maxsilver

I do. Michigan has 700mhz deployed in multiple areas -- and, by three different carriers T-Mobile has some Band 12 running in West Michigan today.

Depends on the group, the situation, their network, etc. Some of them go back to a server in a CLEC's datacenter "point to point", and then out to the "pure internet". (That's what my numbers above are from. This is my recommended, ideal, perfect-world situation). Sometimes they microwave to a central point, where a rack is hosted, and that goes to the internet over a CLEC connection. (This is more typical, there's a WISP here on three cell towers, that all backhaul over microwave to a downtown rooftop, which hosts server equipment and backhauls to CLEC's IP Port over fiber from there). It's not a nice controlled datacenter, with nice steady lines -- but you get basically the same setup, it's a perfectly safe/acceptable setup, and it can sometimes be a lot cheaper (fiber lines usually require right-of-ways, and those are taxed. It can get pricy). Sometimes they take their CLEC's IP Port (or "pure internet") and plug it straight into a sector. This is a terrible idea for a bunch of probably-obvious reasons, but I know for sure one of them does it. And I yell at them not to. And they just shrug and do it anyway, because it *technically* works. And when I object, they say, "But I can just throttle subscribers on the CPE firmware. Who cares about QoS? And what's a CALEA?". And then I sigh loudly and facepalm. These things are often passion projects. There's rarely any money in it.

Where's your Aunt located? Your usertag says your in Michigan, I might know offhand of someone local to her area. Otherwise, there's a directory at http://wispdirectory.com/index.php?option=com_mtree&task=listcats&cat_id=28&Itemid=53 (fair warning, it's like a decade out of date. But a number of those entries are still valid and running)

I've never built a tower, I've only spec'd leases for existing. So I can't comment on that. (Although some of the guys I've talked to have built their own. They have a lot of clever ways to do that for crazy cheap. But they're also in the middle of nowhere, and I'm not entirely convinced they followed full building code / filed the proper permits ---- anyway) Here's the approx costs for the most recent site I spec'd out, early last year : $900/month - an American Tower site, in suburban Grand Rapids. (Next to a freeway, near-line-of-sight to one urban/dense neighborhoods, and two suburban neighborhoods). $2,000/month - Backhaul from a local CLEC, usually fiber. 200x200Mb fixed. (No one ever needs all that upload, but most of the LEC's I deal with only sell symmetrical, so we end up getting it anyway) $1,200/month - Equipment Costs, Installation Costs, minor utilities, Insurance (Amoritized across 36-60 months, depending) Total cost: (approx $4,100/month for 3-5 years). Note that I usually work with Wireless Internet Providers ("WISPs"), not cellular, so there are some differences. A few things in our favour : - Our gear is presumably cheaper on some sides (Ubiquiti is presumably much cheaper than Nokia, or ALU gear) - Our gear isn't always cheaper -- sometimes the gear costs the same (our DragonWave's are presumably identical to their DragonWaves, if they're still using those for microwave links). - Our people are probably cheaper (most of these guys do the tower installs themselves, or have a full-time on staff tech who's certified + insured for climbing. Labor is almost never outsourced, unlike cell carriers) - Our spectrum is "free" or "almost free". (I've only ever worked with, or seen people use unlicensed ISM bands for most access. A few of the backhaul links are FCC licensed, but it's not like cellular spectrum, the cost for that is much cheaper and non-exclusive). - Bigger carriers sometimes put expensive stuff on the ground near sites. (Usually racks. Occasionally generators). WISPs put almost nothing on the ground. Ground space costs extra (anywhere from $100 - $600/month, depending on how much you want). It's also not usually available, if other carriers are already on a cell site. - Our gear usually has lower weight than big carriers. And we use less sectors. (This matters, both for leasing costs and weight / wind load). - We use CLECs almost exclusively. (Competitive local exchanges). T-Mobile, to my knowledge, also does this, only with much bigger CLECS (like Zayo). Sprint only uses ILECs (incumbent, monopoly providers, AT&T, Comcast, Charter, etc -- according to what I've been told here at S4GRU). - - - However, T-Mobile (or Sprint) should get better, cheaper pricing than we do, in almost every other factor: For instance, the American Tower sales rep calls me every six months, asking me to lease more sites. If we did, he claims we'd get cheaper leases on all of our sites. None of the groups I work for have the capital to have more than a couple of sites at any time -- and even if they did, WISPs generally need less density since they don't provide "coverage", but pure line-of-sight to specific buildings. But an large entity like Sprint/T-Mobile should need lots of sites and high density, and should get cheaper lease costs on every site because of this. Similarly, our lease length are pushed as short as we can to reduce risk (Usually 3 years, as this is the shortest lease we can usually get). But that bumps our cost up too. An entity like Sprint / T-Mobile can probably jump straight into 5-year or longer leases, which reduces the monthly cost even further. We pay retail for our gear, since most of these guys buy in small quantities. An entity like Sprint / T-Mobile is probably buying thousands of units at once, they are (hopefully) negotiating volume discounts for their stuff. Our insurance is sort of silly expensive. Cell providers, being a large enough entity, probably doesn't need the same kind of insurance we have, and save a lot of money there. (If a single tower of gear gets wrecked, a WISP might loose their multiple month's of margin overnight. If a single tower goes out from T-Mobile, they can just pay cash money to replace the whole thing pretty quickly -- the have such huge scale it probably doesn't even effect their financials). - - - I'm by no means an expert at this. But I've seen regular folks with kids and mortgages set this stuff up, and I've run the numbers myself. If they can do it (in a state with crazy high unemployment and 20% lower-than-national-average wages), there's nothing preventing anyone else from doing it.

Internet commenters are ignorant. I wouldn't worry about them. And I see no evidence they're on any kind of "tight watch", just because they want to sell their shares. We've been conditioned to think that networks are pricy to run. It's simply not true. They aren't cheap but they aren't expensive, and I've got numbers from sites I've spec'd, along with a dozen other local wireless guys here who all have the numbers to prove it (even using the best American Tower sites, or fiber backhaul). According to T-Mobile's own ARPU, T-Mobile should only needs about 90 smartphone subscribers to break even on an average tower. That shouldn't be hard for them to do in a lot of these small towns. Even if they're lacking 700mhz. - - - Harrisburg, IL, for instance, has a population of about 9,000. From each carriers coverage map, it looks like there is no native Sprint or T-Mobile coverage there today. T-Mobile could throw two LTE sites up in the town -- if they get just 180 of those 9,000 people to switch (just 2% of the market) -- they break even. Those numbers aren't unreasonable. In fact, they aren't much different than a small suburb in a big city. (Hudsonville, Mi has 2k less pops than Harrisburg, IL. T-Mobile covers Hudsonville just fine with wideband LTE on three sites). I don't think most people will ever leave Verizon or AT&T in these rural areas. But find it highly unlikely T-Mobile's network expansion will be a "net loser" for them in any significant manner -- it's likely T-Mobile will pick up a few people in these rural towns and that's all they really need to cover the costs. Your absolutely right -- there should be greater density there (and almost everywhere). I also agree with you -- I'd love to see them densify native coverage first. But I don't think they'll bother with that until they have to (where "they have to" means, their average speeds start dropping below AT&T because of congestion). But based on the maps you posted, I don't know that they actually need to do a lot of work. They only appear to be one or two sites behind Sprint in most of those areas. That adds up, but it's not a crazy stretch. And that's all they really need to stay competitive. Additionally, for the most part your talking about hyper-rural areas. The cities you show all look fine. (For example, from the photos you submitted, Decatur has roughly identical site density as Grand Rapids, even though Grand Rapids has 2.5x the population.)

For mom-and-pop shops, I would somewhat agree. (Although if any meaningful number of them could afford licensed spectrum, Ubiquiti would probably make gear for them) But I certainly fail to see how that applies to Dish. They have no trouble finding vendors to make them custom satellite dishes or custom advertisement-skipping DVR's. They could get the custom gear they need, if bothered to try. Strong buildout requirements might make "lack of equipment" an issue for tiny providers. But DISH has no excuse on that front. Agree, 100%. Not to mention rural WISPs who could do some amazing things with licensed spectrum. Which is why most people are advocating for Title II + Municipal Broadband. Which would give independents equal footing on every one of those networks, enforced by law. It's not free money like the ARRA's mess. Every example I'm aware of of local Municipal service with Title II (or similar) protections has been extremely popular, and worked really well. If you can point to any examples where that did *not* happen, I'd love to hear them. When there is competition, I'd agree. For instance, mobile is (currently) fairly competitive, in my opinion. But in other areas (landline broadband) -- there's zero competition in almost every market. It makes a lot of since to me for local municipalities to be allowed to build their own last mile, so long as it's protected by Title II. Especially since they often do a good job with it.

Thats...that's not how government broadband typically works... This is getting offtopic, but I've also built and worked for an independent ISP. I've also met with half-a-dozen other ISP's across Michigan. Title 2 + Municipal Fiber would be a godsend to basically all of them. (Even the ones that already paid to put their own fiber in the ground). Gigabit Fiber to homes, $65/month, from your pick of 8 different providers (on top of regular Cable + DSL services). "Government fiber" is real, it exists today, it's not "controlled" by the government in any noticable manner, and it basically rocks. - http://www.utopianet.org/pricelist/ While there are plenty of growing pains with programs like that, it's clearly the right direction to push for. We need people to promote this. Not pretending it's some sort of "government boondoggle"

No, no no no. That's not the takeaway from this. The FCC should welcome new entrants into telecom. But spectrum sold should always have strong buildout requirements. All the FCC needs to do is have and enforce strong buildout requirements. Serious new entrants will meet them. Fake new entrants won't, and will loose their spectrum.

Finally! I'm glad someone is trying to hold Dish accountable for this mess. Or at least pretending to care.

Yeah, since we got 1x800, I've never actually *needed* it -- I no longer loose voice/txt coverage in my market, and there's full roaming as soon as I leave the city. And since Sprint WiFi calling doesn't work very well anyway, I rarely try to use it anymore. A number of major markets don't have 800. My biggest concern would be for folks there

Yes, the call just drops. Sprint's WiFi calling isn't quite as reliable as I'd expect it to be either. They do have it, it does "work". But it can be pretty flaky at starting up, authenticating, etc. It's more of a beta feature than a launched service. On top of those problems, there's the fact that it's missing from every other type of phone (iPhone / BlackBerry / Windows Phone -- all support WiFi calling, but not on Sprint). And even some of the other Android phones haven't been updated to support it. I think it's the right decision, to not advertise WiFi calling right now. The 1x800 voice/text experience is much better anyway, let that take the limelight on voice/text.

To be fair, "1-2mbps" is actually somewhat accurate representation of average speeds in West Michigan during peak hours for a lot of urban Michigan sites on B25 / B26. (such as ones within the city limits of Muskegon). His description of it isn't quite accurate, but if you run speedtest on those bands, you'll regularly see 2mbps-ish speeds in those areas. However, Muskegon's an officially launched Spark market -- it has plenty of B41. And B41 in Muskegon routinely pulls down 7-20+mbps. Complaints about B41 coverage are pretty fair (site density isn't great). Complaints about B41 data speeds are...lies, for the most part. Muskegon's B25/B26 congestion is higher than normal, because they have a higher-than-normal amount of Sprint subscribers, since T-Mobile didn't exist there until last year and even today, only offers slow <1mbps "3G" data there (EDIT: Coverage they care so little about, they didn't even remember to include it on their new coverage map... ) Muskegon is also a very poor city, even by Michigan standards, which exacerbates the "non-Spark devices on network" situation...

Which is kind of ironic, since if he *was* actually in Muskegon, he'd have access to Spark right now...

Who/what is "Joseph A Sofio"? That's not another bidder for Dish, is it?

I don't think it's that surprising. Everyone knew T-Mobile needed that spectrum -- and (after looking at the different bids) AT&T appears to have pushed the bid up to prevent anyone else from getting it. T-Mobile almost had the I block, until AT&T outbid them at round 48.

You could easily convert it from http://transition.fcc.gov/Daily_Releases/Daily_Business/2015/db0130/DA-15-131A2.txt I could convert it to CSV later tonight when I have some time, if that's helpful.

AT&T bought most of the good spectrum here (as expected -- they sort of need it). And Dish is squatting on more spectrum, per usual. But I'm kind of surprised by T-Mobile. I expected them to buy nothing here, but it looks like they got the H block and some G block in a lot of northern Michigan.

I've noticed something similar. Verizon has been pushing density + small cells like crazy over here, even our smallish suburbs have gotten some. They literally have dual overlapping coverage on both bands in a mall here (one tower of which they constructed themselves, this year) plus a DAS in the mall, plus a small cell just for a one particular Costco and street between those towers. They're building lots of capacity they don't even need yet. Which is awesome to see, but I'm sad it's only Verizon doing that here so far. My guess is that Verizon is determined to retain that #1 network ranking. And since T-Mobile is pushing data speeds (and winning on them, in a handful of markets), Verizon's upping their speeds through better density (better AWS reception).

No one actually knows. We're all just throwing out ideas here. Rumors and speculation. They could fall back from Sprint LTE to T-Mobile HSPA+ , or T-Mobile LTE to Sprint 1x. It's completely possible to do this, even on a single RF path. We know this is a fact, because it's already live on network today (for instance, T-Mobile LTE already falls back to MetroPCS 1x CDMA for compatible subscribers *and* to T-Mobile HSPA+ for those subscribers, by using dual MNCs ). There's lots of neat tricks like that could be used to make it happen. However, it is extremely unlikely Google will use this, because it's hard to orchestrate (technically both from the device side and from the network side, as well as politically from the business standpoint of convincing two competitors to work together). It's also carries odd side effects (for instance, if you drop from LTE to HSPA+, what happens when you need to drop from HSPA+ to 1x? Call drop?) But it's not impossible. - - - No one actually knows what Google will do. But the cool thing about Google is that they are big enough that it becomes plausible that interesting things could occur. No MVNO has ever been big enough and cared enough to make Sprint + T-Mobile support anything interesting from a network sharing standpoint, even though they technically can do so. Google's fun to talk about, because they're large enough that it becomes plausible to discuss. That doesn't mean Google will do anything interesting. It is very likely that they won't. I won't be shocked if they just pull a Straight Talk / Ting and ship out different SIM cards for each network. But Google could potentially do something interesting, And that makes it fun/interesting to discuss.

Yeah, it wouldn't surprise me if RW pays approx $5/GB. I suspect the 3G plan assumes the average user is using an average of 1GB cell data per month ($5ish in data costs) and the 4G plan assumes the average user is using an average of 3GB cellular data a month ($15ish in data costs)

It depends on how your calculating "margins". Your "119%" number is probably accurate (in costs to Sprint) but appears to *only* count Ting's costs to Sprint. Typically, "margin" includes a lot of other non-Sprint costs. For instance, Ting has to handle device fulfillment orders. Ting incurs some costs in advertising / marketing / etc. Ting also has the best-of-industry customer service (far better, in my opinion, than postpaid Sprint and T-Mobile, on both personal and business sides), which is particularly expensive. Ting also has the best website / online experience (in my opinion) which they developed with their own software team, which is another somewhat expensive endeavor. Ting's also pretty gracious on credits -- if a mistake was made, and you call them, they waive costs and fees pretty easily. I suspect once you add those costs in, Ting's true margins drop from "119%" to roughly 40-70%, which is still very high, but not ridiculously unreasonable. This is just an assumption on my part though. - - - I do believe you are absolutely correct on your Ting vs Republic Wireless comparison. Ting happens to charge a very high premium for data (more than any other consumer MVNO I'm aware of). For Republic Wireless, it wouldn't surprise me if most of their subscribers never hit the 5GB ceiling. Your numbers sound appropriate for a "worst case" scenario. But I bet a lot of their subs hover in the 1-3GB monthly data usage, which would bump their margins back up a bit on the non-maxed-out users. Especially since they reward subs for WiFi usage in a non-antagonistic way.

http://en.wikipedia.org/wiki/Coupon-eligible_converter_box Wikipedia claims there was a shortfall in funding, some of it came from the FCC via auctions, but some of it came from other places (including part from the American Recovery/Reinvestment Act)

Ting almost certainly pays less than $12 per gigabyte. (My guess is around $7/gb) What Ting charges has very little to do with what they pay. Google's $2/GB is very low, but not ridiculously low. Most group's specific agreements are under strict NDA's, so I can't *prove* this with a source, but it's not uncommon for MVNO's to pay $4-$8 per gigabyte depending on their size / scale / negotiated agreement / etc. (and those prices are from 2013 -- I have to assume they've dropped some since then) Prices also drop if you buy in bulk, and pre-purchase ahead of time. (Which most MVNO's can't afford to do, but Google presumably could). It's also typical for Sprint / the parent provider to only get 25-40% of the "sticker price" of any plan you see an MVNO offer https://gigaom.com/2012/06/25/why-are-mvnos-so-hot-right-now-thank-the-carriers/ (although this fluctuates depending on how much of the Sprint "platform" they choose to use alongside simply offering service. Sprint specifically has a product that's a "MVNO in a box" -- with literally everything done for the MVNO already -- that they typically charge extra for). No consumer MVNO is paying $10-$15 per gigabyte unless they've made a huge mistake. Some enterprise MVNO's for M2M are locked into agreements like that. But they typically sell $5-$20/month plans for industrial/commercial machines on tiny data allotments, which is significantly different. (Think data plans for semi truck GPS's, or plans for ATMs, plans for ZipCar/Car2Go trackers, remote weather monitoring stations, etc, where monthly usage is measured in the 5 to 90mb range)

Ah, ok. Sounds good! I know it's all just rural / small town, but I really hope they add PCS HSPA+ up there (instead of PCS LTE), in addition to whatever B12 (and potentially B4) LTE they run. In some ways, Northern Michigan is a decade behind the rest of the world. The population age skews a lot older and those folks are still using a lot of older devices up there. It's also heavily AT&T country (no TMO presence + poor Sprint presence + ATT ILEC = lots of extra AT&T cellular subscribers). I think there's still a solid business case to spend money deploying new HSPA+ in those particular counties.

Why the "exception"? They have enough spectrum there.