News for hackers. Stuff that matters.
Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by grandalf on Friday October 13, 2017 @10:16PM (#15470323)
Mobile carrier consolidation has been one of the major failures of antitrust regulators. Looking back to the Sprint/Nextel merger, plans and pricing both declined in customer value severely in the months after the merger closed, with in-plan minutes going from the thousands down to the hundreds for the same price.

I would strongly prefer that any consolidation such as this be allowed only if one of the firms return all of its spectrum licenses to the FCC to be re-auctioned. It's fine to merge to consolidate infrastructure, but the unfair aspect of market power is derived from the natural scarcity of spectrum.

The hardware cost of launching a mobile carrier in an urban environment is falling rapidly, so only spectrum scarcity is preventing upstarts from disrupting the most lucrative urban markets.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by jessaustin on Friday October 13, 2017 @10:33PM (#15470369)
If competition or "customer value" were actually a goal of FCC, they would open more unlicensed space, on more useful frequencies, rather than auctioning, or re-auctioning more rigidly regulated spectrum. This "natural scarcity" is a property of the deployed technology, not of the universe.

[EDIT:] To clarify, the problem isn't so much the portion of the spectrum designated for mobile phone use, as the terms under which that portion is so designated. More of that portion should be available on an unlicensed basis, or at the very least on a non-exclusive licensed basis. The tiny fractions of the spectrum that have been available on those terms in recent history have seen more innovation and less artificial scarcity than more strictly licensed frequencies have.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by grandalf on Friday October 13, 2017 @10:36PM (#15470379)
> would open more unlicensed space, on more useful frequencies

Which spectrum do you think should be reallocated to mobile use?

> This "natural scarcity" is a property of the deployed technology, not of the universe.

What frequencies do you think should be used for mobile voice and data that are not already?

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by wahern on Saturday October 14, 2017 @12:09AM (#15470703)
Required reading:

  https://spectrum.ieee.org/telecom/wireless/the-end-of-spectrum-scarcity
  https://www.salon.com/2003/03/12/spectrum/
Basically, the argument is that interference (as understood in spectrum policy) is a myth. The channel capacity of a frequency band is a limit between two distinct physical points--it's not some global (or even local) limit. If you put two pairs of broadcasters and receivers in the same room using the same band, the laws of nature don't require that the available capacity to each pair be cut in half. It's the same as it was before for each pair; any reduction in throughput is merely a limitation of their particular technology. As sensitivity improves; as algorithms improve; as processing power increases; you can asymptotically approach your original (uncontended) throughput even as more transmitters join the fray.

The above articles are fairly old, but their arguments have born out, and the fundamental science isn't contested. MIMO, beamforming, software defined radio, and other approaches are now widespread and have permitted orders of magnitude more _aggregate_ throughput for the same geographical area, even as the number of transmitters has dramatically increased. And things will only continue to improve for the foreseeable future.

It's not literally true that interference is a myth. And there are fundamental physical limitations that we'll eventually bump up against. But those fundamental limits are beyond the horizon. In the context of historic and contemporary spectrum policy, interference is indeed a myth.

The above technologies have come to fruition slower than free spectrum proponents argued. But their argument has always been that as long as government treats spectrum as scarce and grants legally protected monopoly rights, there's less market incentive to develop those technologies.

The fundamental argument isn't that we need to add more bands to mobile, but that we need to remove (or at least loosen) the regulations which artificially constrain contention. If market participants had to deal with more band contention, there'd be more incentive to improve their technology to restore the service quality. As long as you don't have malicious people intentionally trying to disrupt the system, the argument is that market incentives alone would spur technological improvements at least as fast as contended usage grew.

Of course, for the deployed technology at any particular point in time, more spectrum is better. But the argument here is about optimal long-term policy, not about the best way to ensure first responders can communicate 6 months from now.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by rayiner on Saturday October 14, 2017 @12:09PM (#15472761)
> Basically, the argument is that interference (as understood in spectrum policy) is a myth.

The FCC deals with the world as it is, not with theory. When over the air television was designed, you couldn’t put the computational power into a TV to do true kind of interference rejection a modern WiFi system can do. Heck, even a more modern system like HDTV has miserable interference rejection. Transmit something else on the same frequency as a TV channel and see how much of a myth interference is.

You’re right of course that the FCC’s model of splitting uses by frequency band treats the issue as one dimensional, while in reality there are opportunities for spatial and temporal reuse. But you need computational horsepower to take advantage of that, and the necessary computational horsepower is really just now becoming available.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by grandalf on Saturday October 14, 2017 @12:58AM (#15470854)
This is an interesting argument. I'm aware of it but wasn't aware that the other person was referring to it in his post.

I certainly think it would be interesting to have the FCC allocate a band that could be used in a less regulated way such as you describe.

It's a very interesting area of SDR and game theory and I've actually put a bit of thought into how such an ecosystem might work after hearing a talk about the DARPA Spectrum Challenge.

So I don't disagree with your point at all, and think that the FCC should allocate some spectrum to this sort of deregulated scenario. Perhaps if it also required that all protocol and modulation approaches be open sourced that would help participants to coordinate usefully.

I think that we are far from the limits of physics across most of the network, but there are likely some congestion patterns in dense areas that create complex intermod scenarios that would make a system like this somewhat brittle until the ecosystem evolved to find a cooperative equilibrium of some sort.

In addition to MIMO, I'm interested in things like using TDM and awareness of harmonic and intermod characteristics to create various emergent latency "contracts" that correlated with different frequency ranges and time slices, so that participants that did not need low latency could create significantly less interference than they otherwise would, and participants could cooperate based on the data rate and latency requirements of other nearby nodes.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by NotSammyHagar on Saturday October 14, 2017 @01:29AM (#15470929)
why would it be different than wifi's 2.4 and 5 ghz bands? those are open, but don't penetrate far because they are high freq. it works but there's tons of things making it unreliable.
Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by rayiner on Saturday October 14, 2017 @12:11PM (#15472775)
WiFi works in part because it doesn’t carry very far. The unclicensed band strictly limits output power, which limits range which limits the potential for interference.
Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by raverbashing on Saturday October 14, 2017 @03:24AM (#15471166)
https://en.wikipedia.org/wiki/Channel_capacity#Channel_capac...

Technology may help, and the technological effects mentioned in the articles do happen but there is a fundamental limit

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by throwaway613834 on Saturday October 14, 2017 @04:40AM (#15471341)
>> It's not literally true that interference is a myth. And there are fundamental physical limitations that we'll eventually bump up against. But those fundamental limits are beyond the horizon. In the context of historic and contemporary spectrum policy, interference is indeed a myth.

> Technology may help, and the technological effects mentioned in the articles do happen but there is a fundamental limit

Parent mentioned exactly that in their comment.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by raverbashing on Saturday October 14, 2017 @07:00AM (#15471645)
> But those fundamental limits are beyond the horizon.

And that is the part I disagree

Especially when you have a large amount of people in the same place (stadium, busy street, etc) Or even your office wifi

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by user5994461 on Saturday October 14, 2017 @11:31AM (#15472593)
Parent is wrong.

There is a fundamental limit and we are at it in many scenarios and we've been for a very long time.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by grandalf on Monday October 16, 2017 @04:14PM (#15486294)
Just curious how one would define that limit. Would it be correct to define it as:

Bits of data delivered at or above minimum tolerable latency per hz of RF bandwidth per second?

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by Aloha on Saturday October 14, 2017 @01:12AM (#15470889)
There is a rational talking point relating to a higher noise floors causing smaller range and higher bit error rates. For CDMA network P/N solution is a real thing.
Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by rayiner on Saturday October 14, 2017 @11:58AM (#15472704)
The FCC is limited by the technology that’s available. LTE networks of the capacity and scale we have today couldn’t be deployed in an uncontrolled, unlicensed environment. The technology probably exists today to build something like LTE that can work in an uncontrolled environment (though, it certainly didn’t 10-15 years ago when the FCC started down this path). But the FCC’s job isn’t to develop technology, it’s to figure out how to let people deploy the technology that exists in the market.

I think in the long run the technology is getting to the point where the FCC can get out of the business of spectrum allocation and let cognitive radios mediate spectrum access amongst themselves. We’re not there yet, though, and the FCC’s current approach is quite reasonable as a stop gap.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by Aloha on Friday October 13, 2017 @11:51PM (#15470619)
If you weren't doing licensing, you'd still need to do coordination, and give the coordinators teeth to go after people who don't follow the rules - in effect licensing by another name.
Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by mcny on Saturday October 14, 2017 @09:12AM (#15472002)
No it is not the same because there is no exclusivity.
Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by Aloha on Saturday October 14, 2017 @01:40PM (#15473229)
you can't really have two carriers, operating on the same spectrum, in the same place, and have an expectation that anyone is providing reliable service.
Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by noobermin on Friday October 13, 2017 @10:35PM (#15470376)
I hate to bring politics into this but I will: how is it some people moan about the "growth of government" while left and right we see merger after merger without the government doing a thing to stop these mergers? Where is the over-regulation I keep hearing about?
Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by Turing_Machine on Friday October 13, 2017 @11:07PM (#15470478)
"Where is the over-regulation I keep hearing about?"

It's baked into the cake, as grandalf noted. More regulations benefit larger players, who can afford the parasitic staff of lawyers, accountants, and other compliance people to deal with it.

While the big guys don't like regulation, as such, as long as the regulation is applied evenly across the board it doesn't really affect them that much (they just raise prices). It does do a wizard job of discouraging upstart competition, though.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by ZenoArrow on Saturday October 14, 2017 @08:06AM (#15471807)
> "While the big guys don't like regulation, as such, as long as the regulation is applied evenly across the board it doesn't really affect them that much (they just raise prices). It does do a wizard job of discouraging upstart competition, though."

Regulation can exist for good reasons though. Take the pharmaceutical industry as an example, do we want to ensure the drugs that are on the market have gone through clinical trials, even if doing so limits the group of companies that can afford to pay for these trials?

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by Zach_the_Lizard on Saturday October 14, 2017 @08:20AM (#15471832)
If by reducing regulations we increase drug deaths but also decrease deaths from newly treatable illness, not to speak of the benefits of cheaper drugs, it seems the calculus can come out against the regulations depending on how the numbers shake out.

That is to say, the regulations should be subject to a cost-benefit analysis and only those with a positive ROI should remain in place.

At the very least we should be allowing drugs to be sold that have been trialed and approved in other first world markets.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by Turing_Machine on Saturday October 14, 2017 @01:24PM (#15473153)
Many have argued that the regulatory incentives are skewed with respect to pharmaceuticals.

1) Company A's drug is approved, but causes 10,000 people per year to suffer sudden death.

2) Company B's drug is not approved, even though it would have saved 10,000 lives per year.

The problem here is that scenario #1 makes the regulatory agency look bad, while #2 generally does not. Those patients (and even their doctors) usually won't know that there was a treatment available.

This creates an incentive for non-approval.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by ZenoArrow on Saturday October 14, 2017 @12:02PM (#15472727)
> " the regulations should be subject to a cost-benefit analysis"

Not everything related to pharmaceuticals boils down to a cost-benefit analysis. If companies sell a drug with known side effects, but cover it up to make a profit, that might make business sense to its shareholders, but it's not something that benefits society at large.

To give an example, Bayer knowingly sold products that were known to give people hepatitis C and HIV.

https://en.wikipedia.org/wiki/Contaminated_haemophilia_blood...

"Contaminated haemophilia blood products were a serious public health problem in the late 1970s up to 1985.

These products caused large numbers of haemophiliacs to become infected with HIV and hepatitis C. The companies involved included Alpha Therapeutic Corporation, Institut Mérieux (which then became Rhone-Poulenc Rorer Inc., and is now part of Sanofi), Bayer Corporation and its Cutter Biological division, Baxter International and its Hyland Pharmaceutical division. Estimates range from 6,000 to 10,000 haemophiliacs in the United States becoming infected with HIV."

https://articles.mercola.com/sites/articles/archive/2006/08/...

"Recently unearthed documents show that the drug company Bayer sold millions of dollars worth of an injectable blood-clotting medicine -- Factor VIII concentrate, intended for hemophiliacs -- to Asian, Latin American, and some European countries in the mid-1980s, although they knew that it was tainted with AIDS.

The company stopped selling the drug in the United States in 1984, but continued to sell it overseas for an additional year."

Reducing government regulation of drug companies increases the chances of similar underhand practices, as companies can claim they were following the laws they were intended to comply with.

> "At the very least we should be allowing drugs to be sold that have been trialed and approved in other first world markets."

That still limits the pool of companies that can afford it to a select few. It's very expensive to bring a new drug to market, even if you limit the trials to a single market.

To give an idea of how much these clinical trials cost...

http://journals.sagepub.com/doi/abs/10.1177/1740774515625964...

"Therapeutic area was an important determinant of clinical trial costs by phase. The average cost of a Phase 1 study conducted at a US site ranged from US$1.4 million (pain and anesthesia) to US$6.6 million (immunomodulation), including estimated site overhead and monitoring costs of the sponsoring organization. A Phase 2 study cost from US$7.0 million (cardiovascular) to US$19.6 million (hematology), whereas a Phase 3 study cost ranged from US$11.5 million (dermatology) to US$52.9 (pain and anesthesia) on average. Across all study phases and excluding estimated site overhead costs and costs for sponsors to monitor the study, the top three cost drivers of clinical trial expenditures were clinical procedure costs (15%–22% of total), administrative staff costs (11%–29% of total), and site monitoring costs (9%–14% of total)."

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by grandalf on Friday October 13, 2017 @10:39PM (#15470384)
Well, in this case what we are seeing is the growth of a regulatory state that exists mainly to help select firms generate profit.

This is known as "regulatory capture" and infrastructure services that the government can justify regulating for any reason are vulnerable to it.

I'd say that you are correct in the broader observation that both parties strongly favor big government that helps their pet industries avoid competition.

And along similar lines, both parties are perfectly happy with mobile carrier consolidation, which is too bad, because the poor spend disproportionately on mobile services and are most likely to be victimized by misleading plans or deceptive sales tactics (which are extremely common).

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by convolvatron on Saturday October 14, 2017 @10:39PM (#15475581)
a market is broken. you can message 'we need to control this market' or 'its only broken because the stupid government has its fingers where they don't belong'.

which story plays better with the incumbents? ignore any rational discussion of outcomes.

Re:SoftBank Close to Announcing Merger Deal for Sprint, T-Mo... by dnautics on Saturday October 14, 2017 @05:26AM (#15471445)
>I would strongly prefer that any consolidation such as this be allowed only if one of the firms return all of its spectrum licenses to the FCC to be re-auctioned.

That's a really good suggestion!