AI/ML Enhancements in 5G-Advanced for Intelligent Network Automation

Artificial Intelligence (AI) and Machine Learning (ML) has been touted to automate the network and simplify the identification and debug of issues that will arise with increasing network complexity. For this reason 3GPP has many different features that are already present in Release-17 but are expected to evolve further in Release-18. 

I have already covered some of this topics in earlier posts. Ericsson's recent whitepaper '5G Advanced: Evolution towards 6G' also has a good summary on this topic. Here is an extract from that:

Intelligent network automation

With increasing complexity in network design, for example, many different deployment and usage options, conventional approaches will not be able to provide swift solutions in many cases. It is well understood that manually reconfiguring cellular communications systems could be inefficient and costly.

Artificial intelligence (AI) and machine learning (ML) have the capability to solve complex and unstructured network problems by using a large amount of data collected from wireless networks. Thus, there has been a lot of attention lately on utilizing AI/ML-based solutions to improve network performance and hence providing avenues for inserting intelligence in network operations.

AI model design, optimization, and life-cycle management rely heavily on data. A wireless network can collect a large amount of data as part of its normal operations. This provides a good base for designing intelligent network solutions. 5G Advanced addresses how to optimize the standardized interfaces for data collection while leaving the automation functionality, for example, training and inference up to the proprietary implementation to support full flexibility in the automation of the network.

AI/ML for RAN enhancements

Three use cases have been identified in the Release 17 study item related to RAN performance enhancement by using AI/ML techniques. Selected use cases from the Release 17 technical report will be taken into the normative phase in the next releases. The selected use cases are: 1) network energy saving; 2) load balancing; and 3) mobility optimization.

The selected use cases can be supported by enhancements to current NR interfaces, targeting performance improvements using AI/ML functionality in the RAN while maintaining the 5G NR architecture. One of the goals is to ensure vendor incentives in terms of innovation and competitiveness by keeping the AI model implementation specific. As shown in Fig.2 (on the top) an intent-based management approach can be adopted for use cases involving RAN-OAM interactions. The intent will be received by the RAN. The RAN will need to understand the intent and trigger certain functionalities as a result.

Ericsson White Paper - 5G Advanced: Evolution towards 6G - https://t.co/9ZKnC8PDxX#Free5Gtraining #Ericsson #5G #B5G #5GAdvanced #6G #5GNR #5GC #Timeline #XR #RedCap #NetworkAutomation #EnergySavings #DeterministicNetworking #IoT pic.twitter.com/eeI4SKvIRC

— Free 5G Training (@5Gtraining) July 21, 2022

AI/ML for physical layer enhancements

It is generally expected that AI/ML functionality can be used to improve the radio performance and/or reduced the complexity/overhead of the radio interface. 3GPP TSG RAN has selected three use cases to study the potential air interface performance improvements through AI/ML techniques, such as beam management, channel state information feedback enhancement, and positioning accuracy enhancements for different scenarios. The AI/ML-based methods may provide benefits compared to traditional methods in the radio interface. The challenge will be to define a unified AI/ML framework for the air interface by adequate AI/ML model characterization using various levels of collaboration between gNB and UE.

AI/ML in 5G core

5G Advanced will provide further enhancements of the architecture for analytics and on ML model life-cycle management, for example, to improve correctness of the models. The advancements in the architecture for analytics and data collection serve as a good foundation for AI/ML-based use cases within the different network functions (NFs). Additional use cases will be studied where NFs make use of analytics with the target to support in their decision making, for example, network data analytics functions (NWDAF)- assisted generation of UE policy for network slicing.

If you are interested in studying this topic further, check out 3GPP TR 37.817: Study on enhancement for data collection for NR and ENDC. Download the latest version from here.

Related Posts: 

• The 3G4G Blog: What Is the Role of AI and ML in the Open RAN and 5G Future?
• The 3G4G Blog: 3GPP Release-18 Work Moves Into Focus as Release-17 Reaches Maturity
• The 3G4G Blog: NWDAF in 3GPP Release-16 and Release-17
• The 3G4G Blog: 5G-Advanced Flagship Features
• The 3G4G Blog: 3GPP's 5G-Advanced Technology Evolution from a Network Perspective Whitepaper
• Free 6G Training: Qualcomm Pushing the boundaries of AI research 

What Is the Role of AI and ML in the Open RAN and 5G Future?

Artificial Intelligence and Machine Learning have moved on from just being buzzwords to bringing much needed optimization and intelligence in devices, networks and infrastructure; whether on site, on the edge or in the cloud.

New Study in 3GPP RAN3 on Artificial Intelligence and Machine Learning in NG-RAN - https://t.co/o4o4kq6Wpp via @3GPPLive #Free5Gtraining #3GPP #5G #ArtificialIntelligence #MachineLearning #AIML #RAN #Algorithms #SON #5GC #5GS pic.twitter.com/pvPlG9HpPC

— Free 5G Training (@5Gtraining) October 15, 2021

Qualcomm has been very active in talking about AI/ML in webinars and on their site. A detailed blog post looking at 'What’s the role of artificial intelligence in the future of 5G and beyond?' is available here. It was posted in time for a Light Reading webinar where Gabriel Brown, Principal Analyst – Mobile Networks and 5G, Heavy Reading and Tingfang Ji, Senior Director, Engineering - Wireless R&D, Qualcomm discuss the topic. The video is embedded below and slide deck is available here.

Standardizing AI/ML in cellular communication systems. From #Qualcomm presentation, The essential role of AI in the 5G future - How machine learning is accelerating wireless innovations in the new decade and beyond - https://t.co/fwxtoKIZKR#Free5Gtraining #3G4G5G #5G #AI #ML pic.twitter.com/YS5w8KXQuw

— Free 5G Training (@5Gtraining) September 27, 2021

Louis Scialabba, Senior Director of Marketing at Mavenir, looking at AI and Analytics spoke at Layer 123 conference on the topic, 'AI/ML for Next Gen 5G Mobile Networks'. His talk is embedded below and a blog post by him on the topic, 'The RIC Opens a New World of Opportunities for CSPs' is available here.

Related Posts: 

• Free 6G Training: Qualcomm Pushing the boundaries of AI research
• The 3G4G Blog: 3GPP Release-18 Work Moves Into Focus as Release-17 Reaches Maturity
• The 3G4G Blog: 5G-Advanced Flagship Features
• Connectivity Technology Blog: 5G Indoor Precise Positioning
• The 3G4G Blog: ITU Standardization Bureau on Machine Learning for 5G
• The 3G4G Blog: AI your Slice to 5G Perfection
• The 3G4G Blog: Qualcomm Explains 5G Millimeter Wave (mmWave) Future & Integrated Access and Backhaul (IAB)
• The 3G4G Blog: Qualcomm Demoes Sub-band Half Duplex (SBHD)
• Connectivity Technology Blog: Super Uplink to improve 5G Coverage and Speeds
• Free 6G Training: One XR Device to Rule Them All! 

5G Non IP Data Delivery and Lightweight M2M (LwM2M) over NIDD

Earlier this year, MediaTek had announced that its MT2625 NB-IoT chip has been validated for LwM2M over NIDD on SoftBank Corp.’s cellular network across Japan. This achievement marks the first global commercial readiness of LwM2M over NIDD; a secure, ultra-efficient IoT communications technique that is being adopted by operators worldwide. The benefits of LwM2M over NIDD include security improvements, cost-efficient scalability and reduced power consumption.

LwM2M over NIDD is a combination of the communication technology "NIDD (Non-IP Data Delivery)" that does not use an IP address in LTE communication NB-IoT for IoT and the device management protocol "LwM2M (Lightweight M2M)" advocated by the Open Mobile Alliance. It's been a while since I wrote about Open Mobile Alliance on this blog. OMA SpecWorks is the successor brand to the Open Mobile Alliance. You can read all about it here.

OMA SpecWorks’ LightweightM2M is a device management protocol designed for sensor networks and the demands of a machine-to-machine (M2M) environment. With LwM2M, OMA  SpecWorks has responded to demand in the market for a common standard for managing lightweight and low power devices on a variety of networks necessary to realize the potential of IoT. The LwM2M protocol, designed for remote management of M2M devices and related service enablement, features a modern architectural design based on REST, defines an extensible resource and data model and builds on an efficient secure data transfer standard called the Constrained Application Protocol (CoAP). LwM2M has been specified by a group of industry experts at the OMA SpecWorks Device Management Working Group and is based on protocol and security standards from the IETF.

You can get all the LwM2M resources here and the basic specs of 'Lightweight M2M 1.1: Managing Non-IP Devices in Cellular IoT Networks' here.

The 5G Americas whitepaper 'Wireless Technology Evolution Towards 5G: 3GPP Release 13 to Release 15 and Beyond' details how Current Architecture for 3GPP Systems for IOT Service Provision and Connectivity to External Application Servers. It also talks about Rel-13 Cellular IoT EPS Optimizations which provide improved support of small data transfer over control plane and user plane. Control Plane CIoT EPS Optimization transports user data (measurements, ID, status, etc.) via MME by encapsulating user data in NAS PDUs and reduces the total number of control plane messages when handling a short data transaction. Control Plane CIoT EPS optimization, designed for small infrequent data packets, can also be used for larger data bursts depending in UE Radio capability.

User data transported using the Control Plane CIoT EPS Optimization, has special characteristics, as different mobility anchor and termination nodes.

Therefore, the Preferred Network Behavior signaling must include information on:

• Whether Control Plane CIoT EPS optimization is supported
• Whether User Plane CIoT EPS optimization is supported
• Whether Control Plane CIoT EPS optimization is preferred or whether User Plane CIoT EPS optimization is preferred

These optimizations have enabled:

• Non-IP Data Delivery (NIDD) for both: mobile originated and mobile terminated communications, by using SCEF (Service Capability Exposure Function) or SGi tunneling. However, it has to be taken into account that Non-IP PDUs may be lost and its sequence is not guaranteed
• For IP data, the UE and MME may perform header compression based on Robust Header Compression (ROHC) framework
• NB-IoT UE can attach but not activate any PDN connection
• High latency communication handled by the buffering of downlink data (in the Serving GW or the MME)
• SMS transfer
• EPS Attach, TA Update and EPS Detach procedures for NB-IoT only UEs, with SMS service request
• Procedures for connection suspend and resume are added
• Support for transfer of user plane data without the need for using the Service Request procedure to establish Access Stratum context in the serving eNodeB and UE

When selecting an MME for a UE that is using the NB-IoT RAT, and/or for a UE that signals support for CIoT EPS Optimizations in RRC signaling, the eNodeB’s MME selection algorithm shall select an MME taking into account its Release 13 NAS signaling protocol.

Mpirical has a nice short video explaining 5G Non IP Data Delivery. It is embedded below.

IoT has not taken off as expected and prophesised for years. While the OMASpecWorks is doing some fantastic work by defining simplified approach for IoT deployment, its current member list doesn't have enough operators to drive the uptake required for its spec adoption. They would argue that it doesn't matter how many members there are as the NIDD approach is completely optional and over-the-top. Let's wait and see how it progresses.

Related Posts:

• 3G4G: 5G Service Based Architecture (SBA)
• The 3G4G Blog: Mobile Initiated Connection Only (MICO) mode in 5G System
• The 3G4G Blog: Embedded SIM (eSIM) and Integrated SIM (iSIM)
• The 3G4G Blog: 5G and Industry 4.0
• 3G4G: Internet of Things (IoT) and Machine-2-Machine (M2M)
• The 3G4G Blog: M2M vs IoT
• The 3G4G Blog: What is Industrial IoT (IIoT) and how is it different from IoT?
• The 3G4G Blog: New 3GPP Release-17 Study Item on NR-Lite (a.k.a. NR-Light) 

Telefonica: Big Data, Machine Learning (ML) and Artificial Intelligence (AI) to Connect the Unconnected

Earlier, I wrote a detailed post on how Telefonica was on a mission to connect 100 Million Unconnected with their 'Internet para todos' initiative. This video below is a good advert of what Telefinica is trying to achieve in Latin America


I recently came across a LinkedIn post on how Telefónica uses AI / ML to connect the unconnected by Patrick Lopez, VP Networks Innovation @ Telefonica. It was no brainer that this needs to be shared.

Big data for social good from Patrick Lopez

In his post, Patrick mentions the following:

To deliver internet in these environments in a sustainable manner, it is necessary to increase efficiency through systematic cost reduction, investment optimization and targeted deployments.

Systematic optimization necessitates continuous measurement of the financial, operational, technological and organizational data sets.

1. Finding the unconnected

The first challenge the team had to tackle was to understand how many unconnected there are and where. The data set was scarce and incomplete, census was old and population had much mobility. In this case, the team used high definition satellite imagery at the scale of the country and used neural network models, coupled with census data as training. Implementing visual machine learning algorithms, the model literally counted each house and each settlement at the scale of the country. The model was then enriched with crossed reference coverage data from regulatory source, as well as Telefonica proprietary data set consisting of geolocalized data sessions and deployment maps. The result is a model with a visual representation, providing a map of the population dispersion, with superimposed coverage polygons, allowing to count and localize the unconnected populations with good accuracy (95% of the population with less than 3% false positive and less than 240 meters deviation in the location of antennas).


2. Optimizing transport

Transport networks are the most expensive part of deploying connectivity to remote areas. Optimizing transport route has a huge impact on the sustainability of a network. This is why the team selected this task as the next challenge to tackle.

The team started with adding road and infrastructure data to the model form public sources, and used graph generation to cluster population settlements. Graph analysis (shortest path, Steiner tree) yielded population density-optimized transport routes.


3. AI to optimize network operations

To connect very remote zones, optimizing operations and minimizing maintenance and upgrade is key to a sustainable operational model. This line of work is probably the most ambitious for the team. When it can take 3 hours by plane and 4 days by boat to reach some locations, being able to make sure you can detect, or better, predict if / when you need to perform maintenance on your infrastructure. Equally important is how your devise your routes so that you are as efficient as possible. In this case, the team built a neural network trained with historical failure analysis and fed with network metrics to provide a model capable of supervising the network health in an automated manner, with prediction of possible failure and optimized maintenance route.

I think that the type of data driven approach to complex problem solving demonstrated in this project is the key to network operators' sustainability in the future. It is not only a rural problem, it is necessary to increase efficiency and optimize deployment and operations to keep decreasing the costs.


Finally, its worth mentioning again that I am helping CW (Cambridge Wireless) organise their annual CW TEC conference on the topic 'The inevitable automation of Next Generation Networks'. There are some good speakers and we will have similar topics covered from different angles, using some other interesting approaches. The fees are very reasonable so please join if you can.

Related posts:

• Internet para todos: Telefonica and Parallel Wireless on a mission to connect 100 Million Unconnected
• Automating the 5G Core using Machine Learning and Data Analytics
• ITU 'Network 2030': Initiative to support Emerging Technologies and Innovation looking beyond 5G advances
• Artificial Intelligence - Beyond SON for Autonomous Networks
• Twitter et al. for Small Cell Planning

5G Network Architecture, Design and Optimisation - Jan 2018

Prof. Andy Sutton, Principal Network Architect, Architecture & Strategy, TSO, BT, provided an update on 5G Network Architecture & Design last year which was also the most popular post of 2017 on 3G4G blog. This year again, he has delivered an update on the same topic at IET '5G - State of Play' conference. He has kindly shared the slides (embedded below) that are available to download from Slideshare.

5G Network Architecture, Design and Optimisation from 3G4G

The video of this talk as follows:


There are many valuable insights in this talk and the other talks from this conference. All the videos from the IET conference are available here and they are worth your time.

Related Links:

• [Video] The UK’s first end-to-end 5G network test - EE
• [Video] 5G Network Architecture Options - 3G4G
• [Video] Beginners: 5G Terminology - 3G4G
• 3GPP-VRIF workshop on Virtual Reality Ecosystem & Standards in 5G
• Ultra Reliable Low Latency Communications (URLLC) 2017 Conference summary

EE's vision of Ultra-Reliable Emergency Network

Many of my readers would be aware that UK is probably the first country to have decided to move its emergency services network from an existing TETRA network to a commercial LTE network operated by EE.

While some people have hailed this as a very bold move in the right direction, there is no shortage of critics. Around 300,000 emergency services users will share the same infrastructure used by over 30 million general users.

The following is from an article in Wireless Magazine:

Steve Whatson, deputy director Delivery, Emergency Services Mobile Communications Programme (ESMCP) – the organisation within the UK Home Office procuring ESN – assured delegates that ESN will match the existing dedicated Airwave emergency services communication network in terms of coverage for roads, outdoor hand portable devices and marine coverage. Air to ground (A2G) will extend its reach from 6,000ft to 12,000ft.

Whatson also pointed out that coverage is not one single piece, but will comprise a number of different elements, which all need to mesh into one seamless network run by the ESN Lot 3 Mobile Services (main 4G network) provider – EE.

This includes: EE’s main commercial 4G network; Extended Area Services (hard-to-reach areas of the UK where new passive sites are to be built under a separate contract and then equipped with EE base stations); air-to-ground; London Underground; Crossrail; marine coverage (to 12 nautical miles); and special coverage solutions.

EE is currently rolling out new 4G sites – it will eventually have some 19,500 sites – and is upgrading others with 800MHz spectrum, which propagates over longer distances and is better at penetrating buildings than its other 4G spectrum holdings. Crucially for ESN, it is also switching on a Voice over LTE (VoLTE) capability, starting with the UK’s main cities.
...
Mission critical networks must be always available and have levels of resilience far in excess of commercial networks. Speaking exclusively to Wireless in early May, Tom Bennett, group director Technology Services, Architecture & Devices at EE, said: ‘We already achieve a very high availability level, but what the Home Office was asking for effectively was about a 0.3% increase against our existing commercial availability levels.

‘Now for every 0.1% increase in availability there is a significant investment because you are at the extreme top end of the curve where it is harder and harder to make a noticeable difference.’

There are very specific requirements for coverage and availability of the ESN network for the UK road system. Bennett says: ‘Mobile is based on a probability of service. No more than 1% of any constabulary’s roads are allowed to be below 75% availability, and on major roads it is 96% availability. A coverage gap in this context is no more than 1km.’

The current Airwave network has approximately 4,000 sites, many with back-up generators on site with fuel for seven days of autonomous running if the main power is cut, along with a range of resilient backhaul solutions.

Bennett says that out of EE’s 18,500 sites it has about the same number of unique coverage sites (ie. no overlapping coverage) – around 4,000. ‘As part of our investment programme, those unique coverage sites will need a significant investment in the causes of unavailability – ie. resilient backhaul and back-up batteries.’

He explains that EE has undertaken a lot of analysis of what causes outages on its network, and it has combined that data with the Home Office’s data on where the natural disasters in the UK have occurred over the past 10 years.

From this, EE is able to make a reasonable assessment of which sites are likely to be out of action due to flooding or other disasters for more than three or four days. ‘For those sites – and it is less than 4,000 – you need generators too, because you may not be able to physically access the sites for some days,’ says Bennett.

For obvious reasons, the unique coverage sites are mostly in rural areas. But as Bennett points out, the majority of cases where the emergency services are involved is where people are – suburban and urban areas.

‘In these areas EE has overlapping coverage from multiple sites to meet the capacity requirements, so if a site goes down, in the majority of cases we have compensation coverage. A device can often see up to five tower sites in London, for example,’ he says.

Having adequate backhaul capacity – and resilient backhaul at that – is vital in any network. Bennett says EE is installing extra backhaul, largely microwave and fibre, but other solutions will also be used including satellite and LTE relay from base station to base station – daisy chaining. On 9 May 2016, EE announced a deal with satellite provider Avanti to provide satellite backhaul in some areas of the UK.

Additional coverage and resilience will be offered by RRVs (rapid response vehicles), which EE already has in its commercial network today, for example, to provide extra capacity in Ascot during the racing season.

‘We would use similar, although not exactly the same technology for disaster recovery and site/service recovery, but with all the backhaul solutions,’ says Bennett. ‘Let’s say we planned some maintenance or upgrade work that involved taking the base station out for a while.

‘We’d talk to the chief inspector before the work commences. If he says, there’s no chance of doing that tonight, we can put the RRV there, and provided we maintain coverage, we can carry out the work. RRVs are a very good tool for doing a lot of things.’

At the British APCO event, Mansoor Hanif, director of Radio Access Networks at EE said it was looking at the possibility of using ‘airmasts’ to provide additional coverage. Meshed small cells, network in a box and repeater solutions are becoming available for these ‘airmasts’, which will provide coverage from balloons, or UAVs – tethered drones with power cables and optical fibre connected to them.

Mansoor Hanif, Director of RAN at EE gave a presentation on this at Critical Communications World 2016 and has also given an interview. Both are embedded below.

Ultra-Reliable Networks – A Mobile Operator Perspective from 3G4G

Feel free to let me know if you believe this will work or not and why.

Case Study: RAN Sharing in Poland

The last post on Network sharing by NEC was surprisingly popular so I thought its worth doing a case study by Orange in Poland on how they successfully managed to share their network with T-Mobile. Full presentation embedded as follows:

Deployment of RAN Sharing in Poland from Zahid Ghadialy

Radio Challenges and Opportunities for Large Scale Small Cell Deployments

Here is another presentation by Keima at our Small Cells Event. Their last presentation on "Twitter et al. for Small Cell Planning" was a hit and has racked up over 9000 views on slideshare.

Radio Challenges and Opportunities for Large Scale Small Cell Deployments from 3G4G

Integration of Wi-Fi with Cellular Networks

Integration of Wi-Fi with Cellular Networks

View more presentations from Zahid Ghadialy

Presented by Tiago Rodrigues in the ITM Optimisation event in Prague on 18th April 2012.

There is another paper by NSN on the same topic that is available here. 

LTE deployment and optimisation challenges

Presented in the 3G, HSPA, LTE Optimisation conference, April 2012 by Ljupco Jorguseski. The ICIC presentation referred to in this presentation is available in an earlier post here.

LTE deployment and optimisation challenges

View more presentations from Zahid Ghadialy

Signalling Load per device and OS

From the presentation by Martin Prosek, Telefonica, Czech Republic in 3G Optimization Conference 2012, Prague.

Signalling can cause many issues:

In the mobile device, Frequent PDP-context establishment is known to drain the battery. Battery life can be improved by supporting fast dormancy in network.

In the network, Signalling flood can create situations reminding DoS attacks. Increased signalling in RAN can cause impacts in core network:

• Radius/Diameter interface overload of AAA servers
• DHCP IP address pools exhaustion

Summary of tweets from #ITMoptimisation

Here is the summary of tweets from the Optimisation conference - #ITMoptimisation for those who missed them:

DAY 1

@patricksteemers: LTE optimization conference started. Kim Larsen kicks it off #TME

@patricksteemers: a typical user spends 80% of mobile data traffic on just 3 cells #TME

@zahidtg: After introducing CELL_PCH the PS Signalling reduced by 50% in TMo Netherlands

@zahidtg: Nice talk by @kimklarsen

@zahidtg: The network state equation by @kimklarsen

@zahidtg: Martin Prosek has already set the stage for my presentation and breakfast briefing tomorrow.


@zahidtg: Signalling load per OS in TEF CZ

@KimKLarsen: pdp context parking helped Telefonica O2 Czech in reducing signaling and improve pdp connection success rate!


@KimKLarsen: Nich Waegner #Qualcomm often overlooked! advanced receivers greatly improve the network capacity and performance.


@zahidtg: Smart WiFi offload

@zahidtg: Migrating from 2G to LTE directly sounds interesting


@KimKLarsen: Dirk Schoeneboom is setting the scene for GSM to LTE migration!


@zahidtg: Me -> Its always tricky going from 2G to LTE considering that few devices are available and have quite a few bugs


@KimKLarsen: LTE TDD 2.3/2.6 top runner for main LTE band (i.e., China & India) in terms of LTE users!

@kitkilgour: @KimKLarsen @zahidtg do you have more information on PDP context parking? what it is, how it works?


@zahidtg: @kitkilgour Will have to check with Martin Prosek as I dont know much about it as well (Cc @KimKLarsen)


@KimKLarsen: Dirk Schoeneboom recommends first launching 3G, prior to refarming 2G band to LTE as likely dormant 3G terminal in base.


@zahidtg: Telefonica UK has already refarmed 900MHz from GSM and moved to 3G in the city centres mainly. In future they see 4G on 900MHz as well


@KimKLarsen: Dirk S>Really important to know the terminal mix in the base! Launching 3G today could have substantial 3G base from start!


@zahidtg: In 2010: 75% of 2G data traffic was from 3G devices, mainly to save battery life - Robert Joyce, Telefonica UK


@KimKLarsen: Rob Joyce, deploying UMTS @ 900MHz is between 3 to 2 times cheaper to deploy per unit provided capacity.


@KimKLarsen: TF UK uses their EGSM band for their UMTS 900 (as this was not highly utilized ... Still took 18 month to refarming)


@KimKLarsen: TF UK launch 3G 900 Jan 11 and expect to extend to full network by 201X ;-) X>2012 but maybe also <2015 (as GSM phases out)


@zahidtg: Robert Joyce, Telefonica UK presentation is full of interesting stats, need to blog about it later :)


@KimKLarsen: TF UK experienced ca. 22% data offload to UMTS900 from 2100!


@KimKLarsen: U900 base 3G layer and U2100 the capacity layer ; Keeping traffic there and avoid HO to 900 challenging.


@KimKLarsen: TF UK calls their UMTS900 their Heineken network (they are also where nobody else are) ...;-)


@zahidtg: O2 UK is refarming the 900MHz spectrum and they will be using 5MHz for UMTS but there will still be some GSM channels


@patricksteemers = Great case on refarming GSM spectrum to cater for smartphone use. #TME


@zahidtg: In Panel discussion @patricksteemers reminds us of those good(?) old days when the main discussions were about Killer App(s) :-)


@zahidtg: TF: 75% of data traffic is from laptops but 75% of signalling traffic is from smartphones


@zahidtg: Robert Joyce, Telefonica UK says that on their Unlimited LTE trial n/w, biggest consumer consumes 55GB/week, Avg=2GB/month


@zahidtg: @KimKLarsen asks the audience if anyone will pay extra for facebook access, no one raises they hands :-))


@zahidtg: Martin Prosek, Telefonica, Czech Republic, points out that in the last year the amount of upload has doubled on average


@zahidtg: Thanks @patricksteemers for being an excellent chair. Enjoyed the panel discussion.


@KimKLarsen: @zahidtg amazing!!! I would pay extra! and I know my teenage daughter would! ... Hmmm ;-)


@KimKLarsen: the majority of mobile operators price plans are a sure road towards what they so much fear ... Becoming dumb bit pipes

DAY 2

@zahidtg: Ready for my breakfast briefing :-)

@KimKLarsen: had a really good breakfast discussion on #offload and on-load strategies and how to optimize between the two options.


@KimKLarsen: Yves Bellego, FT, France Telecom / Orange target to have 1/3 of all their cell sites shared within the next few years!


@zahidtg: Good to hear that Bell Canada has 6/7 carriers for UMTS in most places.


@zahidtg: Ljupco Jorguseski, TNO gave an interesting presentation on Optimisation but we will have to wait for technical details


@zahidtg: Its interesting that in #ITMoptimisation there are only 1 (today 2) women out of approx. 80 people. Should have a blog discussion on why...


@zahidtg: Capacity planning in mobile data networks experiencing exponential… http://goo.gl/fb/G03fh


@zahidtg: @patricksteemers just mentioned that open WiFI hotspots can spoof operator networks that can be a major security risk


@zahidtg: LTE Optimisation 

@zahidtg: Adnan Salkic is talking about evolution of WLAN interworking with 3GPP - Simplification of offloading


@zahidtg: Adnan Salkic has an interesting slide on different WLAN solutions being used by different operators. On blog once available


@zahidtg: Tiago Rodrigues of WBA speaking on Carrier WiFI


@zahidtg: 5C's of WiFi - Capacity Crunch. Coverage extension, Customer experience, Cost-effective access, Complimentary


@zahidtg: Need to check the WBA Next Generation Hotspot (NGH) program - Includes 802.1x, 802.11u, EAP-SIM & EAP-TLS/TTLS authentication


@zahidtg: WBA NGH trial phases

@zahidtg: Tiago Rodrigues, WBA says that for 10% of their members, 50% of traffic is being offloaded to WiFi


@disruptivedean: @zahidtg Thanks for your #ITMoptimisation tweets. Maybe ask a question about how other non-offload WiFi models (eg Onload) are supported?


@zahidtg: @disruptivedean Sure, in the Panel discussion in an hour.


@disruptivedean: @zahidtg My view is that offload is only use-case #3 or #4 for smartphone WiFi. Many uses are nothing to do with the MNO.


@zahidtg: David Antunes, Optimus, has loads and loads of useful results but too much info. Has to be analysed offline


@zahidtg: Good to hear CSFB working as it should be. 1.5-2.5s is setup time.


@zahidtg: Additional call setup time to CSFB expected to be shortened by DMCR (Missed what DMCR is...)


@zahidtg: Conclusion of LTE deployment challenges over legacy networks

@KimKLarsen: @zahidtg, good news! there is still a huge market potential for growing the mobile Internet access!


@KimKLarsen: @zahidtg bad news: users want true unlimited data access plans ... Kim> maybe this does not need to be bad!


@KimKLarsen: @zahidtg describes very well the Customer versus Operator dilemmas and competing interest!


@KimKLarsen: great illustration of the future smartphone traffic types: private wifi, 3G data, elastic wifi, on-load!...


@zahidtg: O2 puts wifi on the menu at McDonald’s - http://bit.ly/I5YymX - Mentioned in my talk


@zahidtg: Telstra closed their Carrier-Wifi network http://bit.ly/I5YQud when they launched 4G WiFi hotspot - http://bit.ly/I5YZh6 - Mentioned in my talk


@KimKLarsen: my take away from Randall Schwartz, Wireless 20/20, is wifi offload networks are not a magic bullet but depends on market!!


@zahidtg: ABI Research estimates that carrier Wi-Fi can deliver data at 5% the cost of adding cellular capacity http://bit.ly/HSQp6i - Mentioned in my talk


@KimKLarsen: really enjoying Randall Schwartz techno-economics insights into wifi offloading strategies...This stuff really excites me!


@zahidtg: Yes Randall Schwartz, Wireless 2020 presentation is good but loads of info and it would need offline analysis


@zahidtg: Only half an hour for panel discussion at #ITMoptimisation . @KimKLarsen on the panel so I am only one to tweet


@zahidtg: The Panel topic is: Optimisation strategies for the successful deployment of next generation LTE - #ITMoptimisation - other ques not allowed


@zahidtg: @KimKLarsen "Whatever you deploy should be agreeable with the next gen technology. Should support both HSPA+ and LTE"


@zahidtg: David Antunes, Optimus "Backhaul is probably the most important issue in Optimisation"


@zahidtg: Stephane Teral, "General advice is that nice to have such event to discuss various optimisation issues"


@zahidtg: Stephane Teral, Infonetics Research very impressed with the event, thinks this is the right size and balance


@zahidtg: @KimKLarsen "Wifi gives some benefits but not that extreme benefits, wifi is not small cell, lesson learned in data mining is that mobile users are not mobile at all"


@zahidtg: Optimus did interesting experiement; users were asked to do various tasks on phone and were happy with 800Kbps connection


@zahidtg: Question asked if there is a case for Carrier Wifi except for offload.


@zahidtg: Stephane Teral says that when the user offload, it should onload to carrier wifi so they can be tracked


@zahidtg: David Antunes, Optimus says that if they offer true unlimited data offloading is useful to improve network capacity


@zahidtg: @KimKLarsen says that Carrier Wifi can help solve international data roaming issues (before being forced by EU)


@disruptivedean: Only if the user chooses the carrier's WiFi, surely? MNO shouldn't be tracking use of private, open, ...


@zahidtg: @disruptivedean #ITMoptimisation Sorry time constraints, cant ask another question :-(


@zahidtg: Ok, last presentation at #ITMoptimisation by Mark Nash, iPass - Commercializing Wi-Fi


@zahidtg: what role will WiFi play in 4G

@zahidtg: Mark Nash, iPass: By 2015 there will be around 5.8million public Wifi hotspots available


@zahidtg: 3G data roaming prices - rip off?

@disruptivedean: @zahidtg Monthly subscription model inappropriate for roaming, except for handful of regular travellers


@zahidtg: @disruptivedean I wont mind paying that amount even for 2 days as I know, will not have to worry about bundle allowance


@zahidtg: Mark Nash, iPass points out that so many operators are losing business because of stupid roaming rates.


@zahidtg: iPass closing remarks

@zahidtg: Thanks to fellow tweeter @KimKLarsen. Looking forward already to #LTEWS next month. 


POST EVENT

@zahidtg: The concept of 'PDP Context Parking' #ITMoptimisation http://goo.gl/fb/HwVoY


@zahidtg: Operators strategy for supporting the ‘Mobile Data Explosion’ - http://goo.gl/fb/DyjZD


Note that tweets have been edited for clarity

PARTICIPANTS

@patricksteemers = Patrick Steemers
@zahidtg = Zahid Ghadialy
@KimKLarsen = Dr. Kim Larsen
@kitkilgour = Kit Kilgour
@disruptivedean = Dean Bubley

Capacity planning in mobile data networks experiencing exponential growth in demand

Key note presentation by Dr. Kim Larsen at Informa’s 3G, HSPA & LTE Optimization conference in Prague.

You can follow the complete discussion on Twitter using tag #ITMoptimisation

Capacity planning in mobile data networks experiencing exponential growth in demand from Dr. Kim (Kyllesbech Larsen)

Related Post:

• Summary of tweets from #ITMoptimisation

Mobile Energy Efficiency (MEE) Optimisation project

Recently read that Telefonica, Germany has identified that it can save €1.8 million per year with the help of GSMA's MEE Optimisation service. Here is a detailed case study from GSMA:

Case Study: Mobile Energy Efficiency Optimisation

View more documents from Zahid Ghadialy

Also, found a presentation that explains a bit more about what MEE (Mobile Energy Efficiency) is:

Mobile Energy Efficiency Explained

View more presentations from Zahid Ghadialy

Maybe a good idea for other operators to start looking into how they can be saving with this initiative as well.

More details on MEE here.

Testing and Optimising LTE

JDSU: Testing and Optimising LTE

View more presentations from Zahid Ghadialy

An interesting presentation from JDSU

Optimising Networks for Video Delivery

Interesting presentation from Alcatel-Lucent

Phillippe Keryer - Alcatel Lucent - BBWF 2011 Keynote Day 2

View more presentations from ceobroadband

Use the slideshare link to download the presentation.

Service Layer Optimization element to Improve Utilisation of Network Capacity

Picture Source: Simple LTE for 4G Mobile Broadband

The following is an extract from 4G Americas whitepaper, "Optimizing the Mobile Application Ecosystem":

Applications have diverse requirements on the mobile network in terms of throughput, relative use of uplink vs. downlink, latency and variability of usage over time. While the underlying IP based Layer 3 infrastructure attempts to meet the needs of all the applications, significant network capacity is lost to inefficient use of the available resources. This inefficiency stems primarily from the non-deterministic nature of the aggregate requirements on the network from the numerous applications and their traffic flows live at any time.

This reduction in network utilization can be mitigated by incorporating application awareness into network traffic management through use of Application or Service Layer optimization technologies. A Service Layer optimization solution would incorporate awareness of:

1) device capabilities such as screen size and resolution;

2) user characteristics such as billing rates and user location;

3) network capabilities such as historic and instantaneous performance and;

4) application characteristics such as the use of specific video codecs and protocols by an application such as Video on Demand (VOD) to ensure better management of network resources.

Examples of Service Layer optimization technologies include:

* Real-time transcoding of video traffic to avoid downlink network congestion and ensure better Quality of Experience (QoE) through avoidance of buffering

* Shaping of self-adapting traffic such as Adaptive Streaming traffic through packet delay to avoid downlink network congestion

* Shaping of error-compensating flows such as video conferencing through use of packet drops to avoid uplink network congestion

* Shaping of large flows such as file uploads on the uplink through packet delays to conserve responsiveness of interactive applications such as web browsing

* Explicit caching of frequently accessed content such as video files on in-network CDNs to minimize traffic to backbone

* Implicit caching of frequently accessed content such as images in web content on in-network caches to improve web page retrieval speeds

Service Layer optimization technologies may be incorporated in the data path in many locations:

1) the origin server;

2) the UE device;

3) as a cloud-hosted offering through which devices and/or applications and/or networks route traffic or;

4) as a network element embedded in a service provider’s network.

Further, in a service provider’s network the optimization function may be deployed in either the core network and/or edge aggregation locations. When Service Layer optimization entities in the network are deployed at both core and edge locations, they may operate in conjunction with each other to form a hierarchy with adequate level of processing to match the traffic volume and topology. Such a hierarchy of network entities is especially effective in the case of caching.

The 3GPP standard network architecture defines a number of elements such as QoS levels that are understood and implemented in the network infrastructure. However, much of this network capability is not known or packaged for use in the Service Layer by application developers. One approach to resolving this discrepancy may be to publish standard Service Layer APIs that enable application developers to request network resources with specific capabilities and also to get real-time feedback on the capabilities of network resources that are in use by the applications. Such APIs may be exposed by the network to the cloud or may be exposed to application clients resident on mobile devices through device application platforms and SDKs. The network APIs being defined by the Wholesale Application Community are an example of the recognition of the need for such Service Layer visibility into network capabilities. Future versions of the WAC standards will likely incorporate and expose network Quality of Service (QoS) capabilities.

Pic Source: Aria Networks

Why does Optimization matter? A good answer to this question is provided in Telecoms.com article as follows:

For many people, says Constantine Polychronopoulos, founder and chief technology officer of mobile internet infrastructure specialist Bytemobile, the definition of optimisation as it relates to mobile networks is too narrow; restricted to compressing data or to the tweaking of the radio access network in a bid to improve throughput. While these are key elements of optimisation, he says, the term ought to be interpreted far more broadly. “The best way for us to think of optimisation,” he says, “is as a set of synergistic technologies that come together to address everything that has to do with improving network and spectrum utilisation and user experience. If you stretch the argument, it includes pretty much every thing that matters. This holistic, end-to-end approach to optimisation is the hallmark of Bytemobile’s solutions. Point products tend to be costly and difficult or impossible to evolve and maintain.”

And optimisation matters, he says, because the boom in mobile data traffic experienced in some of the world’s most advanced mobile markets represents a serious threat to carrier performance and customer satisfaction. US operator and pioneer iPhone partner AT&T is a case in point, Polychronopoulos says.

“If you look at what’s been said by Ralph de la Vega (president and CEO of AT&T Mobility) and John Donovan (the firm’s CTO), they have seen a 5,000- per cent increase in data traffic over the past two years. The data points from other operators are similar,” he continues. “They see an exponential growth of data traffic with the introduction of smartphones, in particular the iPhone.”

Operators may have received what they’d been wishing for but the scale of the uptake has taken them by surprise, Polychronopoulos says. The type of usage consumers are exhibiting can be problematic as well. Bytemobile is seeing a great deal of video-based usage, which can often be a greater drain on network resource than web browsing. Given the increasing popularity of embedding video content within web pages, the problem is becoming exacerbated.

Dr. Polychronopoulos is keen to point out that there are optimisation opportunities across different layers of the OSI stack—Bytemobile offers solutions that will have an impact on layers three (the IP layer) through seven (the application layer). But he stresses that some of the most effective returns from optimisation technologies come from addressing the application layer, where the bulk of the data is to be found.

“An IP packet can be up to 1,500 bytes long,” he says. “So at layer three, while you can balance packet by packet, there is only so much you can do to optimise 1,500 bytes. At the top layer, the application can be multiple megabytes or gigabytes if you’re watching video. And when you’re dealing with those file sizes in the application layer, there is a whole lot more you can do to reduce the amount of data or apply innovative delivery algorithms to make the content more efficient,” he says.

By optimising content such as video, Polychronopoulos says, significant gains can be made in spectral and backhaul network utilisation. A range of options are open to operators, he says, with some techniques focused on optimising the transport protocol, and others designed to reduce the size of the content.

“With video, we can resize the frame, we can reduce the number of frames, we can reduce the resolution of the frame or apply a combination of the above in a way that does not affect the video quality but greatly improves network efficiencies,” he says. “So if you go to a site like YouTube and browse a video, you might download something like 100MB of data. But if you were to go through a platform like ours, you may download only 50MB when the network is congested and still experience not only the same video quality, but also fluid video playback without constant re-buffering stalls.”

It is possible, he explains, to run these solutions in a dynamic way such that data reduction engages only when the network is congested. If a user seeks to access high-volume data like video during the network’s quiet time, the reduction technologies are not applied. But when things are busier, they kick in automatically and gradually. This could have an application in tiered pricing strategies. Operators are looking at such options in a bid to better balance the cost of provisioning mobile data services with the limited revenue stream that they currently generate because of the flat rate tariffs that were used to stimulate the market in the first place. Being able to dynamically alter data reduction and therefore speed of delivery depending on network load could be a useful tool to operators looking to charge premium prices for higher quality of service, Polychronopoulos says.

If it is possible to reduce video traf- fic in such a way that data loads are halved but the end user experience does not suffer proportionally, the question arises as to why operators would not simply reduce everything, whether the network was busy or not. Polychronopoulos argues that in quiet times there are no savings to be made by reducing the size of content being transported.

“The operator has already provisioned the network one way or another,” he says, “so there is a certain amount of bandwidth and a certain amount of backhaul capacity. When the network is not congested, the transport cost is already sunk. When it becomes congested, though, you get dropped calls and buffering and stalled videos and the user experience suffers. That’s where optimisation shines. Alternatively, media optimisation can be factored in during toplevel network provisioning when the savings in CAPEX can be extremely compelling.”

While LTE is held up by some within the industry as the panacea to growing demand for more mobile broadband service, Polychronopoulos is unconvinced. If anything, he says, the arrival of the fourth generation will serve only to exacerbate the situation.

“LTE is going to make this problem far more pronounced, for a number of reasons,” he says. “As soon as you offer improved wireless broadband, you open the door to new applications and services. People are always able to come up with new ways of inundating any resource, including bandwidth. We’re going to see more data-driven applications on mobile than we see on the typical desktop, because the mobile device is always with you.” And while LTE promises greater spectral efficiency than its 3G forebears, Polychronopoulos says, the fact that spectrum remains a finite resource will prove ever more problematic as services evolve.

“We’re reaching the limits of spectral efficiency,” he says. “Shannon’s Law defines the limit as six bits per Hertz, and while we may be moving to higher-bandwidth wireless broadband, spectrum remains finite. To offer 160Mbps, you have to allocate twice the amount of spectrum than in 3G, and it’s a very scarce and very expensive resource.”

Operators have been wrong to focus exclusively on standards-based solutions to network optimisation issues, Polychronopoulos says. In restricting themselves to 3GPP-based solutions, he argues that they have missed what he describes as “the internet component of wireless data.” Internet powerhouses like Google, Yahoo and Microsoft (which he dubs ‘the GYM consortium’) have established a model that he says is a great threat to the mobile operator community in that it establishes a direct consumer relationship and disregards the “pipe” (wireless broadband connection) used to maintain that relationship.

“The operators have to accelerate the way they define their models around wireless data so that they’re not only faster than the GYM consortium in terms of enabling popular applications, but smarter and more efficient as well,” he says. Dr. Polychronopoulos then makes a popular case for the carriers’ success: “The operators have information about the subscriber that no other entity in the internet environment can have; for example, they know everything the subscriber has done over the lifetime of their subscription and the location of each event. They don’t have to let this data outside of their networks, so they are very well positioned to win the race for the mobile internet.”