Lesson finished
Unlock all 27 lessons of 5G CoreFull course ₹40,000 · paid
Telco Cloud Placement and Mapping 4G Nodes to 5G Functions
A free lesson from Skill-Lync's 5G Core course, on Grow by Skill-Lync · Last updated 16 Nov 2026 · Session: “Week 1 Session 2”
What will you learn in this free 5G Core lesson?
The lesson shows where virtualised network functions sit across RAN, edge and core clouds, including UPF for multi-access edge computing. It then maps 4G elements such as SGW, PGW and EPDG to 5G functions and interfaces, and ends with 5G identifiers and SSC modes.
- Placing RU, DU, CU and UPF across edge and core clouds
- Extra 4G nodes: SCEF and ANDSF for Wi-Fi offload
- How 4G components map to 5G network functions and N interfaces
- 5G identifiers such as GUTI, SUPI and DNN, and SSC modes 1 to 3
What comes after lesson 3?
Later in Skill-Lync's 5G Core course (paid): lesson 4, Week 1 Session 3; lesson 5, Week 2 Session 1; lesson 6, Week 2 Session 2. The full course has 27 lessons (12 h 47 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.
Hello and welcome to the session. So in the last session, we had gone through various components of LTE and IMS code. And also we had seen the evolution from 2G to 3G to 4G to 5G. So in this session, we will see when we virtualize all the components. We'll start with 4G here. So when we virtualize all the components, so how we can place those components in end-to-end telco cloud. So in this diagram, we have end-to-end telco cloud. Okay. So in this diagram, we have a few components mentioned, which we have not covered so far. So some of the components are specific to OpenRAN and 5G. So that we will cover in the later sessions or slides. So for now, for your understanding, I will tell you that in OpenRAN or in 5G, So the RAM is actually disaggregated into couple of network functions. So for example, in 4G, if we consider virtual E0B. So if we compare with OpenRAN in 5G, then it is disaggregated into radio unit, RU, plus distributed unit, plus centralized unit.
So three components. So three components. And then we have some of the components of 5G, like a user plane function. So those things we will cover in later slides. Okay. So now if we want to actually place this, and we need to also consider the location as well. I mean, how far it can be deployed from the user, right? So first one is the RAN elements. So which is the actually closest to the user. So this will have your all the E0Bs. Or if we compare with 5G, then it will have the radio unit. So radio unit will actually control all the signaling towards the users. So that will be all the antennas will be present in this location.
So this is the first one. Then to this, this is not a cloud because here all these components will be based on hardware. Next is the edge data center, edge cloud, or also it is called far edge cloud depending on the implementation. So in the edge cloud, you want to keep the components which you want to be closer to the user location. So DU and CU distributed unit and centralized unit, which is part of the RAN in 5G will be deployed in the edge cloud. So as this is a cloud. So as this is a cloud, then you will have VNF manager, as we had discussed in the ITC MANO architecture. So you have a VNF manager, then you will have UPF. UPF is basically user plane function and to support the MEC use case, multi access edge computing.
So UPF will actually host. So basically UPF will have all the uplink classifiers to select if traffic has to go to the internet or it has to be terminated to the edge data center or to the caching server or to the CDN server. So for example, if we want to implement this MEC for the OTT platforms, so all these videos, for example, Netflix videos can be hosted in the CDN server, which can be located close to the edge data center. So basically that CDN server will be closer to the user point of location. So user need not to actually download it from the internet.
So that is why actually we have a UPF here as well. So this is edge data center, which can have distributed unit, centralized unit, which are part of the RAM plus user plane function for the MEC use case. Then we will have transport. Then this is the central data center. So it will have all the 4G core nodes. Whatever we discussed like MME, S gateway, P gateway and PCRF. So 4G core nodes are virtualized, not based on the containers, but these are based on the VMs.
Then we will have 5G core nodes. So these are the VNFs, virtual network function. These are the container network function. These are based on microservices like EMF, SMF, NRF, PCF. So we will cover all these nodes for now for you just to understand how deployment can happen. So this is the national data center or centralized data center. So for example, if we consider Bharti. So now they can have actually few national data center located centralized and then they can have edge data center, which is located in multiple places, which is closer to the location.
So here actually they can implement RU, DU, UPF. And in the national data center, they can host a 4G core function, 5G core function and IMS core function. So this is how it is deployed. And again, they need to have end to end security. So there is a new concept of network slicing in 5G. So network slicing is grouping of resources which can actually provide a certain characteristics or which can fulfill certain use case requirement. So for example, 3G can have different set of network functions.
4G can have different set of network functions and 5G can have different set of network functions. And in 5G, there are three slices defined in 3GPP. So first one is EMBB. So EMBB stands for enhanced mobile broadband. Then a second one is a URL CC. This is a ultra reliable low latency communication. And then third one is the MMTC massive machine type communication. So these are the three standard slice defined in 3GPP spec, which can be supported by different group of network functions. And then for controlling the RAN, there will be RAN controller, then edge data center controller, transport controller, then telco DC controller, mobile core controller and IMS controller.
And this will be end to end orchestration, which will be connected to the domain controllers. So this is a very high level view of any telco cloud deployment where customer has virtualized everything and how they can actually manage various components. So now we will break these things and we will see how these are deployed in detail. Okay. So before that, we will see what is the need for 5G. And now we will see the new concepts in 5G as part of the evolution. Okay. So these are some of the reasons why 5G is required and what are the new opportunities that comes with 5G.
So world has gone mobile. So changing customer expectation with AI and VR 10X mobile traffic growth from 2013 to 2019. Then ubiquitous access to apps and services. I mean, all the apps actually talk to internet all the time. So basically it sends and receives data almost all the time. Then traffic growth in access driven by video. So with the launch of a lot of OTT players. So now everyone actually uses OTT platform. Right. So now there is a lot of demand in video. For example, if you want to browse a video on Netflix, then definitely actually you cannot download such a heavy content all the way from internet.
So they need to actually deploy CDS servers which are closer to the user point of location. So that actually they don't consume a lot of bandwidth on the internet because obviously that will be very expensive as well. And also it will impact the user experience. So now there is a tremendous growth in access. So access is basically the location which is close to the user. Then rise of the cloud computing. So cloud computing is basically to access any computing resources or to access any deployment remotely.
Then changing SP architecture service delivery methods. Then changing enterprise business models because of the rise in the cloud computing. Then digitization leading to internet of everything. So now everything is connected to internet be it people process data things. So these are the new opportunities and to actually fulfill these opportunities or to monetize these opportunities. We need 5G. Okay. So now we will see a 5G core architecture. Okay. So let's start with what is different in 5G. So let's perform a quick comparison between what is supported in 4G and what will be supported in 5G.
So to start with latency in 4G is a 10 milliseconds while in 5G it can be less than 1 milliseconds. Then data traffic was 7.2 exabyte per month. And now it can go up to 50. You can l spots a little bit. But online the demand rate is 1 million. Your已经 has been up to 44m. Then, peak data rate in 4G, the peak data rate supported was 1 Gbps. In 5G, peak data rate supported is 20 Gbps. So, that is a massive improvement. Then, available spectrum in 4G was only 3 GHz. In 5G, it is 30 GHz. And connection density in 4G was only 1000 connections per square kilometer.
In 5G, it can go up to 1 million connections per square kilometer. So, now, we see this comparison. So, now, actually, based on the differences on these comparisons, we can have actually new use cases. So, first one is the latency to support ultra-low reliable communication. Then, this is the peak rate. So, this is for enhanced broadband. And then, for the huge available spectrum and huge connection density, actually, it will provide the massive machine type communication. So, now, we have three use cases for 5G. First one is ultra-reliable low communication. Then, enhanced broadband. And then, massive machine type communication.
So, next is frequency bands in 5G. So, even in the earlier slides, we had seen various frequency bands that is supported for 2G, 3G and 4G. So, let's perform a quick comparison with 5G. So, first one is the sub-1 GHz. So, sub-1 GHz is basically a frequency spectrum less than 1 GHz. So, basically, the same LTE spectrum can be used for 5G as well. And second one is the 1 to 6 GHz. And third one is the millimeter wave. Or this is very high frequency from 30 to 300 GHz frequency. So, as the frequency has increased so much, it means wavelength will reduce because wavelength and frequency are inversely proportional.
So, if wavelength reduces, then actually, it cannot travel far. It means, actually, it has to be applied in the small pockets because its range will be less compared to these two frequency bands. So, now, if 5G is deployed using millimeter wave, then actually, it has to be deployed as the Wi-Fi access point. So, it will be deployed in each building. Then, only actually, it will able to cater to all the devices in that locality. So, range will be very small. But at the same time, the penetration will be very good.
It means, actually, even there will be very good coverage even in the basement. But range will be small. So, this will be the new way of deploying 5G. Okay. And next one is the 5G use cases that we have already discussed. So, we have three standard 5G use case. First one is the enhanced mobile broadband EMBB. The second one is the massive machine type communication MMTC. And third one is the URL-LCC, ultra-reliable low-latency communication. Okay. So, now, we will see the evolution from 4G core to 5G core.
So, this is the 4G architecture that we have already discussed with certain changes. So, this is UE. This is a 4G RAN. E-UTRAN, evolved UTRAN. Then, we have MME that we have already discussed. Then, we have HSS. So, this is the combination of S-Gateway and P-Gateway. And with one difference. So, here, you can see that this is the control plane. S-Gateway, P-Gateway control plane. And this is the S-Gateway, P-Gateway user plane. So, there is a segregation. And this is called CUPS, control plane and user plane separation.
And CUPS is basically a path from 4G to 5G. And then, we have PCRF that we already discussed. And this is the application function. Application function, you can consider as the IMS core that we had discussed. So, only difference is we have combined S-Gateway and P-Gateway. And then, we have separated the S-Gateway, P-Gateway control plane with S-Gateway, P-Gateway user plane. And other entity is SCEF, Service Capability Exposure Function. So, basically, SCEF is used in 4G to actually expose the 3GPP services to the external device like IoT device.
So, for example, IoT device can actually ask for the user location from SCEF. And SCEF can talk to MME and HSS. And actually, it can provide the user location to the IoT devices. And this is the data network. So, this is the internet. So, now, we know all these entities. Now, we will see how these entities will evolve into 5G network function. Okay. So, before that, let's introduce a few more 4G components first. So, ANDSF, Access Network Discovery Selection Function. So, ANDSF, what it used to do in 4G is, it used to send the UV rules to the ANDSF client.
Which used to be configured on handset in order to tell the handset that when handset should select the Wi-Fi offload. For example, service providers like Bharti and also Jio, they had implemented Wi-Fi offload. I mean, it can be trusted Wi-Fi offload. That is SAMOV or untrusted was EPDG. So, basically, Wi-Fi offload is instead of using the cellular radio, use the Wi-Fi to connect to the 4G core. So, because access is provided by 4G core, just that radio is the medium to connect to the core. So, instead of using radio, actually, UV can also use the Wi-Fi of their home.
And Wi-Fi is basically connected to EPDG, Evolve Packet Data Gateway. And EPDG is connected to SP Gateway. But how handset will decide if it has to select to the Wi-Fi or it has to select to the cellular radio? So, that intelligence was provided by ENDSF. ENDSF used to send the rules to ENDSF client in 4G. But this was not very popular because this was not supported in iPhone. It was just supported in some of the Samsung devices. And as it was not so popular, even service providers did not actually implement this.
Okay. So, now, let's see how the components in 4G are evolved to 5G. So, this is UE. Then EU-TRAN will become 5G radio. And SP Gateway will become 5G user plane function. So, this has become user plane function. So, S Gateway, B Gateway user plane in 4G is equivalent to user plane function in 5G. And then this is the access network, 5G access network. It is called new NG-RAN. NG is next-gen RAN. And this is the UE. So, now, all the blue blocks are the 5G blocks.
And this is the data network. And you can see the interface names are getting changed. So, now, everything is next-gen. That is why it is called N. N is next-gen. So, interface between UPF and Internet is N6. So, these all components will evolve. This is now called G node B. Next-gen node B. G node B to UPF interface is N3. Okay. So, now, actually, this EPDG. This EPDG will evolve to N3IWF, non-3GPP interworking function. So, in 4G, we had EPDG or SAMO. Now, it is called N3IWF.
And N3IWF will have the same interface as G node B. So, N3IWF will have N2 interface towards this will become EMF and towards UPF N3. So, this will become N3IWF and N3IWF as it has the same interface. Now, we can say that this RAN supports the multi-access because it supports a cellular radio plus now it also supports the Wi-Fi. So, to support the MEC use case, multi-access edge computing. So, now, this UPF is connected to the Internet. But if UV wants to actually browse any data, for example, Netflix data.
And Netflix, they have enabled a CDN server. So, basically, it will go to the UPF. And UPF will have the uplink classifier rules. And UPF will know, okay, so, if this traffic is coming for Netflix, it will send to the UPF which is hosted in the edge data center. So, in the first slide, we had seen edge data center will have UPF as well. So, It will be hosted in the edge data center and edge data center will have all the cache for the videos. So basically, you will browse the videos directly from the UPF.
Internet is this. It is not going till internet. So UPF is connected to the internet. Internet is here. But these are not internet. These are the private connection within the service provider. So it will just go to the UPF and UPF, it will go to the cache, CDN servers, and it will download it from here. And using, it will have the uplink classifier. So now UV can actually browse few data from the CDN server for the videos. And other data can go all the way to the UPF and to the internet.
Okay, now we will see how these MME and other components in 4G will evolve to 5G. Okay, so in 4G, MME used to handle both mobility management plus session management. But in 5G, now MME's evolution is EMF. EMF is a access and mobility management function. But now EMF will just handle mobility management. Session management is with SMF. SMF is the evolution of S-Gateway and P-Gateway control plane has become session management function. And END-SF plus PCRF has evolved to PCF, policy charging function. So earlier END-SF used to send policies to UV.
Now PCF will send policies to UV. And also PCF also has PCRF. So it will send all the QS and charging rules to SMF as well. And now HSS has broken into AUSF. AUSF is authentication server. Then UDM is the unified data management. And UDR is the unified data register. And SCF has become a network exposure function. And all these network functions are actually part of the service-based architecture. It means they communicate using HTTP with JSON as the data representation. Earlier in 4G, we had different protocols like GTP, diameter, radius.
But in 4G, everything has been actually evolved to HTTP. So all these native functions, they talk to each other using HTTP REST APIs, except between SMF and UPF, this protocol is still UDP-based. And this protocol is basically evolution of GTP and it is called PFCP. So apart from this N4, all other entities, they actually talk using HTTP. And then other component that we have, that is the new component is NRF, Network Repository Function. So Network Repository Function is used to select or discover any network functions.
For example, if EMF wants to talk to SMF, so EMF will first actually register its service to NRF. And for example, if EMF wants to talk to SMF, all the network functions first will register to NRF and if any network function wants to actually consume the service of any other network function, it will go and ask NRF, I need to consume the service of SMF. So NRF will provide the IP address and port number of this SMF to EMF and then EMF can talk to this SMF using REST API.
And other component is NSSF, Network Slice Selection Function. So in case we have multiple slices deployed, then NSSF helps to identify which slice to be used. And then there are three use cases. As we discussed, EMBV, UR LCC and IoT or MMTC. And for security, as we discussed, are two different PLMN. So each network is identified by PLMN, Public Land Mobile Network. So for example, if AT&T wants to talk to Verizon, so obviously there should be some security. So SEPP, Security Edge Proxy Protection is the security between two different PLMNs.
So these are the abbreviation for the various 5G network functions. Okay. So now we will see the roaming architecture. So now we understand the architecture of 5G, but there could be home architecture and roaming architecture. So roaming architecture is basically if you are a subscriber of Bharti Airtel, India, and if you are roaming in US, then for your handset, that will be the roaming scenario. So we will see how roaming is handled. So first architecture is the home routed. So this is the visiting PLMN and this is the home PLMN.
So for example, this is US and this is India. And this UV is basically has the Airtel India seam and it is in US. So now we will see how the traffic flow will look like. So in US, this UV will connect to the NGRAN, NextGenRAN and NextGenRAN will talk to its own EMF. So these all network components or network functions belong to visiting PLMN. It could be AT&T or Verizon and this home PLMN is Bharti or Airtel. So UV will talk to NextGenRAN, to EMF, then to visiting SMF and visiting UPF.
So home routed is the most popular roaming architecture because Airtel, right? Airtel would like to still have the control on the charging even when its subscriber is actually roaming in other countries. That is why home routed, it means even if you are in any other visiting PLMN, any other country, still your traffic will flow via my network function in the home PLMN. So still, all the traffic will flow via this is AT&T or Verizon. It will come to Airtel India, SMF and UDM. So the call will be still processed by Airtel India network functions.
That is why actually Airtel will still able to charge accordingly. If actually, if call was terminated by AT&T, SMF itself, so Airtel wouldn't have got a chance to charge it accordingly. And as charging is very important, all the service providers wants to control the charging irrespective where the subscriber is roaming. But disadvantage is obviously, if you are in US, but still if call has to come to India, obviously experience will be very bad because there will be a lot of delay. And that is why actually we hear delay.
Whenever we call a US number or if someone is roaming in US, you see delay because actually mostly it is routed via home. So advantage is actually, it gives more control to the service providers for charging, but disadvantage is quality of experience for the user is not that great. So other option is local breakout or it is also called LPO. So it means actually process the call in the visiting PLMN itself. Home PLMN does not want call to come all the way to home PLMN except for some policies and authentication because obviously AT&T or Verizon in US will not store any authentication credentials for the Indian subscribers because they don't know who all will come.
So apart from policy and authentication, once those are done, session will be completely handled in the visiting PLMN in US or in the roaming PLMN. So this is called local breakout LBO. Okay. So next one is the 5G identities. So first one is the SUPPI subscription permanent identifier. So all subscribers within the 5G core are allocated globally unique 5G SUPPI. So in 4G equivalent to SUPPI is IMSI and SUPPI is generally in the form of traditional IMSI or NIAI. So International Mobile Subscriber Identity and Network Access Identifier.
And SUPPI when sent across RAN, it is never sent in clear text. So one difference in 5G is in 4G, UV used to send IMSI if it did not have any globally unique temporary ID assigned by the network. But in 5G, UV will never send its SUPPI in clear text. If it does not have any temporary ID assigned by the network, it will convert this SUPPI to SUPPI. SUPPI is the concealed identity. Subscriber concealed identity. So it will encrypt the SUPPI into SUPPI using the public key.
And public key is stored in the. So, it performs the asymmetric encryption and utilization of IMSI ensures that various roaming and interworking scenarios are being supported. Because in 4G, IMSI is used. That is why actually SUPI and IMSI can be interchanged to support the interworking scenario with 4G. And next one is the 5G GUTI. So, GUTI is the globally unique temporary ID which is used in 5G as a means by which to keep the subscribers SUPI confidential. So, GUTI is assigned by EMF. When UV first registers to the network and it is the temporary identity. So, it might change.
So, EMF allocates the 5G GUTI at the time of network registration which is comprised of the GUAMI Global Unique EMF ID and 5G TIMSI Temporary Mobile Subscriber Identity. So, this is the format. So, 5G GUTI has got two components. First is the GUAMI Global Unique EMF ID. So, this uniquely identifies the EMF which has assigned this GUTI to the UV. And other component is 5G TIMSI which uniquely identifies the UV to which this GUTI has been assigned. And then GUAMI is broken into MCC MNC, then EMF Region ID, EMF Set ID and EMF Pointer.
Then GUAMI serves as a global unique identifier for each EMF within the 5G network. Between different regions EMF Set ID's can be duplicated. So, entire one is GUAMI, then inside that we have MCC MNC, then this is EMF Region 1, EMF Region 2. Inside EMF Region 1, there could be multiple EMF Set and inside EMF Set, there could be EMF Pointer. So, this is the summary of all the 5G identifiers. So, first one is 5G GUTI that we already covered, then GUAMI, TIMSI, SUPPI, then we have PEI, Permanent Equipment Identifier equivalent to IMEI, NLT.
Then GYPSI is equivalent to MSISDN which is the 10 digit number for the handset or for the subscriber. Then SUKI is subscription concealed ID. So, SUKI is basically encrypted version of SUPPI. Then DNN is equivalent to APN in LTE. DNN is data network name. So, one thing is in 5G, subscriber never sends SUPPI in clear text. In 4G, subscriber can send IMZ in clear text. So, if it does not have this global unique temporary ID assigned, then it will send SUKI, subscription concealed ID. Okay, next one is the, again this is new in 5G, session and service continuity, SSC.
So, in 5G, there could be three SSC mode, SSC 1, 2 and 3. So, one is basically if UE has SSC mode 1, so IP anchor remains stable to provide continued service towards UE even after its location is updated. So, in 5G, we can have use case for URLCC, ultra reliable low latency communication. It means UPF has to be close to the UV point of location. So, for example, if this is UV and IP1 and it is actually moving away from this UPF. So, now, if actually there is a requirement to support URLCC use case, then actually it is good idea to actually provide the IP address from a different UPF which is close to the UV location.
So, in that case, SSC can be 2 or 3. It means IP can change. But if SSC mode is 1, even if it is moving away from this UPF, now it has come close to other UPF. Still, actually SSC IP will never change. So, in 4G, this was the case. IP used to never change in 4G. Even if UV used to move away from the network. But in 4G, as we have ultra low latency scenarios, IP can change. If it is 1, it will not change.
If it is 2, it will first break and make. If it is 3, it will first take a new IP, then release the old IP. So, these are 3 SSC mode. So, 2 and 3 are new to 5G. And in 4G, there was no SSC mode because only one was supported. And this is the DNN. DNN, as we discussed, it performs the same function and follows the same format of APN, access point name. So, basically, it has got two components, APN network identifier and APN operator identifier.
That's all in this session. And we will go through the next topics in the next session. Thank you.