Monday 16 September 2013

#5G: Your Questions Answered

This is our view on what 5G is, please feel free to add your comments here or if you want a much wider audience to discuss your comments, please add them to the Cisco Communities here.


21 comments:

Unknown said...

Hi Zahid

First of all thanks for sharing this information, I am one of the big fan of your Blog.

My question is, as you mention in 5G there will be "perfect Cell coverage" does it mean that there will be more improvement in TTI bundling or is there any new way to achieve perfect cell coverage?

If there is some improvement in existing one then what would be?

Thanks
Shailesh

Zahid Ghadialy said...

Hi Shailesh,

Glad you like the blog and thanks for the feedback.

Regarding the perfect cell edge coverage, its one of the issues that need resolving in '5G'. TTI bundling is specific to the way LTE works and may again be used in future technologies but there are other technologies like Relays, Small cells, etc. that I have in mind when we talk about a much more improved coverage. Couple this with CoMP and you have effectively solved the problem.

Again, these are just ideas. What will happen finally, we cant predict yet.

BR,
Zahid

Anonymous said...

Hi Zahid,
Very nice post. I guess the best definition of 5G is that there are none :)
However what i would for sure bet on is 5G will need to encompass:
- lots and lots of data (big and better data)
- lots of devices (human and machines).
- lots of nodes (hyper-dense networks)
- cell edge is not an issue => however latency and QoE are key

Regarding cell edge issues, if you are familiar with the phantom cell approach and massive MIMO, the notion of cell edge will gradually disappear. So i would not put much emphasis on it

PhantomBen

Zahid Ghadialy said...

Hi PhantomBen, good to hear from you. With regards to Phantom Cell (blogged here), its more of an LTE-A concept, rather than '5G'. I would hope that we would perhaps have something new/better by the time we are ready for '5G'.

This would also depend on what access technology is used. I certainly agree that "the notion of cell edge will gradually disappear".

Anonymous said...

That is right: As per DOCOMO Phantom cell likes concepts will remove the user-control boundary. My point was just to go beyond cell-edge KPIs :)

Another paradigm for 5G and even beyond - not very much discussed - is proactive networking. We have just submitted a work on the topic and will be happy to share it with you in due time
please keep up the very nice blog. It does help a max

Theodore Stoner (via Het Net group on Linkedin) said...

I am suprised a specific reference to smart antenna technologies is not included on your list. mmWave technologies and Massive MIMO rely on these active antenna techniques. I am not familiar with FBMC, so I downloaded and IEEE tutorial on the subject. Thanks for the list.

Here is a link to the tutorial if you are an IEEE subscriber (excuse the obnoxious URL):

http://ieeexplore.ieee.org/xpl/articleDetails.jsp?arnumber=4481344&contentType=Journals+%26+Magazines

Manoj Das (via Het Net group on Linkedin) said...

IMHO, active antenna (Or Digital Beam Forming antennas-DBF) is a sub set in the context of 3D Beamforming and Massive MIMO. Please share your thoughts. :-)

Theodore Stoner (via Het Net group on Linkedin) said...

In general yes I do agree, however, I think it is such an important technology that it may be worth mentioning in its own right.

I still think we have alot of growth available in the cm bands before jumping up to mm microwave bands.

The new 3.5 GHz band targeted for small cells in the US is well positioned to exploit improved MIMO (much smaller antenna arrays) and Congnitive Radio. It will be interesting to follow the development of spectrum sharing rules for this band.

Manoj Das (via Het Net group on Linkedin) said...

Thanks Theodore, after the declaration of 3.5 GHz band for small cells, Cable operators are eyeing to deploy voice, data and video over 3.5 GHz along with traditional WiFi which they are building aggressively right now. Of course the incumbent WSPs are also well positioned to take advantage. Although 3.5 GHz band is very attractive for small cells, it has few limitations such has it can not be used at certain coastal "Exclusion zones" defined by NTIA. NOAA radars operate in the same band.
Here is the presentation by PCAST (President’s Council of Advisors on Science and Technology.

http://www.whitehouse.gov/sites/default/files/microsites/ostp/gorenberg_ppt.pdf

Zahid Ghadialy said...

Theodore, the paper you referred to "Multicarrier communication techniques for spectrum sensing and communication in cognitive radios", is it not the same as we have covered in slide 13?

Aldo Kanakis said...

mm wave is necessary if we want to increase the number of active antenna elements in a MIMO setup. The general rule says that the minimum distance between two active antenna elements should be half wavelength.

Now, if the carrier frequency is low (for coverage purposes) the wavelength is large and only a limited number of active antenna elements can be installed in a mobile device (max of 12cm). If we move to the mm wavelength, massive MIMO becomes feasible and hence 3D beamforming.

Although in principle I agree that cm wave has a lot more to give, 5G should rely to a more flexible, higher channel capacity and lower BER scheme to deal with the "crazy" cell capacity demand from the users.

Don't forget, the ultimate goal is to deliver Ultra High Definition Video to a connected device (SIM equipped 50" TV, laptop, tablet, mobile phone etc.).

Theodore Stoner (via Het Net group on Linkedin) said...

Triantofyllos,

As a wireless carrier my goal is to deliver HD content to at least 4 screens in the home, I prefer to do this from a small cell on the street vs a femto cell inside the home. I do not want my customer going to a competitor for broadband access, if it can be avoided. Nor do I want to go on premise to bury wire, etc. If you add in the building penetration losses, this more feasible <10GHz. Also a 50"" TV and even 10" tablets offer more antenna realestate than is currently used by our 4G standards.

Manoj Das (via Het Net group on Linkedin) said...

Hello Triantafyllos, you are only considering two dimensional spatial separation between antenna elements. One should consider polarization separation as well as some engineers are taking more innovative approach such as angle of rotation of the waves etc.

Aldo Kanakis said...

Theodore, don't forget that 5G is a BEYOND 2020 technology. Standardization has not even started yet. In year 2013 you want to deliver HD content to 4 screens and this should be your goal. In 10 years time that should not be enough. You will be looking to deliver UHD video to 4 or 5 or 10 screens. I totally agree with you that small cells and relays will play significant role in the future of mobile telephony but with bandwidth demand, sky is the limit.

20 years ago I could not even imagine downloading a full quality mp3 song in less than 30 seconds when in the old Napster years an average sound quality song would take hours or even a whole day to download. 10 years ago I could not imagine that my broadband speed could exceed 8 Mbps and 5 years ago I would bet that mobile broadband could be faster than home broadband. Bandwidth demand is only increasing and setting up limits can only prove you wrong. With predictions talking about a couple of hundred billion connected devices by 2020 the bandwidth demand could exceed every prediction.

For that reason we need to utilize every bit of spectrum available. High frequency enable higher bandwidth and also the capability to install hundreds or even thousands of antenna elements in a single device --> massive MIMO --> 3D beamforming --> best known (nowadays) spectral efficiency.

Aldo Kanakis said...

Manoj of course you can have more benefit from polarization and wave rotation etc, however the main benefit you get from MIMO is increased dimensionality of the system and the only way to achieve that is to increase the number of antennas. The question is how can we increase the number of antenna elements if we cannot fit them in a device? The best way to fit it is increasing carrier frequency. I don't say that we need to operate exclusively on high carriers, as this will bring us to the "poor wall penetration properties" dead end, however for 5G (or beyond 2020 mobile communications) in my opinion you cannot have the one without the other.

Theodore Stoner (via Het Net group on Linkedin) said...

Triantafyllos,

I will concede that in 10 year’s time we may have mostly exploited the cmW bands and need to focus on mmW. It is hard to get an operations guy to focus that far ahead.

Zahid,

I went back and looked at your slides, yes the tutorial discusses the same techniques you have shown on slide 10 and slide 13. The paper provides further depth for the motivated reader.

Aldo Kanakis said...

Theodore,

I totally agree with you. It is a bit pre-mature for operators to even consider recruiting 5G dedicated staff. 5G needs to go through the whole research, standardization and basic design cycle before we consider it a complete new technology worth to invest on.

Although LTE and 4G are just starting making impact on the global cellular communications market, we already receive questions and concerns on 5G. You see, with the rapid expansion of small cells and HetNets in 4G, people cannot even consider of how could 5G be different from 4G. I don't know what is the situation in the U.S. but I can assure you that in Europe and UK there is unexpected activity in the research and development of 5G. EU and UK governments along with the global telecommunications industry provide huge funds to academic foundations for the research of 5G.

10 years down the line a lot would change that we cannot even image yet. Let's give it sometime...!

Ravi Sinha (Via Het Net group on Linkedin) said...

Really like the information. I have seen a part of it one and half year back, when Professor Ted Rappaport and Prof Paulraj had a discussion on futuristic network. I can see a lot of valuable points covered under this document, and curious to know the companies doing some real work even POC to demonstrate some of the functionality. Mainly would like to know about FBMC. If you guys have any further info kindly share.
Thanks
Ravi Sinha

Robert Syputa (via Het Net group on Linkedin) said...

5G will:
A) Somewhat be more a marketing demarcation than a transition point between 'generations' of networks.
B) The evolution will be more integrated evolution from 4G than has been normally attributed to generations of networks. This is because 4G 1) has reached a stage in wireless signaling in which close to theoretical (practical) limits for shannon-hartley have been reached, causing evolution to occur more in the multi-dimensional architecting of networks and the corollary technologies that support and extend that. These include tiered multi-carrier/band methods, SONs-SDN, both centralized BS and highly distributed small cells with dispersed storage and content/application server capabilities.
C) Cognitive radio methods are finding their way into 4G and will become a major thrust of 5G. However, this might be so integrated into the core signaling and network methods that discrete cognitive functions are put into use separately so that there isn't a single demarcation point that can be called "5G cognitive radio". LTE-Advanced has the capability for sensing and estimation of signal conditions including of sub-channels. That offers the potential to build cognitive radio methods into 4G or to provide the demarcation line to call it 5G whenever marketing departments decide it is time.
D) Multiple carrier aggregation has much headroom: use of MIMO-COMP including '3D Massively MIMO/massively small cell architectures with multiple band/channel aggregation offers almost endless possibilities considering where wireless is today (the real state, not the 'state of the art wish list'). We are discussing methods long in the works but still years away from what had been imagined when IMT-Advanced/LTE-A was being initially thought out.
E) We can no longer think about wireless networks as stand-alone development: that is thinking about where the advantages will come from based on signaling methods exclusive of the comprehensive ICT network. I know that is not how this discussion is considering it, however, in the industry there is still a tendency to focus on just the wireless technology and architecture while giving little or not drift to stuff like smart routing across tiers, distributed storage-server that offers huge potential for relieving traffic while reducing end-to-end latency/jitter. Developments in rugged solid-state storage-routers that can be implemented within the macrocell or as part of the more distributed 'massively smallcell' tiered network should be considered into the equation because it can shift the other parameters of wireless network development. Note: roughly 80% of content/applications are local-time based: therefore, localized storage-server-routing offer great potential for reliving congestion.. Or front-haul via fiber to optimized BS server farms.

Aldo Kanakis said...

I totally agree with you Robert, it seems like 5G is not going to be a brand new technology, a brand new air interface or a set of new techniques. It is most likely going to be the convergence of all mobile and wireless communication systems into one. It may sound small development but I personally think it is a revolution... at the end of the day seamless communications is the ultimate goal. Some improvements in the air interface and networking will set the foundation for a new massive building ready to host all the older mobile and wireless technologies. Small cells, M2M etc are gradually coming into the picture with 4G and I personally believe that 4G is going to be used as the "crash test" of the mesh networking and integration of the various technologies with 4G before the standardization, design and launch of the 5G.

Mohammad Khatib (via Het Net group on Linkedin) said...

All the work of SG13 takes us in the direction of convergence, so I agree 5G is definitely going to be the convergence of all wireless technologies including M2M, but I am a bit doubtful about the air interface. But selection of a brand new air interface does not seem like out of question at least that is what the history of 1G, 2G, 3G and 4G indicates. All these came up with new air interface. So I would be mentally prepared for that.
Considering the 2020 timelines of 5G and the current penetration density of 4G, the question that comes to mind is about its availability on time, or is 5G going to be "go-slow". Not sure.