You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
I've recently been reading a report claiming that the effects of 5G will be negligible. In its most contracted form the argument is - 1) 5G presents little that is fundamentally new over 4G, 2) the spectral efficiency of 5G will be equivalent to 4G
and will not drive capacity gains. Is this credible?
I'm no expert in this field, but while the spectral efficiency (b/s/Hz) of 5G might be equivalent to 4G, 5G is set to use small cells and mm-wave frequencies up to 60 GHz -- where there's loads of bandwidth/ channel. That's the source of the high data rates available -- Gb/s downloads.
Perhaps the report is referring to the 5G initial deployment, otherwise I don't really understand their claims.
They claim that this presents a catch 22 - either i) use available mmW spectrum which will add capacity even at similar spectral efficiency levels to 4G but which is not cost-effective due to meagre propagation and interference issues requiring a super-dense small cell network , or ii) use mid-band spectrum which should be more cost-effective but is legacy tech
needless to say, I have no idea what to make of it(!)
It was interesting to see TSMC mention AI and 5G as the top semiconductor drivers moving forward at their recent symposium. In the past TSMC had always mentioned multiple market segments: HPC, IoT, Mobile, Automotive, etc...
I agree completely though, AI will touch all market segments and increase the die sizes accordingly. 5G will be an AI enabler so it is in the critical path, absolutely. I'm not sure however how quickly 5G will impact the customer. It could be months or years. Here is a nice article from business insider if you want a quick refresher:
How about this: 5G is a telco gimmick to get more spectrum, squash wifi and choke off access. Domestically if they wanted to deliver faster access they would 1. Lay fiber. 2. Use their own 4G spectrum better. We are something like 15th nationally in 4G throughput. 5G won't work without a lot more fiber in the ground anyway as signals don't go far at high frequencies. Once you have an antenna on every building, you have fiber too so why not just connect to that? The way INTC puts it we will have to lay fiber along every road to get to those tiny antennas everywhere for cars to work. Meanwhile 30% of our population relies on DSL speeds for home access. Sheesh.
On the other hand if you were a telco and though all the spectrum they want for '5G' suddenly gets opened to new public wifi bands, oh my you would have competition. That is what they want to avoid.
Overall, I suspect they are just too negative, and are attacking the (hyped) idea that 5G is completely revolutionary. But it looks like there will be more of a path to 5G via LTE Advanced Pro and similar. Along the way, the new, and very real tech: beam forming antennas, mm-wave radios, massive MIMO will be deployed. And we'll end up with something that looks like 5G in the end.
If you look at the transition from 3G to 4G, it was very much an incremental, gradual process where capacity was first added to heavily populated developed areas and then slowly expanded. If you look at coverage maps, 4G still doesn't have the same coverage as 3G. I think something similar will unfold for 5G, where again you'll see coverage rolled out to add capacity specifically where it will be needed, and that capacity will come from utilizing higher frequencies and small cells vs spectral efficiency. I look at places like arenas and public gathering places as places where you will initially see 5G, and maybe after that hospitals and business centers where IoT capabilities might be utilized. But I expect it'll be rolled out incrementally and strategically over the span of a decade or more - similar to the 4G roll out.
I'd also add that 5G today feels like it's at where 4G was in 2007-2008. People were talking about it, carriers were experimenting with it, but large roll outs didn't really start until 2011-2012 or so.
Si photonics will find its way into WDM coherent products like Acacia is doing while just like in the data center products, it’ll co-exist with InP counterpart for a long time because one it like CMOS in IC offers inferior performance than III-V counterparts and per perhaps more importantly it is a small part (though important) of the whole products that doesn’t make the biggest impact on BOM at component level.
Like other 3GPP generations, 5G will be introduced gradually. Right now we just have New Radio supporting Enhanced Mobile BroadBand; other proposed services (such as millisecond latency for live media) will need new forms of packet routing in core and access networks, see Packet routing for the other two corners
Yes, seriously OTT. The first link is a bit more sober, and I think you can get from it the message that the first phase will give a bit more capacity on Internet access (especially with all those GHz of extra spectrum in the mm wave bands, though to use them operators will have to put a lot of fibre in the ground) but not much else.
BTW the piece on "5G WiFi" at the end of their FAQ seems to confuse Fixed-Mobile Convergence (a part of 5G that afaik 3GPP are still working on) with 802.11 in the 5 GHz band.
