Hey Nakul, I think you said that Massive MIMO can have a lot of antenna, like 500 of them. But I have a feeling that my phone probably cannot have 500 antennas, can they? That's absolutely right. Your phone is very unlikely to have 500 or so antennas. In fact, your 5G phone at least in the first couple of generations, will have maybe four or eight antennas. But that is okay. Technical reason is that 5G algorithms have been designed, assuming that the phone will have far fewer antennas than the base station or the cell phone tower work so that works out. But another reason behind why your phone cannot afford to have 500 or so or antennas is because every antenna need certain energy to operate. You have to feel that antenna with certain battery power and because your phone operates on a limited battery supply, and we're trying to make phones go longer and longer on the same charge. If you were to hypothetically try to operate 500 antennas on the same battery power, you would run out of batteries before you knew it. Think about it this way. Let's say you have one flashlight to operate that, you need one battery. But if you had another flashlight to operate that you wouldn't need another battery. If you had n flashlights you would need n batteries, so the total power you will require in our example would be somewhat proportional to the number of flashlights you have. Just like that, the amount of power phone would require in order to operate in a Massive MIMO mode would be roughly proportional to the number of antennas it has. That is another practical reason why it makes sense to keep the number of antennas on your phone to a minimum and in the first couple of generations, it'll be somewhere between 4 and 8. I see. You also mentioned beamforming. When there's like 5G network might be sending a lot of different beams to send data and information to my phone. How does my phone know how many beams are coming in or which one should I be recepting or listening to? Another good question. Your phone technically doesn't need to know how many beams there are being transmitted by the network. As long as it gets at least one beam that is strong enough with the data that the UE is requesting. Think of it this way. Let's say you are standing in the middle of a city square. You want to go from point A to point B and you call a taxi company, the taxi company might have 100 different taxis out there. It doesn't really matter to you which of those 100 cars shows up to ferry you from point A to point B, as long as you get at least one car that is willing to take you from A to B, that is all you need. It doesn't matter how many other taxis there are. Just like that. Once again, I'm simplifying things here a little bit. But in a similar spirit, a phone doesn't need to know how many beams the network is transmitting at the same time. As long as your phone is able to receive at least one of those beams at a signal strength that is good enough. Your phone can definitely function over 5G without knowing how many other beams there are. That said, however, there is a provision using which the network can let the phone know as to how many beams it is transmitting. But the phone performs some of the crucial functionalities long before the network tells the phone that, I am transmitting x number of beams right now. The phone treats that as a nice to have information rather than must-have information. So no, your phone would not need to know how many beams there are as long as it has at least one beam strong enough to provide the necessary information to the UE. Another thing that you mentioned is a beam-sweeping. Is it correct to say that gNodeB will be sending the same information at different directions at different times. Is my understanding correct here? That's another good and insightful question. I have received that question previously from people who we are trying to understand how beam-sweeping actually works in reality. We need to go back a few modules and try to recollect the difference between two types of messaging that we saw go over a cellular network. One is your data or user plane, so to speak. Data includes any pictures or videos you may be trying to upload or download, your email, your web browsing traffic. Any information essentially that you can see on your phone screen is treated as data or user plane. The signaling that happens in the background without you knowing about it between the phone and the next one to establish that data session is called control signaling and those are the two messaging types we previously learned about. It is very crucial to keep that difference in mind when we learn about beam-sweeping. At high level beam-sweeping is primarily used for some form of control information. Whereas your application their data will not be beam-swept. The way it works is the network uses beam sweeping to ascertain where your phone is, whether it is in this direction or this direction or that direction. Network users control signaling to sweep the beam to find out where your phone is located. And once it has determined with some certainty where your phone is, beyond that point, network does not need to Beam sweep for you anymore because it now knows where it is. So all your email pictures, videos, your individual data will now be sent only on a beam in that particular direction, using techniques so that the signal can only be successfully recorded by your phone and your phone alone, no other phone. So beam sweeping is not exactly applicable to your personal data, so you don't have to be worried about your personal data being showered all over the cell. That is not going to happen in a 5G network. A simpler analogy to think of it is that of doctors or other staff being paged in a hospital. Let's say you are sitting on the hospital floor and you often hear announcements like paging Dr. Smith, paging Dr. Gray, something like that. So the entire floor knows that Dr. Smith has been paged, but what message he has been paged for, he'll find out only when it gets to the front desk and says that, "I'm Dr. Smith. What is the message that you have for me?" No other person will know what that message specifically was. The only thing that they'll know is that Dr. Smith was paged for some reason but the actual message we'll remain private between the attendant and Dr. Smith. So this is one analogous way of looking at beam sweeping. Network will sweep only until it finds out where you are, but after that point, the network will only send your private signals in your particular direction. I see. So if I'm sitting here checking my 5G phone, trying to receive my email. You're saying is that the network will start sweeping and find out where I am sitting right here. And the only sending that message to that beam towards me so I can receive that message? You're absolutely correct. Got it. Then I'm just curious, do you foresee any other challenges? That's coming from massive MIMO since this sounds like a really complicated technique. It is indeed a really complicated technique to realize in practice theory sounds relatively simpler, but when you actually implement that, there a lot of complications that arise. One thing is something I've already shared with you, the consideration about power. The more antennas you have, in simplest terms, the more power you will need in order to efficiently operate those antennas. If you have, let's say 200 antennas, you will need X amount of power. But if you have 500 antennas, you will need a substantially higher amount of power. So that is one challenge regarding the implementation of massive MIMO, another challenge that I foresee is the sophistication of the signal processing algorithms that underlie the whole massive MIMO and beamforming paradigm. Those tend to be somewhat challenging to implement, validate, and optimize. So infrastructure vendors will probably need one or two generations of the product in order to fully solidify those Beam Forming and massive MIMO software and hardware challenges could be an issue as well. There is another problem that plagues massive MIMO. The actual problem is somewhat technical, but let me try and abstract that out for you. The problem is about what engineers call a reference signals or pilot signals. How can you look at it in simple terms? Well, imagine first day of classes in your school. It's the first day so when the class begins, everybody stands up and introduces himself or herself. If there are only 10 students in the class, your class-wide introductions. We'll finish very quickly. If there are 50 students in the class, everybody is going to need some time to introduce themselves, so you will need much more time than the class with 10 students. Imagine a class with 100 students, you will probably spend a good part of the first day in introducing everybody rather than trying to get to the class material on the first day. So as you increase the number of students in the class, the time and effort that it takes for everybody to introduce themselves to one another goes up in that terms, just like that. The more antennas you have, the kind of signaling network needs to exchange with the phone to introduce each antenna to the phone, so to speak, in our analogy, that amount of resources and energy are required to make that happen also scales up in proportion. So that is the third, and from engineering perspective, probably the most vexing problem about massive MIMO. So to summarize, massive MIMO, although it is incredibly powerful and full of potential, it won't be very easy to implement because A, you wouldn't need more power to have more antennas. B, software and hardware implementation has to catch up to ideas on people and the reference signals or the introduction that every antenna has to make to different phones. Data requires massive amount of resources as well, in proportion to the number of antennas. So these are some of the problems or challenges rather that massive MIMO will face in the initial phases.
