Yesterday I watched this very inspiring
live chat
that I would like to recommend to anyone who is interested in how machine
learning techniques (aka "AI") can help to optimize and troubleshoot the Radio
Access Network.
[The real contents of in the video starts at approx. 42:00 min]
My key takeaways from this fireside chat are:
Verizon Wireless has enough data
(100… 500 time series KPIs per cell) that they use to feed anomaly detection ML
algorithms and this generates a huge number of alarms, but only a few actionable
outputs.
The “big elephant” (Nick Feamster) is to identify if these alarms
indicating real problems that can and have to be fixed or if they just indicate
a new behavior of e.g. a new handset or a SW version that was not present in the
training phase of the ML algorithm and hence, its pattern is detected as a new
“anomaly”.
For Bryan Larish (Director Wireless AI Innovation, Verizon) the “big
open problem” is “that it is not clear what the labels are” and “no standard
training sets exist”.
[For more details watch the video section between 52.00
min and 57:32 min and listen to Bryan’s experience!]
In most cases Verizon seems
to need subject matter experts to classify and label these anomaly alarms due to
“the huge diversity” in data pattern.
According to Bryan only for very few
selected use cases it is possible to build an automated loop to fix the issue.
Especially the root causes of radio interference are often mechanical or cabling
issues that need manual work to get fixed.
All in all it is my personal
impression at the end of the session that anomaly detection is currently a bit
overhyped and that the real challenges and problems to be resolved start after
anomalies are detected.
Nevertheless, as Bryan summarizes: “ML is a very, very
powerful tool.”
However, strategically he seems not to see a lot of value in
anomaly detection by itself, but rather: “Can we use machine learning (results)
to change how we build networks in the future?”
I have written about the 5G positioning techniques not that long back on this blog here and on connectivity technology blog here. With Release-16 now ready for deployment, Huawei has already announced world's first in 5G Indoor Positioning. Their announcement said:
China Mobile Suzhou and Huawei reached a new milestone with the verification of the 5G indoor positioning capability in metro transport scenarios in Suzhou — a major city located along the southeastern edge of Jiangsu Province in eastern China. The verification showed that, even with pRRUs being hidden, a positioning precision of 3 to 5 m can be achieved in 90% of the platform and hall areas. This is the first time that 5G indoor positioning has been verified on live networks in the world, providing valuable experience for the commercial growth of 5G positioning in vertical industries.
Indoor location-based services are in high demand of vertical applications, such as indoor navigation, asset tracking, geofencing, logistics management, and personnel management, which reflects the huge market space of indoor positioning. Currently, indoor positioning technologies are of great variety and most of them need to be deployed and maintained individually, resulting in high end-to-end costs. As a part of the continuous evolution of 5G, positioning has been added to 3GPP Release 16 finalized in mid 2020 to realize indoor positioning by leveraging the ultra-high signal resolution empowered by 5G's high bandwidth, multi-point measurements, and multi-access edge computing (MEC) deployment.
The verification was based on Huawei's 5G digital indoor solution LampSite and leading MEC solution. The LampSite units measure the radio signals of 5G devices and work with MEC to analyze the signal characteristics. Based on the results of the analysis, leading algorithms are used to precisely locate 5G devices.
We wrote about Huawei's Lampsite on Telecoms Infrastructure blog last year here.
A group of Ericsson engineers have written a research paper on 5G positioning recently. It's available on arXiv here. Here is the abstract:
In this paper we describe the recent 3GPP Release 16 specification for positioning in 5G networks. It specifies positioning signals, measurements, procedures, and architecture to meet requirements from a plethora of regulatory, commercial and industrial use cases. 5G thereby significantly extends positioning capabilities compared to what was possible with LTE. The indicative positioning performance is evaluated in agreed representative 3GPP simulation scenarios, showing a 90 percentile accuracy of a few meters down to a few decimeters depending on scenarios and assumptions.
Definitely worth a read if you like hardcore technical papers.
We often get questions about 5G Service Based Architecture. Luckily, we have a tutorial that we can redirect people to. It's available here and the video just crossed 50K views. One of the questions that people often want to understand, is about the Application Function (AF) and how does it fit in the Applications Architecture.
To explain this, we made a tutorial. The slides and videos are embedded below. In that we have used the examples from our XR, V2X and Private Networks tutorials. All links are available at the bottom of this post.
I have been meaning to write a post on RAN functional splits and even make a video. Recently I came across multiple of these things so I am taking a shortcut by posting them here.
The first is this basic introductory video from Parallel Wireless where they explain why you need RAN splits providing examples of various functional splits for 4G and 5G mobile networks. It is embedded below:
The next one is slightly detailed video from the book "5G Radio Access Network Architecture: The Dark Side of 5G" by Sasha Sirotkin (Editor). I wrote a review of the book here and Sasha kindly made a video for our channel which is embedded below:
Finally, RCR Wireless published an article looking at the 5G functional splits in detail, by Ankur Sharma, Associate Vice President, Product Management and Strategy, Radisys. The article 'Exploring functional splits in 5G RAN: Tradeoffs and use cases' is available here.
Feel free to suggest other videos, articles, etc. in comments.
I wonder if you have seen as many adverts talking about the 5G revolution as I have. In fact I have collected many of them here. The problem is that most of these promised 5G awesomeness can only be delivered when 5G Standalone networks are launched.
Before going further, if you don't know what 5G standalone (SA) and non-standalone (NSA) networks are, then you may want to check one of my tutorials/video. For beginners here and slightly advanced version here. If you just want to learn about the 5G core, tutorial here.
I believe that the 5G Non-standalone networks are a hack that were designed mainly to show just the 5G icon and in some cases it also provided enhanced speeds. Some operators have realised this and are thinking about the 5G NSA sunset. There are some potential issues with 5G SA speeds that need sorting out though.
GSA recently held a webinar looking at the status of 5G Standalone networks. The video of the webinar is embedded at the end of the post. The webinar summarised the stats as following:
By mid-March 2021, 428 operators in 132 countries/territories were investing in 5G
176 operators in 76 countries/territories had announced they had deployed 3GPP compliant 5G technology in their live networks
Of those, a total of 153 operators in 64 countries/territories had launched one or more 3GPP-compliant 5G services
145 operators in 60 countries/territories had launched 3GPP-compliant 5G mobile services
51 operators in 29 countries/territories had launched 3GPP-compliant 5G FWA or home broadband services
For comparison, there are 807 public LTE networks worldwide
GSA has identified 68 operators in 38 countries/territories that are investing in 5G standalone for public mobile networks
Of those, a total of 7 operators in 5 countries/territories had launched 5G SA networks
Operators in China have deployed/upgraded hundreds of thousands of base stations
T-Mobile has a nationwide network
Plus China Mobile HK, Rain (South Africa) and DirecTV (Colombia)
Also ITC KSA (soft launch), STC KSA deployed, Telstra 5G core deployed, plus various contracts for 5G core systems
Private Networks, Non-public networks (NPN) and Industrial 5G Networks are also expected to make use of standalone 5G networks. As 5G networks get virtualized and open, we will see a lot more of these.
The webinar also highlighted the progress of 5G devices:
There has been rapid growth in the numbers and types of 5G devices being announced and launched
As of end February:
628 5G devices announced
404 commercially available (up from 303 at the end of November)
104 vendors
21 announced form factors
Majority are phones (306 announced, 274 commercial)
3GPP just published 2nd issue of 3GPP highlights here. (Issue 1 is here). The contents of the newsletter includes:
TECHNICAL HIGHLIGHTS
A Release 17 update
Artificial Intelligence and Machine Learning in NG-RAN: New Study in RAN3
3GPP Multimedia Codecs, Systems and Services
Is healthcare the next big thing for 5G?
From IMT-2020 to beyond
PARTNER FOCUS
PCSE - Enabling Operational Mobility for European Public Safety Responders
WBA - One Global Network with OpenRoaming(TM)
ESOA - Fulfilling the promise of Anytime, from anywhere and on any device & networks (ATAWAD)
TCCA - Trusted standards mean trusted communications
GSA - mmWave bands for 5G
NGMN - Global alignment for the benefit of end users as new focus areas emerge
5GAA - Study of Spectrum Needs for Safety Related Intelligent Transportation Systems – Day 1 and Advanced Use Cases
A LOOK INSIDE
Ensuring device compliance to standards
Release 17 timeline agreed
Initiative to remove non-inclusive terms in specifications
New Members listing
The 3GPP group structure
CALENDAR
Calendar of 3GPP meetings
NEWS IN BRIEF
In this post we are looking at the Initiative to remove non-inclusive terms in specifications. Quoting from the newsletter:
3GPP groups have started the process of replacing terminology in our specifications that is non-inclusive. The entire leadership proposed jointly a change request (CR) to the specification drafting rules (TR21.801), following an initiative led by several individual members.
In their joint proposal to the TSG SA#90-e meeting, the leaders wrote: “While there are potentially numerous language issues that could be considered offensive, there are two that are most acknowledged and focused on in the industry and applicable to the 3GPP Specifications. These terminologies are “Master / Slave” and “Whitelist / Blacklist” that are often used in 3GPP and other telecommunications technical documents.”
What next? - Change requests will now follow on any Release 17 reports and specifications that need their content brought in line with this policy.
Further reading:
SP-201042: Tdoc from the leadership - Inclusive Language in 3GPP Specifications
SP-201142: Change Request to Specification drafting rules.
SP-201143: Liaison Statement on: Use of Inclusive Language in 3GPP.
5G & Security are both big topics on this blog as well as on 3G4G website. We reached out to 3GPP 5G security by experts from wenovator, Dr. Anand R. Prasad & Hans Christian Rudolph to help out audience understand the mysteries of 5G security. Embedded below is video and slides from a webinar they recorded for us.
You can ask any security questions you may have on the video on YouTube
In my last post I described how NG RAN resources can be divided into network slices.
Now I would like to show how these network slices and the traffic they carry can be identified.
The key to this is a parameter from the NG Application Protocol (NGAP) called the Single Network Slice Selection Assistance Information (S-NSSAI). When configuring virtual network functions in NG RAN there are lists of S-NSSAI exchanged, e.g. between gNB-CU CP and AMF during NGAP Setup procedure, to negotiate which network slices have to be supported in general.
When it comes to connection establishment starting with NGAP Initial Context Setup for each PDU session that is established its individual S-NSSAI is signaled.
The S-NSSAI - as show in the figure below - consists of two parameters, the Slice/Service Type (SST - 8 bit) and the optional Slice Differentiator (SD - 24 bit). The exact format and numbering ranges are defined in 3GPP 23.003.
3GPP 23.501 defines a set of default values for SST as listed in the following table:
Slice/Service type
SST value
Characteristics
eMBB
1
Slice suitable for the handling of 5G enhanced
Mobile Broadband.
URLLC
2
Slice suitable for the handling of ultra-
reliable low latency communications.
MIoT
3
Slice suitable for the handling of massive IoT.
V2X
4
Slice suitable for the handling of V2X
services.
So when looking back at the figure it emerges that for each subscriber represented by an IMSI the SST allows to identify which services are running.
On the other hand allows to see if in which virtual network the subscriber is active. In my example I have defined that the resources are shared among a Public MNO that I consider the owner of the network hardware and two different private (campus) networks. While IMSI 1 and IMSI 2 are not allowed to use any other network slice the IMSI 3 is allowed to "roam" betweent the public slice and the two private network slices. This explains why a slice-specific authentication functionality as defined in Rel. 16 is necessary.
Back in Feb 2020, ETSI announced the launch of a new group dedicated to specifying the fifth generation of Fixed Network (ETSI ISG F5G). The press release said:
We are entering an exciting new era of communications, and fixed networks play an essential role in that evolution alongside and in cooperation with mobile networks. Building on previous generations of fixed networks, the 5th generation will address three main use cases, a full-fiber connection, enhanced fixed broadband and a guaranteed reliable experience.
For home scenarios, emerging services such as Cloud VR (virtual reality) and AR (augmented reality) video streaming or online gaming introduce the necessity for ultra-broadband, extremely low latency and zero packet loss. Business scenarios such as enterprise Cloudification, leased line, or POL (Passive Optical LAN) require high reliability and high security. Other industry sectors have specific requirements on the deployment of fiber infrastructures including environmental conditions such as humidity, temperature or electromagnetic interference.
The ETSI ISG F5G aims at studying the fixed-network evolution required to match and further enhance the benefits that 5G has brought to mobile networks and communications. It will define improvements with respect to previous solutions and the new characteristics of the fifth-generation fixed network. This opens up new opportunities by comprehensively applying fiber technology to various scenarios, turning the Fiber to the Home paradigm into Fiber to Everything Everywhere.
ISG F5G considers a wide range of technologies, and therefore seeks to actively cooperate with a number of relevant standardization groups as well as vertical industrial organizations. ISG F5G will address aspects relating to new ODN technologies (Optical Distribution Network), XG(S)-PON and Wi-Fi 6 enhancements, control plane and user plane separation, smart energy efficiency, end-to-end full-stack slicing, autonomous operation and management, synergy of Transport and Access Networks, and adaptation of the Transport Network, amongst others.
The five work items approved last week deal with:
F5G use cases: the use cases include services to consumers and enterprises and will be selected based on their impact in terms of new technical requirements identified.
Landscape of F5G technology and standards: this work will study technology requirements for F5G use cases, explore existing technologies, and perform the gap analysis.
Definition of fixed network generations: to evaluate the driving forces and the path of fixed network evolution, including transport, access and on-premises networks. It will also identify the principal characteristics demarcating different generations and define them.
Architecture of F5G: this will specify the end-to-end network architectures, features and related network devices/elements’ requirements for F5G, including on-premises, Access, IP and Transport Networks.
F5G quality of experience: to specify the end-to-end quality of experience (QoE) factors for new broadband services. It will analyze the general factors that impact service performance and identify the relevant QoE dimensions for each service.
Then in May, at Huawei Global Analyst Summit 2020 (#HAS2020), Huawei invited global optical industry leaders to discuss F5G Industry development and ecosystem construction, and launched the F5G global industry joint initiative to draw up a grand blueprint for the F5G era. The press conference video is as follows:
Then in September 2020, ETSI released a whitepaper, "The Fifth Generation Fixed Network: Bringing Fibre to Everywhere and Everything"
This looks like quite an astonishing reach by @ETSI_STANDARDS
It wants to replicate the dysfunctional, centralised & bureaucratic structures in mobile (@GSMA@3GPPLive & @ITU ) into the fibre domain
It's basically an attempted lever for more large telcos & more "core" control pic.twitter.com/spnKScbigC
In the past, the lack of a clear fixed network generation definition has prevented a wider technology standards adoption and prevented the creation and use of global mass markets. The success of the mobile and cable networks deployments, supported by clear specifications related to particular technological generations, has shown how important this generation definition is.
The focus of the 5th generation fixed networks (F5G) specifications is on telecommunication networks which consist fully of optical fibre elements up to the connection serving locations (user, home, office, base station, etc.). That being said, the connection to some terminals can still be assisted with wireless technologies (for instance, Wi-Fi®).
The main assumption behind the present document foresees that, in the near future, all the fixed networks will adopt end-to-end fibre architectures: Fibre to Everywhere.
The present document addresses the history of fixed networks and summarizes their development paths and driving forces. The factors that influence the definition of fixed, cable and mobile network generations will be analysed. Based upon this, the business and technology characteristics of F5G will be considered.
This table comparing the different generations of fixed networks is interesting too
The present document describes a first set of use cases to be enabled by the Fifth Generation Fixed Network (F5G). These use cases include services to consumers and enterprises as well as functionalities to optimize the management of the Fifth Generation Fixed Network. The use cases will be used as input to a gap analysis and a technology landscape study, aiming to extract technical requirements needed for their implementations. Fourteen use cases are selected based on their impact. The context and description of each use case are presented in the present document.
The use cases as described in the present document are driving the three dimensions of characteristics that are specified in the document on generation definitions [i.1], namely eFBB (enhanced Fixed BroadBand), FFC (Full-Fibre Connection), and GRE (Guaranteed Reliable Experience). Figure 2 shows that:
depending on the use case, one or more dimensions are particularly important, and
all dimensions of the F5G system architecture are needed to implement the use cases.
I will surely be adding more stuff as and when it is available.
I have been asked to explain in a nutshell how network slicing works in NG RAN. The most important facts you will find in the infographic below.
A network slice is virtual part of a network that offers full end-to-end connectivity for particular services and - optionally - for tenants. A tenant is a 3rd-party company that rents a virtual part of a public mobile operator's network. This allows the tenant to run its own private nation-wide mobile network without owning any hardware.
The network slicing is enabled by virtualization and all network functions can be divided into different slices as well. Thus, you can find in the figure the User Plane Function (UPF), gNB Central Unit for User Plane (gNB-CU UP) and gNB Distributed Unit (gNB-DU) all sliced.
It is also possible that a network function is dedicated for a particular network slice as in case of (gNB-) CU UP 2.
In general - and this is the benefit of the cloudification - the NG RAN is a highly dynamic environment in which additional NW functions can be added (and later released) whenever this is necessary. Mostly this will be triggered by the load on CPU and memory resources. Here comes the automation into the games that deals in large parts with load balancing. Ideally automation enables a zero-touch network management.
(click to enlarge)
However, the most precious of all RAN resources, the radio resources, cannot be administrated so flexibly and easily. Indeed, there are several automation instances that deal with radio resource management. Open RAN Alliance has defined the RAN Intelligent Controller (RIC) that is split into the Near-Realtime-RIC (RT RIC) that shall operate with a latency between 10 and 500 ms while the Non-Realtime RIC (NRT RIC) deals with non-time critical task, e.g. typical SON functions like Automatic Neighbor Reporting (ANR).
While the RIC can deal with a lot of problems there is one thing it cannot do: adding physical layer radio resources on demand. The physical resources are limited by the number of remote radio heads/antennas and as long as we have only static beamforming the physical resources covering a geographical sector are also limited by hardware and their distribution must be carefully planned. Thus, I think it is fair to say that the RIC (or a similar proprietary automation function) has to deal with the most complex situations in the RAN.
Radio resources can also be sliced in different ways. My figure illustrates a kind of slicing on the physical layer where different physical resource blocks (PRB) are allocated to different network slices.
However, this is not the only way how the resources of a cell or a beam can be sliced. Beside a split of PRBs it is also possible to slice on the MAC layer where logical channels (slice-specific radio bearers) are mapped onto transport channels or on PDPC layer as it was described and demonstrated by the 5G NORMA project (Chapter 2.1, page 17 ff.).
What in the end will be implemented by the RAN equipment manufacturers is a question I that cannot answer today.
