 Good morning everyone. It's an honor to present the next talk. We'll talk about the 5G hype, why it is justified or why not. So, welcome our next two speakers. I know it's early in the morning, but nevertheless, a warm applause for Peter and Heurekus. Good morning. We are Heurekus and Peter, and now we will tell you something about 5G. Everybody is talking about it. We need it absolutely, etc. But when using the 5G, this term, we like talking about wood instead of talking about trees. And we want to speak about what is possible with 5G, and the 5G we are talking about today will be about the following. I draw an overview about what we will present today. The 5G that is already existing today. I will explain this in detail later. This is the network, as it is at the moment. The black parts are the LTE networks, which already exist. And the orange parts are the new parts, which will be added for 5G. You see mainly in the radio access, there is a new 5G port, which will be added. 5G and devices. And the important thing about the 5G that already exists, it always co-exists with 4G. It cannot be standalone. That's why it's also called 5G NuA Radio Non-Standalone Architecture. They use the acronym NSA, which is... I don't really like to say that. You have to get used to saying this. And while you are getting used to this acronym, Peter will talk about the most important interface in mobile communication, which is the air interface, the interface over the air. The 4G to 5G air interface. I will begin with 4G, because 5G basically is a complicated 4G air interface. And I will begin with how to get data on a radio interface. For this, you use a carrier, which you can usually turn on and off and by turning it on and off, you change the amplitude and the phase. You can do it with four different states from QPSK. And then add more details to it until 64 or 265 QAM, where you can send 8 bits per symbol. So when I have a bunch of carriers and subcarriers, they have to feed them with data somehow. So the carrier will get part of the data from a serial to parallel converter. But there's a problem with these subcarriers. When I have too many of them, or many of them, then they have frequency ranges besides the actual carrier frequencies, which also have power in it. And this comes from turning the carriers on and off by changing the phase. And in 5G or 4G, they are sent all at the same time. So all these non-wanted parts of the frequencies are turned on and off at the same time. And we choose the carriers at the right direction so that the next carrier is at the minimum of the adjacent carrier. For instance, we use OFDM, the subcarriers are in a distance of 15 kHz. This gives a simple time of 66.7 microseconds. So all 66.7 microseconds, these subcarriers are turned off and then they come again with new information. That's basically the same with 5G. But we can also have other subcarrier frequency. Okay, and there's the symbol time. The symbol will take a certain time in the air, 66 microsecond macros cells. Part of the signal will come via reception and part via direct reception. And the guard period, the end device will ignore signals during the guard period. For this transmission we use OFDMA. OFDM has been around for a long time. It's Bluetooth or Wi-Fi. However, all subcarriers are used for one end device. And only when that device has been served, the next device is served. OFDMA is Article Frequency Division Multiple Access. So guards are introduced and spaces and different kinds are served. It's a little harder to calculate and to process than OFDM. OFDMA is more complex than that. And so OFDMA went at us on the air and a number of subcarriers. In LTE with a 20 MHz band width we have 1200 subcarriers. If the client knows there's information within the subcarrier, all that signaling would take a lot of signaling. And so it's divided into resource blocks with 12 subcarriers over frequency and seven symbols. In LTE that amounts to half a millisecond in 5G, that is 12 subcarriers. However, with a 30 kHz subcarrier spacing, the block gets longer and the time shorter. We'll look at that in detail. Reference signals are a super thing. Certain subcarriers sort of peek out here in this complex of time frequency, this time frequency block. We have a third dimension reference signals because of their position where they carry the cell-identified number of enode B and genode B. And because of the power, the end device can measure how strong the enode B is. Reference signals are sent with 15 to 18 dB power. It doesn't sound like a lock. However, the signals aren't alone. There's 1200 subcarriers in the LTE system and a receiver can be very... In GSM, we have 150 kHz, 200 kHz band width, and here it's 15 kHz, so they can be much more selective. We can also go down to 20 dBm sensibility for some reference signal. This is several resource blocks chained, a resource grid from LTE. And in this client state, 1.4 MHz band width, 6 resource blocks. You can recognize the resource blocks in different colors. Green is the broadcast channel. A few parameters are described. And this one exists not yet for 5G because you don't know yet where to position it. It could be positioned anywhere. So the synchronous elements are orange and red. The gray area is for the client the open flow. So to search the data in the white areas, for the resources that a client system shall be looking at. This now is a 20 MHz system and 100 resource blocks. Here they are quite small, the resource blocks. The time that we put onto it is 10 milliseconds. So this is already LTE. In specific advanced standards there is the possibility for multi-cast and public warning systems. But I have the feeling that LTE advanced will now be overhauled by 5G. Yeah, you just have to build them in 5G. If I now take again a resource block, I can calculate the maximum data speed. It has 84 elements. So it's reference signals, so 80 remains for traffic. And if I modulate each of those subcarriers, I can... Each of these subcarriers can carry two-to-ight data bits. When using, for instance, the 64 QAM modulation, you take one of these rows with 12 subcarriers, then these are about 960 kilobits per second per subcarrier. So this then times 100, because I have 100 of these resource blocks on top of each other. Then I have expected speed of about 96 megabits. Let's call it 100 megabits. And if we add MIMO, you kind of double the data rate, but approximately times 1.6. That's the approximate speed we can have in such a system. So what is MIMO? MIMO, in principle, it is the transmission of several data streams at the same time on the same frequency, but with different, for instance, polarities or... Yeah, that's quite challenging because the enterprise can move. So about each millisecond the channel is re-measured, and then it's checked if you can do MIMO or not or which kind of MIMO. To use MIMO, you need several antennas. For instance, if you want to do four times MIMO, you need four antennas. This will add to about three to four times the data rate that you could normally achieve without MIMO. So now we did 4G, we have all basics to explain you 5G because we will do something with resource blocks and stop carrier spaces. What's the problem? We only have 20 MHz carrier bandwidth. We could aggregate several frequency bands, but the maximum is 20 MHz. We only have potato cells, I call them like that. If one client is creating traffic, then the cell doesn't care where these cells, the signals are sent in the entire cell, which leads to more interference with other systems. The idle to active time is always under 100 milliseconds. This means for a client, it always needs to sleep for sometimes. This time means the device can sleep for 99 milliseconds and only one millisecond it has to get awake to check if something happens for it. With 5G we can change this time based on the new structure. 5G gives a lot of new possibilities. As I said, 5G is a complicated 4G, so it's made such that if somebody has a new idea, then you have to find a device manufacturer who will do something with it. I need a device manufacturer, I need a system manufacturer who will implement the new feature, and I need a network operator who will operate this feature. So that's the structure for 5G, for a feature that doesn't exist yet at this point of time. We have a larger carrier band with up to 100 MHz. We can use beef forming and more MIMO possibilities. The idle to active time can be variable. If you want to do low latency stuff, then 100 millisecond is too low. But for some temperature logger, this is still more than enough. It can transmit once a day or so. What are the frequency bands? These are the frequency bands that exist in Germany. We have band 3, 7, 9 and 20. These are the classic LTE bands in Germany. To increase the capacity of these bands, the band N78 will be added. Some of you might remember the auction, but we are available more and more additionally. We can do fancy IHF stuff with it. Higher frequencies with 5G, as I said, the subcarriers can be wider. If I make the subcarriers wider, however, they need to be tested. So you may think that there are more resources per time. It's actually not right. I just marked it up in this space here. The 15 kHz subcarrier yields the yellow block with a 30 kHz subcarrier spacing. The bands are wider, but they are quicker sampled. You saw the resource grid for 4G that looked orderly. This is the resource grid for 5G. It's not quite as orderly. It's really quite a bit more complicated. The pink blocks are the SSBs. We need them for the beams. We'll get to that. The blue blocks are the physical... It's the PDS, the share channel, or broadcast channel that can go there too. It's two beams that I am showing here. It's relatively complicated, and it could make it more complicated still. If all the resources for multimedia broadcasts are in here, that would get more complicated. Or data for positioning over 5G, like GPS in-house based on 5G, etc. So we're not going to go into that. We'll take a look at the data rate here. The data rate depends on the position of the client. The client has an ND device. The ND device has noise. You'll see the noise down here in the picture. The lower the reception, the worse the SNR gets. To get a high data rate, I get a better SNR. Ideally, the ND device is right in front of the antenna, and I can do 256 QAM there. However, a lot of errors are permissible and will be corrected. In parts, over 50% of the data can be errors. And with larger distances, more error rates are permissible. From that, you can calculate data rates. I've tried here for different scenarios to calculate data rates. First of all, what you see is the orange block. 2.23 gigabits per second. That's the maximum you can get in theory. Get out of 5G in N78 bands. If you do 4x4 MIMO and disregard any physical laws. There are some more realistic data rates. First of all, we don't have 100 GHz in Germany, only 90. We can't do 4x4 MIMO everywhere. So then we move into these yellow areas. And the very last column, the 2x2 MIMO normal use and low traffic. It means you share traffic with other customers. And we get into the range of about 500 megabits per second, which the customer can get or might get under certain conditions. This is a slide from Martin. Yeah, so I've made a slide on what this means in practice. Peter doesn't like speed tests, so I made this slide. So what do we get? So you said 2.23 gigabit per second in a 100 MHz channel, if everything is optimal. But 5G is never on its own. There's always LTE present too. And so even if everything aligns, you get another gigabit per second. What I've seen in practice, if you are optimally positioned, I get 1.3 to 1.4 gigabit per second from a channel plus LTE. However, that is not really a meaningful number. This is an empty cell. This is the capacity that everyone has to share. And so to set this interrelation, I've taken a look at what the Wi-Fi here at the Congress does. Right now it's about 3 gigabits per second with 10,000 devices connected. This is a channel that can do 1.3 gigabit per second, maybe not 10,000 channels. You get an indication where we're going with 5G, not too bad. Alright, so band N78, what do we do here? We have a time division DDD and FDD systems. FDD means that the uplink where the N devices are uses a different frequency domain than the downlink. So all bands between 700 and 2.6 GHz are... The 1500 band doesn't have an uplink. An N device at this size can't send at 15 GHz because it'll compete with the GPS. TDD is band N78. When we do TDD, we send and receive on the same frequency, just like DECT. The idea, however, is that if you do TDD, the resource downlink uplink can be flexed. If there's much downlink, there's less uplink and vice versa. This would be such a structure. So we have only downlink slots, one special slot and a little bit of uplink. So the whole thing is a little bit... a lot of specifications, how this distribution of downlinks and uplinks could be made. So you could think we could share the traffic dynamics, but no. Because we don't have only one provider in the country, so there are several. So this is now difficult if I have several antennas on one tower because they are on different frequencies, but they are very close. So if one antenna is sending and the other one is receiving a few megahertz below or above, the sending would be disturbing the receiving of the antenna. So it is important that all the providers in N78 use the exactly same structure of uplinks and downlinks. So the station must have a GPS added. So the providers have to actually talk to each other. They can't do whatever they want. So the clients is a similar case. So if the orange client is sending, then the green one could not be receiving at the same time if they are lying on the same table. So this is a 5G antenna. There are few... There is no IP. This is a specific protocol. These are several antennas. The round items are the small antennas and there are also receivers on it so that the phases can be activated. So now how does this beaming work now? So that it beams in a certain direction and receives from a certain direction. I can do this with phase redirections. So you usually work with different cable lengths. So there is a split. So there is a radio field that has a certain direction. And the phasen redirector already sits within the element here. So here you can see the beams. This is a synchronization signal block, SSB. This is in this N78 band in the middle. It contains a few elements. So I will take several SSB blocks for a beam that variates in the phases. So it's like a light tower that beams in through the area with different SSB blocks. And the beams have different strengths. So after 2 milliseconds, 8 beams have been sent. And the client can actually receive those and initiate a connection. And this is now the connection establishment in the earth. So the Gino B says I take a traffic beam that lights into this direction. And I offer you a few more traffic beams. So you, the client, in this same phase area. And you tell me which of those traffic beams you can best receive. So the client is actually reporting back to the antenna which of the beams is the best for the client. So we are in the same cell. So the senders and receivers just change their phases. So you can do actually 2x2 or 4x4 MIMO with this setup. And you just use different channels. So it's quite orderly in this cell with these different beams. So the traffic is only there where actually is a target. There is no further interference in the field. So I can directly provide good quality to customers. So there are further MIMO antenna with a cable. So you have to measure the phases and you have to correct it. So the phase coupler makes a back coupling to measure the phases of the different cables. And with 5G there is then a single user MIMO possible. So one user gets data from different antenna areas. One channel is used for one client. The other channel is used for a second client. And this works in the uplink and in the downlink. So with how can you measure 5G antennas? We can use passive antennas to use normal antenna measurements. But this is more difficult with active antennas because we would have to open the entire antenna thing and check each antenna element, check if it's still there, if it is maybe got wet or etc. That's something that a system technician would have to do. It's not that trivial. But I can do high level measurements as long as the static beams are still working as they should. I assume that the antenna is more or less okay. Now let's talk about the not physical properties anymore. Let's talk about the network architecture. So I made some slides about a network, how it works. You see the slide with this nasty NSA abbreviation acronym. You see all the components are inside a 5G network. The black components are already there for LTE. And the orange ones are the new components which already exist more or less until today for 5G. In the future all black components should be replaced by 5G components. In the middle of the slide there is the core network which is divided in two parts logically. One part transports user data payload which is connected to the internet. This is done via gateways. It's done with usual routers as you know them. Software is a bit special but it's basically IP. And you have also the management stuff on the other side. Which has for instance subscription database where any subscription each client has an entry. Each customer has Cypher keys etc. So everything is based on IP in these days. So in the end we need only one cable which transports all these data to the HSS. Only one cable to a mobile antenna. So there are about 20.000 antenna elements but core elements we don't need that money. For 5G you know the 4G E node B plus also the 5G node B. But I always work together. The LTE part is always the master and the 5G part is added as a speed booster only. This is what is called non-standalone architecture. We don't have standalone 5G networks. It's something that was done because it was easier in these days to implement it like this. Usually we also need better... I call the radio space stations because it's called like this in GSM. And we also need better links from the core network to these space stations. These days we usually have gigabit fiber connections but in 5G we might have bigger data rights. So in theory we have three sectors. We have three times a data right. And we need about 10 gigabit links in these days. State of the art is 10 gigabits. You have to exchange the transceivers, the fibers stays the same. Alright so here's a couple of flow diagrams. What's happening in the network when a 5G connection is established from plane mode, from flight mode. Looks complicated. It might be actually but all of this happens under 100 milliseconds. On the left hand side there's the user equipment which is the end device, the smart phone whatever. EnoB is the 4G base station. The MME is the mobility management entity in the core network. The HSS database and the gateways on the right hand side which transport payload. So exiting the flight mode even with a 5G end device we do the 4G part of the procedure. It searches the broadcast information of nearby stations such as for the best station. Then goes to a random access procedure because with LTE and 5G things work differently than from Wi-Fi where in Wi-Fi everyone just checks whether they can send. Here it's very clear who can send and who may receive at any one time. End devices cannot just send and receive at random. There is an access procedure. The end device says hey I need a channel to tell you where I am and who I am. That's the RSE connection setup procedure. The TET request is sent. This contains all right I'm end device with number so and so. I'd like to have internet access. The 4G base station passes this to the MME which looks up a record in the database then starts the authentication and suffering procedure. So first authenticates the device where it actually is the device it claims to be. Then ciphering is switched on such that at least trivial eavesdropping attempts aren't possible. While all this is happening the location of the end device is being passed to and written into the database or at least a rough area because if later on to save battery life I take away the channel the network still needs to remember where I am so the rough location is put to the database. Now while this is happening on the left hand side the capabilities are exchanged the end device capabilities depending on the age of the end device it might be capable of less or more. This information isn't just kept in the base station in E node B but also in the MME. Depending on how many carriers can be bundled what modulation schemes are possible the highest data rates varies. The bottom part on the MME acquires an IP address from the PDN gateway the PDN gateway access the internet usually you get a private IPv4 address with NAT of course isn't so great but in the mobile network that's not too bad because at least you keep the script keys out that will drain your battery. And finally the MME will send back the initial context setup request containing the IP address which is then passed on to the end device and a so-called default bearer is initiated seen from the smartphone it looks like a logical network interface so an Android IF config might show you how a new IP interface has just been put into existence there can be several of these on an end device for the default bearer that there's a default bearer specifically for the telephone stuff that's a different interface from the IP and finally the payload can flow also a measurement config is sent so that when the signal power drops the network can decide and do a handover all of this in 100 milliseconds so this was the 4G part only now we get to 5G so when the 4G base station notices this is actually a 5G device and I have a 5G cell it takes the 5G G-Node B in addition so then measurements are done on the 5G frequencies the end device does that that goes back to the 5G G-Node B and it'll take the IP data stream over and tell the 4G G-Node B that a switch can be done a switchover can be done and then you'll get your data by both the 4G and the 5G node Bs that's why the bottom blue arrow is wider than the upper one the simultaneous reception of 4G and 5G is called simultaneous bearer because my packets come from both node Bs the 5G G-Node B simply splits the packets most of it it transmits itself and the rest goes via the X2 interface to the 4G card which uses the LTE to transmit and the end device will then recombine the packets in the end device in uplink we're not doing this as a practical matter everything is being transmitted in LTE so end device to network is doing everything where LTE or one could do 5G as well however the advantage of LTE is it's a lower frequency and so the reach is higher it reaches better however you need to then share the channel with others if you do everything in 5G you might have the channel for yourself because there's not so many end devices but however the reach is not as good and so if you are too far from the base station and it will reconfigure and use LTE for the uplink you can do both it depends with the uplink using only 4G or 5G that's actually only half true because there's acknowledgments on lower layers for the data packets that I get download data packets and uplink acknowledgments for this to happen quickly without data loss and you have to do this for both 4G and 5G because you got the data via split-barrow so your payload might go just through 4G but however acknowledgments have to go via both networks and the problem is you only have one budget for transmission power and now you have two transmitters and so each transmit only gets half of the available power that will limit your reach so this is the story about when do I actually show a 5G logo this is more complex with 3G, 4G, LTE etc the end device would usually know which network it's in with 5G, 5G is only the speed booster and if you do it just so the 4G and 5G might just switch back and forth on the display it could be confusing and so the idea was that LTE has system information called upper layer indication just so you don't know what it's for and the upper layer indication bit if that is said it means there is also a 5G cell sending and the end device if it hasn't been told it must not access 5G that's the access restricted bit it can then use that bit and display the 5G logo even though the 5G part may actually not be active and so the advantage then is whether the 5G part is there and is active or not it doesn't matter you don't toggle all the time then there is a handover scenarios if you have 4G and 5G they have to be there at the same time but the scheduler is totally independent from each other so if I do a handover from step 1 to step 2 it can happen depending on the infrastructure that you first switch to the 5G part and the 4G part still remains on the old station so I will receive my data from two different cells so my data from two different directions so if only if my client then acknowledges that the 4G can also be reached on the new cell the data will come actually both from the same cell then I have here a picture for the perspective, future perspective because of the 3.4G hertz that have only a certain range we have to increase so how do we do that so do we just remove LTE and put in 5G so this is good for the guys with new 5G devices but the ones with old devices have a problem then so another possibility to solve this how the Swisscom is doing it is called dynamic spectrum sharing so the idea is to configure 5G in the same way as 4G to have 5G and 4G in the same channel to control channels so the yellow one for 4G and the blue one for 5G so the clients can then be assigned in the different control channels so depending on the number of end devices for 5G and 4G I can split my channel the disadvantage is that I now need to control channels which costs bandwidth which costs 10-15% capacity so this is a real disadvantage only to have 5G and 4G at the same time so it seems that the first scenario might be chosen so if I then have 5G in the lower bands to make it also available outside of cities so we will not just throw away the 4G core network but we will operate them both so 5G and 4G at the same time for a certain amount of years so also the clients will then still use the 4G network and only move slowly to the 5G core network so you again have the two areas the user plane so the routers and then the mobility entity has been split into two functions so access management function and session management function and the database has been split actually in three blocks this has been done because LTE originally thought that one entity is one hardware box but now virtualization also has reached the telecom providers so for the 5G core network the telecom providers want to virtualize so there are no hardware entities anymore but functions now and that's why this has been split up so now we are nearly at the end here the good thing for 4G, 5G all the specifications are public you don't have to subscribe to anything you can just access them and don't load the specifications from there we have put on all the slides we have put references from the specifications to find those references in the publicly available specifications so now we are finished thanks very much for listening and have fun and now time for questions so we have about 10 minutes for questions the whole then come to one of the mics but we begin with the signal angel question from the internet how far should clients be apart from each other so they don't interfere with each other answer one to two meters but this case won't happen because the co-operators will operate with a constant schematic scheme don't think about it it won't be a problem on my desk there are usually four to five end devices so one meter would not be feasible but as you said they don't interfere with each other but it's of course better to put them a bit apart from each other next question the signal strength below 6 GHz is about 20 dBm so the question is about health issues on millimeter rives can we deploy it actually we always wear our 5G amulet nothing can happen to us there are regulations maximum values with maximum transmission values these are regulated up to the terahertz range and the rest is more like a religion and you can people fear things that they don't know and other people can want to make money with it and it's a difficult question short edition we talked about the sub 6 GHz range because this is what's actually rolled out right now but there's also the millimeter wave thing in 5G where for instance 30 GHz this is done in the US but the big problem is that the range is very limited so just by having a hand in front of the device or having a wall in between you have serious reception problems so nobody really began to do something with the meter rives and people in Europe are still waiting to see the experience from the US ok, microphone 4 please so what are the maximum physical speeds that end devices can move at in 5G so my example is if you take the train in other countries 4G works, it's the same with 5G from Paris to train you have 300 km per second and you still get your data right through works with 300 km per hour there's certain parameters to make the network robust against Doppler shift and also robust against jitter, these parameters are relevant for the entire cell and so downgrade the performance for the entire cell however trains not a problem to up to speeds of 400 km per hour, 400 km per hour still works so airplane actually aircraft use 500 km per hour still works alright, signal angel question from the internet please the internet wants to know about the authentication of base stations against e-node b so if you have a base station how can you authenticate with the base station if you get one well there's some partial answers usually what you do is you build a VPN tunnel between the base station and the core network so that certainly has an authentication and then the MME and the base station will also authenticate with each other I don't have the details of that however do you know so VPN tunnel first and then after that everything is encrypted so the authentication of the network against the end device was that the question the question is having the base station so if you bring your own base station someone who's like the big attacker, the bad guy well so sure you could do that however there's some barriers in place that's political so what we're showing here is the standard why there's no authentication or the authentication itself that's politics alright let's take microphone 7 please so what I would like to know is that we will be simultaneously operated why shouldn't we also use 3G at the same time we have so this and also what you said is that all operators of 5G networks need to be synchronized in their frame structure so for me as a layman that's like why that's unnecessary overhead that will cost performance I'm going to take the first part of that the lower frequencies the UMTS frequencies are already at 2.1GHz so when I said the lower frequencies that of course means the UMTS frequencies too however even UMTS frequencies have limited reach so if I go to lower bands like band 20 that's 800MHz or 1800MHz or 900MHz over a long time 5G will have to move there too and of course also in 2.1GHz it might just be easier in 2.1GHz because 3G not so many people are using so the second question why should we be synchronized well it's simply a technical requirement frequencies in the 3.6GHz space are very closely spaced and that's just technology if I sell frequency space to operators they will have to synchronize that's the limits lower frequencies don't give you any miracle speeds I mean you'll see in the slides the speeds for 5G 10MHz carriers even 700MHz might be 1.3 speed improvement okay two more minutes my question is how do private 5G networks work do they also use a mixture of 5G and 5G or is it just 5G well there's 500MHz that have been reserved when we talk about private networks that's like a campus network or a plant an industrial plant so you don't have to do a mix 4G, 5G mix you can just do a 5G as well okay one hi how much energy does all the signal processing need just without the radio so what's the power draw of the base station well that's difficult you can take a look at the internet that question comes up a lot so what the antenna sends in a 20MHz band it's 20W or 40W or maybe 100W and you have three sectors base station everything together that's 3-4kW that it needs so the actual energy transmitted is the smaller part of the energy that's used alright we're at the end thank you thanks for listening for the English translation