Unlock all 52 lessons of Network Planning for Cellular and WiFi SystemsFull course ₹40,000 · paid
5G Key Changes, dB and dBm, and Network QoS Metrics
A free lesson from Skill-Lync's Network Planning for Cellular and WiFi Systems course, on Grow by Skill-Lync · Last updated 16 Nov 2026 · Session: “Week 1 Session 2”
What will you learn in this free Network Planning for Cellular and WiFi Systems lesson?
The instructor covers 5G drivers and four changes: FR1 and FR2 bands, network slicing, edge computing and WiFi convergence. The lesson then teaches linear versus dB and dBm scales, for example 10,000 times gain as 40 dB, and defines throughput, spectral efficiency, latency, tonnage and oversubscription.
- FR1 and FR2 millimeter wave bands
- Network slicing and edge computing
- Converting gain and power to dB and dBm
- Spectral efficiency and oversubscription rate
What comes after lesson 2?
The next free lesson is lesson 3, “Cell Capacity Calculation Example and LTE Frame Structure” (33 min). Later in Skill-Lync's Network Planning for Cellular and WiFi Systems course (paid): lesson 4, Week 1 Session 4; lesson 5, Week 2 Session 1; lesson 6, Week 2 Session 2. The full course has 52 lessons (25 h 45 min of video); see the course overview or its price and syllabus.
Lesson transcript
Machine transcript of the lesson audio (lightly cleaned). Technical terms may contain errors; the video is the reference.
Continuing the recording for part two on Introduction to Telecom Industry, which is week one on the course, Network Planning for Cellular Networks. Okay, so now that we have seen how the IMT lays the foundation on the specifications and how different standards compete for it, now we have a clear understanding of how 4G was evolved. Now let's move on to and touch upon a little bit on 5G. So this is the fifth generation of standards that are seeking further improvements in the different parameters that we have been reviewing so far.
Again, the fundamental parameter here that seeks attention is the throughput, which is the capacity or mobility. We can also call it as throughput or capacity. Throughput is more like on a per-user basis and capacity is like an overall basis. And we have mobility is another aspect, which means ability to move within a cell or from one cell to the other cell. And connectivity at scale is another thing that we're looking at, like how do we develop a technology or improve the standards in such a way that it's much more scalable.
That means it's not just for the urban people, also for a rural setting. Like, you know, how do we take it at scale? So that's another thing that 5G has been considering. Latency is an important aspect, too. So far, the constraints or specifications on latency have been very minimal in the previous generations. But 5G takes a bigger leap forward in terms of latency requirements. The reason for this is because of real-time communications. Nowadays, with the Zoom calls, the video calls that we have for conferencing, it is required to have much better latency performances, which means, you know, the other side perceives one side without too much of a delay.
And there's also required in terms of gaming aspects and also when you bring in AR, VR technologies where latency becomes a key issue. So this could be a growing need. And reliability is another one since there will be vehicle-to-vehicle communications involved and overall downtime of the network needs to be reduced. So when all those things kick in, the reliability needs to be improved. And there is energy savings because 5G also caters IoT devices, which are very energy-limited or power-limited. So they need to be taken into account as well.
So as you can see, you know, the previous generations were mostly focused on throughput and capacity improvement. But 5G brings in – this is what makes 5G harder. It brings in the remaining aspects here as well into the plate and sets a very hard standard to be met. So the bar is very high, basically, okay, because of the latency requirements and reliability, especially energy saving, where you're expanding your network not just to cell phones or smartphones, but it's also for bringing in vehicles and also making it highly reliable as well.
So that's what makes 5G a little bit more challenging. And hence, that's why it's taking more time to evolve. So I pulled up the roadmap from the 5G standards and seeing, you know, just looking at their timeline as to how things have been evolving and what is planned going forward. So it started in 2017, and as you have seen, the LTE has several releases. After 9.5, we called it 4G. So similarly, after 15 or 16, roughly, you know, we started calling it as 5G. The first name that we have is non-standalone access.
It's a new access scheme that is proposed for 5G, but it says non-standalone. And what that means is it's not going to be a purely 5G base station. It's still going to rely on the 4G base station. So it would be coupled with a 4G base station. So that makes it a non-standalone access, which means it's like a starter pack, you know, with which the base stations can upgrade themselves, you know, providing a hybrid 5G and 4G type of services. In that way, 5G will rely on 4G for certain services as well.
And then slowly over time, it can be a completely standalone architecture. So a few releases were planned in 2018 that were released with NSA architecture. And towards 2019 and 2020, you know, there are further new releases planned. And as you can see, it is divided in a quarterly basis. And the package approval happens like once in a year or so, or depending on when there is closure on the agreements, on the standardization specifications. So there would be an approval process as marked here somewhere. And then we also have how we are also looking at the trend as to how things are evolving over time.
So each of them has its own list of specifications. Feel free to go through this in the 3GPP website. These are all publicly available. You can see how the progress has been made. What are the new features that have been included? Again, these are all written documents, basically. So there is no standard way to put it. It's not like a single table. They're just updated every time. It's like, yeah, every time, you know, there's something new, some addendum, some appendix, and so many things would be there.
So that's why it's beneficial to put it as a free-form documentation because so many people are collaborating. So it requires a lot of reading of this material and trying to understand it. I can guarantee that it won't be very organized. You cannot really take one document and fully understand what's happening in 5G. But it's more like a bunch of documents because there are so many people working there. And it's really hard to some sort of standardize or organize this material in a way that anybody can consume it readily.
So that's why, you know, it will be a little messy even if you look at it. But feel free to look at it and try to understand the processes, how things have evolved. And going forward in 2021 and even in 2020, there were several releases that made small changes to that. And I believe across 2020, you know, 5G deployments have started in the U.S. at least. And they're getting the data, they're getting the feedback, and they're trying to upgrade it. So it's not a fully 100% ready 5G, but it's more like in a testing phase.
But we're still there deploying it. They're marketing those devices as 5G devices. But there's a lot of data being collected, a lot of studies are being done, both by the operators and the big tech companies to understand how their devices are performing and how the cellular network is behaving and everything. So it's an evolution. It'll take a couple more years for it to fully become as robust as 4G is today. That's an introduction to 5G. But let's get a little bit more technical instead of being more administrative as to how things have evolved.
Let's become a little bit technical now in this slide and see what are the technical features at a physical layer 5G is bringing to the table. Okay. So the key upgrades here, of course, 5G in itself can be a separate course if you have to go into detail into all these things. But these are just bullet points in a nutshell, what are the key changes that 5G is bringing as an upgrade to 4G. Okay. The very first one, which is very easy to understand here, is higher frequencies.
Okay. As simple as that. So lower frequencies have certain bandwidth that are currently being used. For the 4G technology, they have certain bandwidth, certain bands that are dedicated for cellular communication, but then they're not enough. As the throughput and the capacity demands grow, they are not enough. We need more bands that have higher bandwidth configurations so that we can support more users in them. Now, that is a very hard stretch because the devices that are available to the users as of today, they do not support those frequencies.
They are operating at these older frequencies, which are much lower. But now this is a major step that 5G is taking, which is bringing in new users with new frequencies. And that requires a new devices to be manufactured. Anyway, so there are not too many frequencies. There are a couple of new bands that are added or categories of bands that are added. We call it as FR1 and FR2. So FR1, I believe it's frequency resource, one frequency resource, so it's something to do with the frequency band.
So FR1 here is a set of bands that are contained within the 6 gigahertz frequency. So 6 gigahertz is a lower frequency. So anything lower than that, we call that as FR1 band. So again, existing 4G frequencies are in FR1 already, but this is just 5G is adding a few more bands to FR1, and we call it FR1 in general. FR2 is a completely different spectrum that could be in 25, 26 gigahertz or even 35 gigahertz. So it's not fully finalized yet, but it's the other set of bands which is extremely high frequency.
We call it the millimeter wave bands. So those bands offer extremely high bandwidths ready for use. It can easily support a lot of gigabits of throughput per user, but the problem is their propagation conditions are not so reliable, which means any person walking in the middle can cause a huge loss, or any object that comes in the middle that spoils the line of sight between these two devices could cause a major problem in the propagation and because of which the link could die or there could be outage.
So this is a very tricky situation to be in. It gives you more bandwidth, more frequencies to play with, but it takes away from you the propagational reliability where you cannot rely on the channel model. So there are a lot of channel modeling studies that have been done, but the way the channels behave, the dynamic nature of the channels actually bring in more opportunities for some adaptive modulation schemes, which means depending on the propagation conditions, you switch to a different modulation scheme. So those are some of the research items that are still going on here, but that's in a nutshell FR2 on millimeter wave bands that are proposed for being used in 5G.
So another new aspect that 5G brings into the picture here is network slicing. So this is, again, a higher level change, not at a physical layer. So a quick background on network slicing is, let's just look at how things work. They're being done in today's technology. The network is a shared piece among multiple applications. So whether it be a video streaming or a gaming event or you're connecting to an industry or whoever is talking to your cellular network, the network is a common shared system, right?
So they all might have different physical layers, but then they all aggregate towards a single network layer where multiple datas are being treated in the same way. Again, these data pieces are coming from different applications, but then the way they're being treated in the network level is the same. You know, they're just a bunch of bits. But what 5G is suggesting is let the network be aware of what type of data it is routing. So the primary purpose of the network layer is routing, okay? So it just takes the data, it identifies where it needs to be routed, it puts them in a queue, or some sort of discipline is followed inside, and then things are being processed or orchestrated in a certain way.
Just that before 5G, it was all orchestrated in the same way for all types of applications, whereas 5G says that let's just slice the network and design different operations or different set of rules or disciplines for different applications. So, for example, if it's a vehicle, then it needs a certain amount of urgency, certain amount of priority, because the feedback needs to come very quickly back to the vehicle as it's moving, you know, there could be mobility involved in there. So that becomes a little bit of a challenge.
So there's a separate network design for that, and it's optimized for that. And then we have another set of network that is designed for, let's say, Netflix or some sort of a video streaming, high-quality video streaming, where it's one way. Mostly the data is flowing. Most of the data is flowing in one direction, let's say in the downlink from the base station to the user. So in that case, you know, a certain type of network architecture can be followed. So there is a different slice allocated for this HD streaming applications.
And then the other application could be something like a data center application, where certain things are being collected from data center, or it could be a live streaming application. And then there's one more category on machine-to-machine type of operations. So machines in one industry or one branch of an industry can talk to the machines in the other branch of an industry and coordinate the production. So this can be completely automated, but then that will require a different type of queuing mechanism or network sharing mechanism that needs to be optimized for that application.
So, again, that would become another slice. So this is what it means by network slicing, you know, how do you slice the network into multiple pieces and then properly allocate them for certain applications, okay? So that's a key change that 5G is bringing in, which was not available earlier. And we also have edge computing. So edge computing is another principle where the computations or certain processing can be done at the edge. So what do we mean by edge, first of all? So if you remember, you know, the network, we broke it down to two pieces, the access piece and the core piece, okay?
Again, some processing happens in the core. Not all processing happens right in the base station except for some of the communication signal processing. But some data processing happens in the core as well. Again, it can happen at a different depth in the core. It can happen at the part of the core that is very accessible to the base station, or it can happen in a core that is very much away from the base station, which means it takes a certain amount of delay for it to go and come back to get the processing done and then come back.
So edge computing is something, you know, where computing can be offloaded to the edge of the cell, which means it does not have to go through this deeper core network for the processing to happen. The processing can happen just at the edge, which means the user can ask for something and get it quickly because it's readily available at the edge. It's some sort of caching mechanism that sometimes our websites follow, right? So, you know, it knows that if you're requesting data time-to-time often, let's say if it requests, you know, a particular website time-to-time often, if you're requesting a particular website in a time-to-time basis, it can take that data and then store it offline, assuming that the data is not constantly updating.
So in that way, you know, you have it ready. You don't even need the Internet to get the data, right? So that's what we call as caching. So instead of going and searching the whole Internet for something that you want, that you are always using, and that's something that's not often changing, it's not a live event, in that case, it can be cached. That's just one analogy I can think of that's very closer to edge computing. So all these computations can happen at the edge, and that way you can avoid this up and down or to and fro journey from the cell to the deeper core.
The next part is Wi-Fi and cellular convergence. In the next week's lecture, we're going to look at how the Wi-Fi industry has evolved. Where we'll be able to fully understand what's happening on the Wi-Fi, just like how we are looking at the cellular industry in this lecture. But again, 5G is attempting to converge Wi-Fi and cellular technologies. The reason for this is twofold. So one is what we call Wi-Fi offload, which means let's say the cellular network is overloaded. You can coordinate with some existing routers and then transfer that data usage to a local router, which can access it in a different frequency.
So that will help ease the load on the 5G network. Or it can also be in a way that, you know, the same frequencies are being used in 5G and cellular as well, because there is an overlap and some bands where you can easily latch on from a 5G to a Wi-Fi band. So, yeah, overall, these are some of the key upgrades that 5G is bringing. I have put a block diagram here to help you understand this better. So you can think of 5G as this cloud here that is somewhere in between.
That's an infrastructure. OK. And as you can see, first of all, it slices it into several applications. This is a mobile broadband slice. You have an IoT slice where you have machine-to-machine type of communication between two industries. And mobile broadband can refer to some sort of communication entertainment or Internet. And then we have vehicular communication. So that can take a mission-critical slice, which means they need higher priority. They're very critical. And then there can be other slices such as, you know, fast-moving trains that need not, you know, request any critical feedback, but maybe just for tracking those trains from a central station.
So, again, there are other applications. So, again, it's not fully defined as to what these other applications are, but at least the slicing aspect is well-defined in this diagram. And the diagram below here talks about the edge computing where, you know, it need not go all the way into the data center if a device or a user requests a certain data, it can be, you know, access the data, you know, in a real-time fashion by loading the data right at the edge. So now this is, yeah, a use case that's typically followed the, you know, not fully followed these days, but it's something that's upcoming.
And it's also very helpful for machine-to-machine type of communication. It helps localize the communication instead of treating it as a global network. You know, you treat it as a local, you form local networks and then treat it there. In that way, it eases the traffic on this global scale, which can be used by the other aspects that really need them. So, yeah, just to conclude, the key changes to 5G here are high frequencies, FR1 band and FR2 band. So it's not just millimeter wave. There's also sub-six part of it, which is what we call as FR1.
And millimeter wave bands are in FR2. And we saw how millimeter wave bands affect or have a risk in terms of the propagation conditions. And then the next aspect is network slicing, which is slicing the networks for different use cases and using it accordingly and optimizing it accordingly. And then the third aspect is edge computing, which is real-time processing of the data at the edge of the network. And the last one is Wi-Fi and cellular convergence. So with that, we get a sneak peek of what's happening in 5G.
Now, you're welcome to dig deep into each of these topics. If you're interested in the physical layer, just look into the higher frequencies, how these frequencies are adding more users and more capacity to the system. If you're more of a network engineering, you can look into network slicing and see how the networks are optimized for certain applications. If you're more of a computational area that you're working on, edge computing is something that can help you there as well. Wi-Fi and cellular convergence, again, is the physical layer idea on how two networks are being converged.
So with that, we have seen the overview of different generations and how we have progressed from 1G to 5G across all these generations and what kept us driven in terms of generating these new standards and implementing them. And we saw where we are currently today and how things are being evolved. And not all the items that are discussed in the 5G slide are readily available, but they are still being under work. Now, with that being said, that gave us a good overview on these different generations.
But now let's get into some of the cellular communication fundamentals. Again, to understand the network planning fully, just these evolution, understanding this evolution is not enough. We also need to understand some of the fundamentals that are very theoretical. And in that way, we can combine both. And then we have a nice platform where we can start defining what network planning is, and then we can take it from there. So some of the fundamentals here, again, this will be really, really basic, but it's important to understand this.
Or if you might have done this in the past, but it's good to recollect that. So a linear versus logarithmic representation. We're going to see this often in our course. Oh, that's a linear answer. Also in the simulations as to when we switch from linear to a logarithmic representation. So when we have large numbers such as hundreds or ten thousands, there are so many digits there. Instead of representing them as 10 to the power of 100 and so it's really hard to say that you can say there are a million.
You can say it in those terms, but you cannot capture the accurate numbers there. So that's why it's good to follow a logarithmic scale sometimes when you're handling with large numbers. So let's say you have an amplifier of gain that amplifies the signal 12,000 times. It's really hard to say a 12,000, 12,500 times, 13,800 times, things like that. So it's better to put it in a dB scale instead of calling it as, let's say for 10,000, you can call it as, it's a 40 dB gain.
The amplifier with a 40 dB gain or equivalently in the linear scale, you can call it as amplitude with a gain of 10,000. Yeah, all these bigger numbers will be translated to two or three digits, including some decimal digits. Okay, think of it as some sort of a compression, okay, that's happening here. Again, power can also be represented in the dB scale. Normally, power, as you know, is represented in milliwatts or watts, and that's in the linear scale. So we have 100 watts, 100 milliwatts, that's all in the linear scale.
Now, this can be represented in a logarithmic scale, what we call as dBm. dB here stands for decibel and m for milliwatt. So when you take the power value in milliwatt and do a 10 log 10 of that, as shown in this formula here. So you take the power value in milliwatt and then do a 10 star log 10. Log 10 here refers to the log base 10. Then what you get is power is in dBm, so decibel for milliwatt, okay. So one example would be like 100 milliwatt is 20 dBm.
You can quickly plug 100 into this one, log 10 of 100 will give you 2, 2 times 10 is 20, so that's 20 milliwatt, sorry, 20 dBm. Here is a chart that shows how mapping is done from a linear scale to any scale, okay. If it's a linear to long scale, you can see how the compression is happening. And if it's a linear to exponential scale, you can see how the compression is happening. So, again, one more aspect here. As I said before, you know, it's not just the powers that you can represent in the log scale.
You can also represent the gain and the log scale. When we say gain, we mean how much is the output power versus input power of a system. Now, this could also be a loss. If the output power is lower than the input power, there could be a loss in the system. So, it's either gain or loss. But in general, you know, if it's output over input, we call it as gain. If it's input or output, we call it as loss. So, the gain in linear scale is given as the output power divided by the input power.
And the decibel scale, you know, you take this ratio and then do a 10 log 10 of that. That will give you the gain in dB. An example for this is like if you stick an amplifier of gain of 10,000, as I said earlier, it will result in a 40 dB gain in the dB scale. Now, there's an easy trick here. You know, if you have all zeros, you can just count the number of zeros and then you can put that number in front of a zero.
So, you have four zeros, just four and then 0, 40 dB. If you have two zeros, it's two zeros, that will be 20 dB. If you have three zeros, it's 30 dB. You have one zero, which is 10 dB. So, power gain of 10 will map to 10 dB gain. Again, if you have some other number there, it's better to use the formula. But that's an approximation, okay? That's just a rule of thumb if you have all zeros. Now, let's look at some of the QoS metrics and the KPI.
As I said before, QoS stands for quality of service and KPI stands for key performance indicators. Now, let's look at some of the performance indicators and the quality metrics that we follow in network planning. Okay. Now, we're not going to see how the network planning is done, but we're going to understand these definitions. That's it. And then later on, we can come back to network planning. So, the first metric we're going to study here is the throughput. As you already know, it's just nothing but the instantaneous or average data rate achieved in a system.
It's typically captured in MBPS or GBPS, so megabits per second or gigabits per second. The next metric is spectral efficiency. So, this is a little complex, not as easy as the throughput, but it's more like how much throughput can you push in a given bandwidth, okay? The right way to define it would be the speed of transmission achieved per hertz, okay? How much speed can you achieve on a per hertz basis, okay? Normally, this is given in bits per hertz, not MBPS per hertz, because it's a very small number.
Bits per hertz is enough to determine it. Let's say you have 10 MBPS and you're using a bandwidth of 10 megahertz, okay? Then it's 1 BPS. If you divide 10 MBPS by 10 megahertz, you would get M and M would get canceled. You will get only 1 BPS per hertz, okay? So, that's how you calculate the spectral efficiency. Area spectral efficiency is an extension to the spectral efficiency concept, but it's just that you're assessing how much bits per second per hertz is achieved in a given area.
So, you have one more division there. So, bits per second per hertz per square kilometer. Now, this can be any area. It could be square meters, but generally, square kilometer is a good assumption because generally, for cell networks, it covers kilometers of area, square kilometers of area. So, you can actually assess how much capacity or how much capability you have in a square kilometer. So, that's one way to capture that. Latency, as you might already be aware of, it's usually coupled with the throughput. It's just that when you're giving this MBPS, what latency you're giving that.
So, let's say you're transmitting a frame from one point to the other, a frame of data. It will take a certain time, a small fraction of seconds or milliseconds, typically in milliseconds, to go from one point to the other point. Now, this delay, transmission delay, could occur because of the propagation itself because it takes time for the light to travel, right, from one point to the other point. So, that itself could be a propagation delay. Or it can also be a processing delay because before it transmits any data, some sort of processing happens in a bulk of data.
So, that could add to that delay. It can also be at a network layer because the network layer, you know, it will take some time for it to route it and do some processing there. So, that will also add to it. So, multiple, you know, fractional pieces of time add up to this overall latency element and then it's typically represented in milliseconds. So, that's what we mean by latency. Then there's this other aspect on tonnage. Tonnage is basically a metric for memory as to how much memory of data is being consumed by a user.
So, it's more like, it's not a speed, but it's more like a memory capacity. So, total capacity that is generally consumed by a user. So, generally, it's represented in TB or GB. So, you might have seen some phone plans or some cellular or even Wi-Fi plans earlier a few years back where, you know, you have a limit. You know, you can use this as a certain data rate, but you can consume max of 1 TB or 1 GB per day or something like that. Or it was even 1 GB per month was some of the plans that we had a few years back.
So, it's generally rated in terms of bytes. And I'll note that this is bytes, not bits. Okay, gigabytes per month. So, that is a factor of 8 when you convert it to how many bits you need. Okay. So, yeah. So, that's your tonnage. And then you have oversubscription rate, which is basically an overestimated user consumption. Why is this necessary? Because we already have an Mbps per user. So, one user could be consuming 1.2 Mbps. But then we always have to oversubscribe it because the user will typically operate less than that.
But let's just overdimension it so that it's actually sufficient, well, sufficiently catered. So, for example, a demand of 1.2 Mbps per user with an oversubscription factor of 1.2 clearly indicates that the user is only consuming 1 Mbps. So, if a user is consuming 1 Mbps, sometimes the cell is designed for 1.2 Mbps because it has to be slightly overdesigned because in some cases, you know, they can hit this target they want. So, that's oversubscription rate. The other parameter here is the busy hour. So, this is more like a worst-case scenario of consumption.
So, typically, in a day, let's say if everybody is working from home or so during the midday, everybody could have a lot of meetings. So, there could be a lot of congestion in the network because everybody is accessing the internet at the same time. And typically, in the night, it's not a busy hour because nobody is accessing the network. Maybe very few people are accessing it. So, that could actually be a time where you can actually perform network tests rather than doing it on the busy hour.
So, that's why it's good to have this concept of busy hour versus a free hour, you know, where someone is, nobody would be likely accessing the network, whereas busy hour is everybody is likely accessing the network at the same time. So, that could put a lot of load on the network. And the reason why it's important is not just for network testing purposes, but also for ensuring that the network is not crashing because of too many users. So, we have to design the network for this type of a scenario, for this type of an hour, because the network could be designed and working very well during nighttime, but during the busy hour, it could be very bad.
So, the network needs to be designed and rated for this type of an hour. So, that's what we call as busy hour. And it's a very important metric to ask for someone, like, what is your consumption during the busy hour? You know, when you're designing a network or coming up with a network plan, it's good to ask these numbers. So, yeah. So, seller networks are typically designed to handle this level of throughput demand. Adoption rate migration is more like a time-based approach that you'll take to this network planning model, which means when you deploy a network, not everybody is going to use it right away from day one.
So, there's going to be an adoption rate there. Not everybody is going to be there. And again, not at the same time. So, there are two metrics here. So, one is a number of people. So, let's say only a percentage, small percentage of a population will buy your service. And then this percentage could. So there is a migration from other companies to your company. So let's say you're working for a company that is deploying something, and there's another company that's already existing there in a brownfield scenario.
Now it will take time for people to migrate from that older scenario to this new technology that you're doing. So you need to take that into account. So you don't have to go right away designing a network for all the users. Maybe you can design it in a smaller way and then have some room for extension. And that way you can slowly add a few more extensions or antennas to your system so that it can support more and more users. Cell loading is another phenomenon which is also referred to as how much of the cell's capability is being fully used.
So normally we want the cell, the cell could be like fully loaded in a busy hour scenario. It's generally defined as a ratio of what is the capacity that is being used to the total capacity that's available. If you're using all the capacity, then the ratio becomes one, which is very near to one. Loading would be very near to one during the busy hour because everybody's using it. During the other, how it could be 50% loaded or even 10% loaded, it could go as low as that.
But the problem with this, with full loading is that when it's operating at its full power, there will be more interference to the neighboring cells. So that becomes a very dynamic situation where when you increase your full potential, the other's potential minimizes. And then there's some sort of coordination that needs to happen to bring this down. So that's why cell loading is a good metric that comes into picture here. Now, one more parameter here at a physical layer level is SNR, signal to noise ratio. It's something that directly defines how much throughput that your system can achieve.
So the higher your SNR, the more the throughput that it can achieve. Normally, if you have an SNR, as shown in this table, it ranges from this value. It's like almost 0 dB to about 15 dB. So for each level of SNR, you have a certain modulation scheme that's associated with that. So generally, for lower SNR, you send it to QPSK, which is a lower modulation scheme that has very small data rates associated with it. Whereas when you go for higher signal to noise ratios, you can send a bigger modulation scheme that generally has higher data rates associated with it.
The noise in a system is generally a limiting factor. That's why SNR considers both signal power and the noise power. And yeah, it is the final metric that accounts for all the system losses and the noise sources when you do the link budget calculation. We also have a separate chapter that we discuss on the link budget calculations. There we'll discuss how exactly this SNR is calculated and how the planning is done based on these SNR calculations. One more extension here before we close this piece of the lecture is SNR versus SINR.
SINR is a generalization. It's a more generic definition of SNR, which is not just signal to noise, but also includes signal to interference plus the noise as a ratio. OK, now the denominator has two terms, both noise and interference. Again, you can also treat interference as a noise because it's unknown. But if it's known, you can work it out and remove that. But if it's most of the times, you cannot predict what the other user is going to transmit. So it will be treated as an unknown.
And that way you have to use in this table instead of SNR, SINR. OK, so again, it can be SNR or SINR depending on the interference. So if the interference level is too much, it's more accurate to use SINR. Again, with this SNR, as we saw in the previous table, you know, we map it to a modulation and coding scheme that determines the net capacity that is achieved by the system. Generally, every technology has an MCS SNR table. So Wi-Fi system has its own MCS table.
LTE has its own MCS SNR table. And when you have 5G, it could still be the same MCS SNR table, but with some changes. So basically, you know, based on the received SNR or SINR, a corresponding MCS is chosen that determines the quality of the network usage. So with that, I'll stop the recording for part two on week one on Introduction to Telecom Industry for the course Network Planning for Cellular Networks.