The real kicker is the last line. T-mobile can software upgrade its LTE equipment to 5G:
> The partnership with Ericsson means that T-Mobile’s already installed base of Ericsson Radio System radios will be able to run 5G NR with a remote software installation.
Ericsson was/is Intel's large foundry client, but with the 10nm delays and 14nm being thoroughly booked with x86-64 and LTE modems for Apple, there isn't room for Ericsson to produce new chips on modern nodes. Worse yet, I hear Ericsson designed specifically for Intel 10nm, which leaves them up a creek for the next few years.
A firmware update to make existing hardware have a longer lifespan and help Ericsson to retain their clients was likely the only out here. If Ericsson had the new chips they were planning on, this firmware update would be much more expensive and less probable to actually see carriers taking them up on it.
Nope, its Ericsson, and Intel fucked them over pretty badly. All of Ericsson's newer chips were designed for 10nm, which doesn't work & Intel is full up on 14nm. Firmware updates to existing hardware and rebrands of existing chips are all Ericsson has to retain customers until they can either get another fab to build their chip (2 to 3 years from now), or Intel sorts out the yield issues on 10nm.
There are a few links in the subscribers only part of the article that compare the last two filings. IIRC these articles age out after a certain period, check back in 46 days?
That's the good thing with almost all telecom equipment by Ericsson and their big advantage. Remote software update/upgrade while in operation with almost zero down time and also the ability to live/actively patch the running code.
It is one of the core features of Erlang (a language developed by Ericsson), being able to rewrite code in production, with errors/crashes that break one thread not affecting another.
Note that they start crashing things in Erlang around minute 7, and fix them shortly thereafter. Meanwhile, their setup keeps calls that weren't directly involved in the crash alive and running unaffected.
I was working on the code for mobile nodes, 3G and 4G. C++ with RSARTE (IBM plugin for Eclipse for Real-Time Applications) [1]. I left but my friends started working on 5G where looks like they want to get rid of RSARTE and use clean C++. At some point of time some people in Sweden started looking into rewriting it in Erlang but after around 1 year it was dropped, don't remember why.
Speed of running the code was likely why C++ was retained, Erlang is a higher level language that uses more resources, and depending on your platform target there may not be enough CPU resources in your average base station to support a software stack in Erlang doing packet forwarding, QoS or especially DPI.
That being said, for communications platforms it has some very useful properties when code starts to break, hence WhatsApp and 2600hz using it for their respective projects.
Great. Now I'll be able to use up my monthly data in 2 minutes instead of 10 minutes.
Edit: I don't mean to poo-poo the advance of tech, but this really highlights the disconnect between what's possible and what consumers actually want/value.
I was under the impression that the real improvement in "5G" systems was in how many concurrent users they can handle, not the amount of bandwidth they can offer a single user.
So, kind of. There’s two parts to it.. By creating more cells you free up the number of users per cell and the bandwidth demand on that cell, providing lower latency and higher speeds. Millimeter-wave technology is another component of 5g that is going to take a while to materialize in phones and to even solve many of it’s fundamental transmission issues. Eventually, mmwave tech will lead to massively faster speeds, but who knows if it’ll ever make sense to put in a phone.
It would be great to see wireless providers becoming competitors to home ISPs. If some day wireless speeds become comparable, the problem of having only one home ISP would go away.
I get 50Mbps with 4G connection for 20euros per month. No data cap. Public ipv4 address.
Router is 4 years old Huawei 4g router and I have installed outdoor antenna on the roof for stable connection. I have been with this setup at home for the last 4 years.
I could propably get faster connection with new router but this is fine.
That still doesn't compare to the 300Mbps I'm getting right now even through Comcast, or the 1Gbps I'll be getting once the local fiber company builds out in my neighborhood in the next year or so.
Couple things- the range of 5G is pretty limited. Getting a cell down the block from your house would seem to be nearly as hard as running a line to your door. Also, with wireless, never forget about bandwidth. Bandwidth is the reason living next to thirty other WiFi networks tanks your speeds (but not your cat5 speeds!)
> Getting a cell down the block from your house would seem to be nearly as hard as running a line to your door
A 5G cell will be 0.1-1.0 km in radius and cover dozens to hundreds of users. In a suburban area, that might be an entire large subdivision. The cost of getting fiber to the cell might be just as much as getting it to a fiber distribution hub in the subdivision, but with 5G once you do that you're done. With FTTH, you have to then run it to each house, which will take (in the aggregate) way longer than it took to run fiber to the subdivision. The branch factor kills you--yeah, it's just a couple of hundred feet to each house, but it's a very labor-intensive couple of hundred feet (under yards, into basements) that you have to do for each subscriber.
Also, all bandwidth is shared--it's just a question of where the sharing starts. GPON--what's typically used for FTTH--involves sharing 2.4 gigabit down/1.2 gigabit up between 16-32 users per PON. That comfortably supports gigabit service with reasonable oversubscription. A 5G small cell will have less bandwidth shared between more users, but will be in the same ballpark.
that's incorrect. contention is the reason being next to all of those other wifi networks tanks your speed. Only 1 device can transmit on a given channel at a time. It's as if everyone is plugged into the same handful of ethernet cables.
> Priced at $50 for people who already have Verizon wireless and $70 for those who don't, it's promising speeds of "around 300 Mbps" up to 1 Gbps, with no data caps.
I currently use my phone provider's tethering plan (10GB at 4G, then throttled to .6MBPS after but with no additional charges) because there is only one ISP where I live and unless I signed a 12 month contract, I was looking at ~$80-90/month in amortized fees since I'm only living here for four or five months. It's not ideal, but other than watching fewer videos and in lower quality I've been able to manage without too many issues.
Where are you from? Here in the EU every cellular provider offers home internet plans & LTE-routers and ISPs have LTE plans (mostly to "accelerate" otherwise slow rural ADSL lines)
Did you have a smartphone 10 years ago? Do you remember what 3G was like? Wasn't so bad, right? Try using 3G in any metropolitan area today, a lot of times it's painfully slow. Why is that? Congestion. The reason to upgrade isn't necessarily to improve the best case performance for a single user it's to provide a better quality of service for a larger number of users. Anyone who's attended an event like comic-con or PAX knows just how easy it is for even 4G LTE to get saturated with connections.
And cell carriers running 3G networks with minimal spectrum. Why leave spectrum on older protocols like 3G, when it could be making your LTE network faster? Hence T-Mobile killing 3G in many areas, with 2G stuffed into the guard bands of their LTE broadcasts. Not much spectrum allocated for 2G either...
Do ordinary consumers actually run into data caps routinely? When I was on T-Mobile, I used hotspot extensively and never paid attention to my data usage, but never hit the soft cap on their unlimited plan.
T-Mobile only enforces their soft cap when the tower you're hitting is at capacity. Basically their soft cap deprioritizes your traffic rather than throttling it.
I think it depends where you are. I’ve just looked, you can get 100GB for £20 a month in the UK (sim-only), cost of data is not a big issue here.
I agree top speed isn’t too much of an problem in itself, but if with contention that speed can drop a hundred fold and still provide a good connection, the leeway is worth having. And presumably there are range and robustness of connection benefits as well.
Higher frequencies... Lower range, but better channel and beam management, so better in dense cell areas. Connections should be more robust due to what's been learned about customer mgmt and use cases.
It depends. As you can read from the replies, 4G in different market has different situations. The main theme of 5G will be bringing more capacity, and lower running cost for carriers ( At least that is what they said, I am not convinced on this matter ) . Top Speed is no longer a major selling point anymore, they want more Data, and 5G will bring anywhere from 5x to 10x capacity improvements. In theory, that should mean US telecoms has incentive to provide much Data for the same price, and EU citizens will enjoy higher speed as their Data package are cheap already.
The good thing abut 5G is, for possibly the first time in history Carrier is well prepared in technology stack than consumer counterparts, may of the carriers are installing and upgrading to 5G equipment now, but they are for now running 4G signals. All of them ( not just Ericsson, Nokia and Huawei ) allow software update to enable 5G signals. Those equipment already provides better LTE performance, making 4G and 5G transition much smoother. Every step from 2G > 3G > 4G were like revolution, 5G was more of a evolution of 4G.
Yeah, I share this sentiment. I live in Canada, we get gouged bad here. Thankfully my employer covers my cell phone, because my bill would regularly be well over $100/month with the amount of data I use.
They increased it last year from 32GB to 50GB a month. Frankly, I think even a 50GB cap on an "unlimited" plan should be illegal, but it's better than most carriers.
It's not a cap, you're just throttled. If they advertise unlimited 4g or whatever speeds then it already is illegal. No one does that though, they just say unlimited data
> Consumers get shafted on these deals. We pay for products with no agreement over what we're actually going to get. It gets tiring :/
Consumers greatly benefit from these deals, because they pay a fraction of what it would cost to get service with an SLA. (Put another way, it costs a fraction as much to build a network with over-subscription than it costs to build a network where each user has guaranteed bandwidth.)
Where I work, we pay Cogent almost $1,000 for sub-gigabit, and that's in a building with several competing providers to choose from. I pay a tenth that for consumer gigabit from Verizon. That's because I'm sharing a 2.4 gigabit PON node with 16-32 other users, and Verizon can assume that nobody will be using the full gigabit more than a small fraction of the time. If Verizon was only allowed to sell guaranteed bandwidth, they could only sell a 75 mbps service. Which would suck for the consumer when they went to go download an iTunes movie or a game on Steam.
Until there is legitimate competition in these spaces you will keep getting "up to X mbps" and "for the first 50GB, subject to change" in every providers terms.
It doesn't help that actually pricing out Internet access is more complex than X cents per byte. For most providers data between the hours of 0100 to 0600 would be free because of how little network usage there is, while data from 1700 to 2300 would be most expensive due to that being when everyone is using the network simultaneously. But throughout the years there have been very few time limited unlimited plans.
And even then, its not actually "expensive" to use the Internet in the evening. Its just saturated, and pressures the ISP to either throttle everyone or expand capacity. Expanding capacity is expensive, but just promising "up to X mbps" is easy and I'm surprised at how eager operators are to adopt data caps over simply throttling heavy data users during peak times.
Well, I'm not really surprised, because the former lets you hit people with surprise bills for ludicrous amounts while the later just saturates your useless customer support lines with complaints.
My old university dorm internet was like that. There were 3 or 4 time zones per day, and only the few hours in the middle of the night were unlimited, otherwise iirc 500GB/month, measured by double and iirc. even triple-counting during prime time and such. For some reasons they were rather oversubscribed, offering 1000BASE-T in the dorms (with some L2 crypto auth), but only had 10Gbit/s fiber uplink in a few dorms. Considering they already had active equipment on both sides, this shouldn't be much of a problem though.
I know, ya'll want unlimited speeds for cheap. I get that. I just want reliability. Currently every internet I've had craps the bed during prime hours. Which hey, is exactly what you're paying for.
Sounds like your on some shifty cable ISP or an older ADSL based ISP. For cable provided internet, its truly shared, and one person plugging a VCR in the wrong way means your speeds may drop significantly.
Meanwhile, older ADSL terminals are often backhauled by a couple T1s, so 6Mbps may be split across more than a few customers, resulting in poor service. Centurylink calls it exhaustion when that occurs.
This is good news as it kept Ericsson on the battle field. Nokia is looking a lot worst though may be it will merge with Ericsson some day. In terms of telecom equipment industry Huawei is now bigger than Nokia and Ericsson combined. And may be I am the only one not entirely comfortable with that.
I don't know from where exactly you get this impression but last time I checked they only had a 1% difference (28% versus 27% and Nokia 23% with 2017 figures [0]). I can tell you that Ericsson is in a much better position than Huawei. And although I would also like an Ericsson Nokia merger, I don't see it happening soon ;-)
Because even if you take out all the revenue from the consumer mobile phone business, which is growing rapidly as we speaks, the revenue from the rest of Huawei is still bigger than Everything of Nokia and Ericsson combined, in both 2016 and 2017.
I've long wondered if the best way to buy software may be to just buy from multiple vendors and see what works best, damned the purchase costs, just to fix the information asymmetry.
I suspect because it is a Chinese company and having key national infrastructure built by a Chinese company leaves some Americans nervous over national security fears.
it sounded right to me, and i got to this [0] after a short google
ericsson doesnt sell mobile phones however, and nokia got screwed over by microsoft in that department.... that's probably at least a big chunk of that revenue.
4G mobile standard was LTE from 3GPP, which competed against WiMax from IEEE and won in the late 00s.
5G mobile standard is NR (stands for New Radio) from 3GPP. There is no competing mobile standard from anyone else.
Biggest change compared to LTE is the formal standard for millimeter wave communication for mobile. Previously it's been used for fixed wireless access and/or satellites. We'll see how it plays out in a mobile consumer environment.
There’s been rumblings in the BUD (big ugly dish) community that 5G will start to ruin C-Band for TRO (TV Recieve Only). Most birds are going mostly Ku but C-Band is used for backhaul & feeds on a lot of soon to be retired: Galaxy & AMC series.
For Amatuer Dx’ing C-Band offers way more unencrypted and English channels than Ku. There’s way more weird channels too. I just pickuped an old 8ft Dish for free off Craigslist simply to watch North Korean TV.
From what I gather one will need a much larger dish for C-Band in the US to overcome spectral interference or that it will pretty much put an end to C-band in the US.
There is such a thing as torrent though. Make initial reputation come from e.g. some crypto mining sufficient to fund the bandwidth costs from loosing the block to an uncooperative peer, and then handle reputation based on share ratio, either theoretical or measured.
Not really. 5G won't replace 4G. 5G just requires to many cell towers to make sense outside of high-traffic areas like hospitals, airports and stadiums. 4G will still be what's used in outside environments. That's how it's been explained to me, anyway.
Those are two separate technology standards. 3GPP2 evolution got merged and we came down to a single "4G" standard with the LTE air interface. It borrowed a lot from WiFi's adoption of OFDMA that allowed a lot of the bandwidth increases. 5G evolution on the air interface though significant, are just an extension of that, using higher frequency bands. But where the meat of 5G lies is in expanding on the legwork done in LTE specs in preparation for IoT. It brings in the ability for multi-bandwidth, multi-reliability and multi-latency applications and SDN paradigms. "Kubernetes for the Network" kind of terms are being thrown about.
Initial deployments though will focus on "FTTH replacement" scenarios.
5G New Radio (NR) is the global standard for a unified 5G wireless air interface. It was ratified some time ago. This allows investments to hardware.
The standardization efforts continue and they add functionality and more interfaces for specialized services and applications. For example: Multimedia Priority Service, Vehicle-to-everything (V2X) application layer services, 5G satellite access, Local Area Network support in 5G, terminal positioning and location and lots of other service interfaces. 5G will be much more than just 4G LTE with steroids.
Someone correct me if I'm wrong but won't this put Verizon in the opposite situation that they're in now? Doesn't Verizon have coverage in most places because they were able to build early and keep their old CDMA towers in place? With the change to 5G aren't all the LTE/GSM carriers now in a better place to update to 5G? Won't this actually cause Verizon to require huge network and infrastructure upgrades?
The move away from CDMA happened with LTE. LTE required a complete network upgrade to add in the eNodeBs to supplement the BTS (basestations) from the CDMA world.
5G (3gpp rev 15 or newer) does require new hardware but a lot of networks have been planning for this. Sprint has been rolling out new MIMO capable antennas for use with 5g. Verizon similarly has been future proofing before the standard was finished. There will be new equipment required no matter what though. Especially with Verizon pushing millimeter in cities.
Yes, there is a set of 3GPP specs that define 5G standard. Slicing of the network to provide low latency is one of the highlights. There are open source implementations such as openCORD that target the same goals.
3gpp release-15 describes parts of what they're calling 5G [1]. And no AFAICT it's not finalized, there's a lot under this umbrella. IIRC fixed/terrestrial wireless (at 60GHz?) is part of what's referred to as '5G'.
> The partnership with Ericsson means that T-Mobile’s already installed base of Ericsson Radio System radios will be able to run 5G NR with a remote software installation.