I spent a day chairing Informa’s Self Organising Networks Conference, which is actually in its second day today and somehow proceeded successfully in my absence.
First off, the day was expertly chaired. No, only kidding. It was expertly organised, with Informa working hard to get its operator contacts to speak and share, and vendors (mostly) avoiding too much puff. Something like SON is quite a tight knit market, many of the vendors know each other well, and the operators know the vendors – so there is little room for too much guff in any case.
So, what did I glean from a day’s
expert chairing? Here are five SON trends I thought were emerging through the day.
1. SON is delivering real world benefits to operators already.
Something like ANR – carried out either in D-SON or with a C-SON control – can reduce dropped call rates to almost zero by mopping up any neighbour conflicts, and eliminating failed handovers etc.
Huawei described completing configuration of a six sectorisation project in Thailand across 2,000 sites in 4 weeks. In one week the Dropped Call Rate (DCR) was reduced to baseline thanks to ANR deployment.
Orange said its trials had seen typically 50% improvement in “missing neighbours” due to ANR, and a 5-10% improvement in DCR.
TIM said that it had achieved almost zero dropped calls – on an already well optimised network – through ANR activation in a C-SON deployment trial.
2. C-SON is happening. And now.
Jiang Wangcheng, President of Huawei’s Wireless Network SingleOSS product line, said that the vendor now has 20 C-SON commitments from operators. Earlier this year that number was… just one.
Orange has an RFP out right now, with selection in 2016, for C-SON – and is doing so as part of its joint procurement initiative with Deutsche Telekom. Fabrice Robin, Director of Radio Engineering and Performance, said that the company sees C-SON being useful for auto configuration, antenna remote mechincal tilting (ie tilting an antenna at a certain time to get different vertical coverage), load balancing and traffic steering, the handling of mass events, and for self-healing (ie managing a network alarm automatically).
P.I.Works said that a C-SON implementation with “an Indonesian operator” had achieved optimisation across a Huawei LTE-TDD macro and Airpsan microcell network and reduced congested cells 92% on SON activation, and achieved CAPEX deferral on those cells by 13 months. A P.I. Works US customer had reduced VoLTE DCR by 7.5% along a particular road, that tended to have plenty of VIP customers driving along it.
3. C-SON brings multi-vendor – even from the RAN vendors.
Huawei claimed that tests of its C-SON CCO (Coverage & Capacity Optimisation) had improved performance in an Ericsson network in France – reducing RRC Radio Resource Control) failures by 75%.
A presentation from Cellcom’s Radio Systems Engineer Orly Ehrlich, identified Nokia’s iSON as providing a multi-vendor SON with several key results. These included a 40% improvement in handover success, and a 50% improvement in some uplink throughputs by suppressing the number of Call Quality Indicators sent by handsets up into the network.
TIM also shared some positive results of a C-SON field trial, although it also identified some issues – for example integration to legacy OSS, and the ability to customise C-SON to enable SON to work in accordance with existing operator policies. For example, TIM was a policy not to allow the automation of certain network parameters.
4. You might think that Het Net management continues to be the go-to macro driver for SON – but it’s more complicated than that.
Does SON need Het Net? It’s more that Het Net needs SON – for mobility management, ANR, interference cancellation and so on, and although SON probably needs Het Nets to happen to really bust through – as we’ve seen, there are plenty of non-Het Net related use cases for SON.
Orange’s speaker, for instance, pointed out that Orange doesn’t really have very many small cells. And TIM’s presenter added that 5G presupposes the Het Net, and it may be here that we see a real need for managing extra capacities between network layers and RATs.
One aspect that was of note, relating to Het Nets, was the increasing mention of WiFi and how that might be managed within the carrier optimisation space – for instance using SON for some element of automated traffic steering – depending on the best available or most suitable (per subscriber) connection. (Also in this vein although more in the R&D phase was research from Dutch research house TNO, as part of the EU Semafour project, that was proposing being able to aggregate unused channels in GSM spectrum to LTE channels to increase bandwidths – using SON to define and manage the rules for doing so.)
5. There was some philosophical musing about what SON might turn into.
This took two forms, really. One was an outward expansion of SON beyond its classical 3GPP-defined use cases into seeing it as an enabling technology for increased automation of the network. This was probably most clearly outlined by Amdocs’ Head of RAN Marketing, Neil Coleman, who saw a host of other use cases enabled by SON functions – such as Customer Experience or more business-orientated process than network processes. Others also picked up on this aspect of SON – its potential to be used as part of a wider strategy to mash together customer quality and network quality KPIs to achieve network optimisation that is revenue focussed, and not just established to satisfy technical KPIs.
In his presentation Coleman had also said he could see SON splitting into multiple categories, and that “operators won’t call all of it SON” – suggesting it will become part of “automated and strategic optimisation” within the network.
And this takes us to the second form of this expansion of SON flowed from the first – and was expounded by Telecom Italia Mobile’s Andrea Buldorini (Senior System Engineer, Wireless Access). Buldorini, tasked to speak on new challenges for SON from LTE to 5G, spoke of the eventual need for Open SON, with NFV being the driver for the increasing availability of open APIs to other network management systems, “breaking down SON processes into functions that are part of a wider process of network and service creation”. Architectural changes would also ensue, for instance within the vRAN, SON and RRM cease being separate entities. If this sounds confusing, what Buldorini was driving at was the creation of “autonomic” functions, exposed via a SON API to advanced optimisation and network applications.
Also treading something of the same road was Huawei’s Mohammed Madkour, VP Wireless Network Marketing and Solution Sales, who described SON as the thread that can, via its inherent closed loop automation capabilities, tie together the 2020 vision of the software defined telco network – again via the exposing of a SON open API to CEM, CRM, OSS and other tools. “Automation is the main and key concept of 5G, and you can’t do that within the current operational model”, he said. Here he saw SON as providing the capability to optimise the network between and within network “slices”.