The Linux Kernel Looks to Eventually Drop Support for WiMAX
phoronix.com
phoronix.com
This news is about dropping support for a specific WiMax chip. It shouldn't affect many current users who would be using a discrete wifi device that connects by wifi or USB.
The branding is pretty confusing. My device/plan had HS mode which is apparently WiMax 2+/TD-LTE, and HS+A mode which is normal LTE and much faster. And, use of the much faster HS+A mode was limited to 7GB per month. Once you go over 7GB, the speed of both modes is throttled to 128kbps for the rest of the month.
The latest version of WiMAX, WiMAX release 2.1, popularly branded as/known as WiMAX 2+, is a backwards-compatible transition from previous WiMAX generations. It is compatible and interoperable with TD-LTE.
Does this mean that WiMAX 2.1 can talk to an original WiMAX network also also talk to an unrelated TD-LTE one as well?
...But when Clearwire started to go downhill, they just... shut it off. There was no way to migrate over to a regular Gmail account, Google obviously had no bespoke method for people to keep using their @clear.com addresses. People just straight up lost their online identity. I had a customer at the time who literally just didn't understand why she couldn't log into her email anymore, she hadn't understood the notices Clearwire sent out.
In comparison, people with aol.com and earthlink.net accounts are still out there. Heck, peoplepc.com email is still out there! I'd argue if you're going to hand out consumer email addresses, you ought to have a plan to ensure they effectively never lose their email addresses. I am a little vexed that Google even let them do what they did, and neither Clear nor Google had a real migration plan when they ended it.
"...starting April 15, 2015, Clearwire will no longer support your CLEAR Email/Clearwire Email services. After that date you will no longer be able to send or receive emails, nor will you have access to any stored emails. This service deactivation affects all CLEAR Email/Clearwire Email accounts. Other applications that come with your Clear Email/Clearwire Email are also impacted and will no longer be available including Google Talk/Hangouts, Google Drive and Google Calendar..."
> I'd argue if you're going to hand out consumer email addresses, you ought to have a plan to ensure they effectively never lose their email addresses. I
This is why regulation is so sorely needed. The "free market" has shown over and over again that it is not to be trusted with unregulated identity providing - I remember a time when there was at least one thread a month on HN where people complained about losing their GMail/Facebook/Amazon... account over "automated" tools (aka AI running rampant) and no way to contact humans.
Not surprised about Google's implicit complicity. They don't really think about this stuff at a corporate level.
Then I got a proper apartment and cancelled the service and later Sprint bought Clearwire and that was basically the end of WiMAX as a wireless WAN option.
Sprint wanted to outsource their 4G network to Clearwire. Instead of investing tens of billions building a 4G network, they wanted someone else to do it for them. Of course, if this had been successful, Clearwire would have become one of the major carriers while Sprint became relegated to the status of an MVNO. Sprint didn't want this and their dealings with Clearwire were terrible.
Sprint kept starving Clearwire for funds. Sprint was paying Clearwire about $4/mo per 4G customer. This meant that Clearwire simply didn't have the money to expand their network.
Sprint not committing to a strategy and being too cheap to build their future is a pattern with the company since their 2005 merger with Nextel. Nextel's technology, iDEN, had no future. Sprint kept it around until July 2013. They were running a network that could barely offer 2G data speeds three years after competitors had launched LTE and 6 years after the iPhone launched. This isn't keeping a compatible network around like carriers will do with LTE for many years after 5G has launched - it was keeping a separate network around, often with vastly different coverage.
To put that in perspective, AT&T shut down a similar network (TDMA) in 2007/2008. The writing was on the wall that the future was data and more efficient technologies and AT&T had been building that and transitioned their customer base. Sprint, by contrast, announced in late 2010 (after Verizon had launched LTE) that their "Network Vision" would include spending lots of money on iDEN. That would be like a company today announcing that their vision for the future meant spending lots on 3G.
Over the years, Sprint became frustrated by Clearwire and tried to find someone else to build a 4G network for them - this time LightSquared. The only problem was that LightSquared's spectrum was meant for low-power satellite communication and is right next to the spectrum used for GPS. Experts and the FCC agreed that using that spectrum for high-powered 4G services would drown out GPS signals. But Sprint spent a long time happy with the thought that they still wouldn't have to build a 4G network.
And they didn't really have a ton of spectrum. Sprint had the least spectrum of the big-4 carriers, but did own half of Clearwire (who had a ton of spectrum). Sprint launched its LTE network before buying Clearwire and they had very little spectrum so they only had a 5+5MHz LTE network (while competitors used 10, 15, and 20MHz channels).
Could Sprint have bought Clearwire sooner? Probably not. They ended up buying Clearwire at a valuation of around $10B which is high given that Clearwire was basically facing bankruptcy.
Plus, that spectrum is hard to put to use. Verizon and AT&T became large carriers through the use of sub-1GHz spectrum that could travel far and cover the suburbs and rural areas. T-Mobile started really doing well as it bought 700MHz spectrum and started increasing its coverage and reliability. The 2.5GHz spectrum that Clearwire had would have significant disadvantages even compared to the 1.7/1.9GHz spectrum. That would mean having to invest in way more cell sites to get coverage. For T-Mobile, this is fine. They already have nearly double Sprint's cell site count, as a merged company they have the subscriber count to build more, and they also have enough sub-2GHz spectrum to support a strong base network in areas where the 2.5GHz spectrum won't reach. Sprint had none of that.
Of course, Sprint spent several years after the Clearwire purchase deciding to try an alternative way of building a network without spending lots of money. It didn't work. Neville Ray (T-Mobile's network guy) said on an investor call that if there were a way of building a great network without spending billions of dollars every year he'd love it, but he hasn't seen anyone make that happen.
It wasn't about a single mis-step on the choice of technology. It was 15 years of incompetence and WiMAX was one of the least terrible decisions.
Ultimately, WiMAX was probably done in by Qualcomm most of all. Qualcomm was hugely threatened by WiMAX which would mean way less in royalties for Qualcomm. Qualcomm got Apple to agree to not create a WiMAX phone in exchange for lower royalty rates (https://www.ibtimes.co.uk/mobile-wars-qualcomms-actions-may-...). I'm guessing Qualcomm pushed on Verizon and AT&T to adopt LTE in the US (as well as operators in other countries).
WiMAX certainly performs poorly against modern LTE, but modern LTE has seen a decade of sustained investment. Initial LTE rollouts weren't nearly the same as we see today. It's also important to remember that Clearwire's WiMAX rollout was poor, but Sprint's 2.5GHz LTE network didn't do much better. It was also a dud leaving customers with poor service and inconsistent coverage. The failure of Clearwire and its WiMAX network might be more a story of underspending and spectrum that is too hard to use in the United States (where the population isn't dense and there's an insane amount of local regulation around building anything).
It wasn't Sprint's technology bet.
I think the key is that it wasn't about LTE vs WiMAX as much as it was about 700MHz vs 2.5GHz. Clearwire's 2.5GHz network would need 13x more cell sites for equivalent coverage which meant rolling out a 4G network was basically impossible. You weren't going to get customers without coverage and you weren't able to build coverage without the customers paying you.
Now, for T-Mobile, the situation is different. They have 600MHz and 700MHz networks providing broad coverage and the 2.5GHz spectrum can give them increased capacity and speed where they need it. T-Mobile talks about it as a layer-cake: low-band to make sure you have coverage everywhere, mid-band to give you better speed and capacity even if it doesn't cover everywhere, and millimeter-wave (Verizon calls it Ultrawide Band and AT&T calls it 5G+) in select areas that need lots of capacity. But it's the low-band coverage that gets enough customers that you can invest in more cell sites to make 2.5GHz work well. And it's the low-band and 1.7/1.9GHz spectrum that makes sure the customer still has a good experience even when they're a bit farther from a cell site or a bit deeper into a building.
Without that base, Clearwire was deploying a network that would just have crap for coverage. And Clearwire was also deploying their network before 4x4 MIMO on phones, before HPUE (high power user equipment), and before massive MIMO network equipment. When they were deploying, the low-band advantage was even bigger.
I think it's situational. Verizon being first to 4G really helped it - they were the first to deploy 4G in what became the dominant way that 4G was deployed with a base layer of low-band supplemented by 1.7/1.9GHz spectrum. T-Mobile is taking that strategy with 5G: get lots of low-band 5G coverage and supplement it with more mid-band spectrum than competitors can come anywhere close to matching. Their customers will end up getting a much faster 5G experience through 2021. Interestingly, Verizon wanted to go the Clearwire route with 5G: deploy 5G at much higher, shorter-range frequencies first. Verizon is backing off that plan and deploying low-band 5G now as I think they've realized that strategy won't work well.
If Verizon kept ignoring low/mid-band 5G and just kept pushing millimeter-wave for the next 5 years, I think customers would care a lot that T-Mobile had nationwide 5G before Verizon. Verizon pivoted and that was key for them. Sprint/Clearwire mostly just kept stumbling. I guess being first matters if it's an actual first rather than a "fake" first. Like, Clearwire launched 4G "first" but didn't really have a plan to continue expanding given their limited revenues and backward agreement with Sprint combined with a lack of spectrum to cover suburbs and such. Verizon launched "first" in that they launched in a way that they could create a nationwide network that gave people consistently good 4G service and that paid off hugely for them.
I think T-Mobile is trying to use that strategy with 5G. Verizon got lots of low-band 4G coverage and then bought more AWS spectrum and got great capacity and speed for users with mid-band. T-Mobile is doing that for 5G.
Sigh.
Did it bring anything new to the linux kernel itself?
(...other than the obvious fact that it implemented WiMAX?)
You are supposed to put your driver in the tree for it to stay compatible, but it gets removed if there are too few users!
It would be interesting if anyone is actually using this at the lower layer with these drivers though. I've never heard of it. It seems like a closed ecosystem where the ISPs always provide the equipment.