.
I think the reason for the slow adoption is simply that NAT works just fine for most people. I’ve got a private address space that can accommodate up to 256 devices and I use maybe ten of those between our phones, tablets, and entertainment devices. NAT isn’t explicitly a security measure but it helps and I have no desire to have all those devices exposed to the wild and woolly. If someone’s going to hack me they’re at least going to have to work at it a little and I’m not worth the effort.
It’s really only servers that need a public static address.Report
It’s really only servers that need a public static address.
I don’t go that far, or at least, I have a broader definition of “server” than most people. There will be some devices in your home that you want remote access to — a DVR, for example, or a thermostat. But you want access to be limited, and you want reasonable security, and… Somewhere here I have a design patent with my name on it for an arrangement that did a reasonable job of those things.Report
Can’t you handle that worth port forwarding and the like?Report
Can’t you handle that worth port forwarding and the like?
To some extent, certainly. Things get bothersome if several devices all want to use port 80, and you have to remember which external ports map to which devices. You also have to depend on the individual devices’ security arrangements — not as big a problem now as it used to be, since devices are typically equipped with a lot more processing horsepower. The system in the patent allowed devices to register with a gateway device which would, among other things, allow a single point of contact for reaching all of your devices and provide robust security. All of the needed registration, code installation, etc, was done automatically.Report
Thanks for these posts. I’ve gotten a lot out of them.Report
Seconded.Report
Very much so.Report
“The correct answer to congestion in IP networks is always to provide more bandwidth,”
This has been trivially easy since fiber optic has been replacing copper, and a moot point for a while after everything was way overbuilt in the Dot-com era (e.g. Global Crossing), but are we finally now reaching a point where ‘just provide more bandwidth’ becomes as vexing a question as ‘just build a bigger interstate’?Report
Only if you don’t understand what’s going on. The USA has been lagging far behind many other countries considered “less developed”, due mostly to existing monopoly ISPs who prefer to price-gouge and engage in noncompete agreements for different geographical areas and who have failed to properly upgrade service for the vast majority of those who pay for their service.
Verizon and Comcast have both been caught deliberately blockading traffic from Netflix in order to try to extort payments from the company, even after subscribers paid for a certain speed. It’s not a lack of bandwidth on their side or technical inability to upgrade, it’s a deliberate choice to interfere with service in violation of ethics and the contract with their customers.Report
+1.
Seriously, we are SO far behind in bandwidth, it’s ridiculous. I’m not talking the rural boonies — I’m talking major cities and the suburbs of major cities. We pay FAR more for FAR less than, well, everyone else.Report
Not sure about “everyone else”. I always hear that cell phone and Internet prices in Canada higher than those in the US due to a highly oligopolistic market (you can buy from Telus, Bell, or Rogers, and that’s it, and they all have high prices and terrible customer service).Report
Australia’s also pretty crappy from what I know – OK, maybe it’s just the former colonies of the UK that were formed by mostly white people that suck at having fast Internet.Report
I exaggerate, of course. Sometimes the answer is IP multicast (there’s a reason one-sixteenth of the entire IPv4 address space was set aside for multicast; I’d love to see one result of this rule making be a multicast requirement). Sometimes the answer is clever caching of popular content in locations where the pipes are already fat.
That said, the research labs continue to set new records for bits/second on individual fibers. New fiber continues to be installed. New xDSL technologies push more and more bits/second on copper loops. Recently Comcast sent me a new cable modem gratis because if all the modems run the DOCSIS 3.0 protocol, they can cram a good many more bits onto the coax. The IEEE continues to issue new 802.11 specs that jam many more bits into the available RF spectrum. Google Fiber (slowly) builds overlay networks. There’s lots of room to expand the capacity of the pipes.Report
https://www.techdirt.com/articles/20060131/2021240.shtml
We’ve already paid for tons of fiber that wasn’t installed.
Not optimistic about getting fiber out of the city anytime soon.
(we JUST got fiber. 3 years late).Report
Thank you for the update. (I hope you’ll continue with these posts as this progresses, too.)
Expect a hugeamount of difficulty in deciding what to do about legacy DSL over the phone company’s loops.
Can you explain this more, Michael?
/asking because I have DSL of courseReport
Local loop unbundling — providing alternate providers with access to the “last mile” distribution network — in the US has only happened on the telco copper pairs, and has changed in lots of ways over the years. Sometimes it’s been the physical copper, sometimes it’s been the xDSL frequencies only, sometimes it’s been the bitstream at the central office, sometimes it’s been mandatory, sometimes it’s not. If your DSL provider is not the local phone company, then you’re enjoying competition as a result of LLU. Will the Chairman’s new rule say that no-one has to unbundle? Will it say that no one is allowed to unbundle?
There are many folks who claim that the reason that Europe has faster and cheaper internet access than the US is because the regulators there generally require LLU, and the resulting competition amongst ISPs requires them to deploy the latest xDSL technology and hold prices down in order to win customers.Report
Ha. No, I have DSL from the local phone company, and have a choice of that or cable.
(I love my local phone company; began when a coffin manufacturer ran a phone line from his water-powered factory to the nearest train station, about 6 miles away. That first town, Woodstock, ME, was also the last town in the US to use the old crank phones. Now, they’ve moved into the secure facilities at an airforce base closed through brac, and they’ve installed broadband service to areas throughout my county, one of the poorest and most rural, and are moving on through other parts of the state; plus secure business stuff, using equipment at a military base closed through brac. Since finance happens here, there’s apparently some tax advantage for banks to do data processing in the state, there’s a big market for highly-secure business stuff; and our teensy-tiny phone company, that began at a water-powered mill that manufactured coffins, has grown to fill the needs.)Report
I find it interesting to see how the regulatory mindset of those looking at the Internet has gone from “well, it’s like a phone” to “no, it’s more like cable TV” to “no, it’s really more like a phone but more than that it’s kind of it’s own thing.”Report
It was easier for them to think of it as a phone when everyone was using a modem and dialing an AOL phone bank.Report
I worked for a telco at the time, and lord, was dial-up data service a nightmare. The phone switches were designed and provisioned assuming certain traffic characteristics. Peak-hour traffic was assumed to be, per line, three call attempts, three-minute average call duration. Data calls with hours-long holding times were killers. Bell Labs was looking at what eventually turned into DSL by about 1981 to try to shift data service off of the voice switches.Report
You’re confusing access and content. First it was clearly access — you dialed a connection to your ISP, or to one of several if you had multiple accounts. Then came always-on [1] and you had a dedicated pipe into your ISP. At that point, the industry was in a “content is king” mode of thought, and the FCC decided that the service should be regulated that way. After 20-odd years, everyone has figured out what some of us thought was obvious from the beginning, that people would pay for access alone because there was so much free content. And here we are, trying to undo a 20-year-old mistake.
[1] USWest, the only company that was both a Baby Bell in its own region and a cable operator outside that region, had ethnographers on staff that did several studies of how people were affected by cable modem service. The results were summarized in some of the early brochures: “High speed is why you try it. Always-on is why you move the computer out of the back room into the kitchen.”Report
No last mile unbundling.
IMHO, every major issue we have in this space can be summed up in two words, “vertical integration.” That and “natural monopoly” are precisely why everyone hates their cable company and can’t do a damn thing about it.
Except under unusual and temporary circumstances any company supplying a service via a fixed distributed infrastructure is going to end up as a monopoly. The only exception may be a very customer-dense environment like an inner city.
Any company that is simultaneously welling you access and content is going to have incentives to screw with the competition in the content space. Comcast and their ilk despise that Netflix, Hulu, and Skype even exist and have zero reasons to deal with them in a neutral fashion.
So if I understand what you mean by no unbundling the last mile we’re doomed to forever fighting the same monopolistic/oligopolistic issues. World without end, amen.Report
Comcast competes with Verizon’s fiber network on cable…
(since I don’t do cable, I can’t tell you about rates).Report
“competes” isn’t really accurate. Here’s a FIOS map.
http://fiberforall.org/fios-map/
Additional note: Years ago I checked to see if FIOS was avail at my address. According to the Verizon map, it was when I put in my zip code. When I put in my actual address, it wasn’t as I was too far from the closes node.
Comcast really has very little competition in most of the state where I live. You want to address the cost of internet service, have the Public Utility commission actual have competing vendors.Report