Putting experience first – how should the SOC evolve as we head towards new services and full 5G SA?
The transition from NOC to SOC has been in progress for several years. At the same time, innovations in Artificial Intelligence (AI) and Machine Learning (ML) have enabled the automation of many assurance processes. With an explosion of 5G services imminent, and the emergence of a host of new KPIs for new services with differentiated QoS, the SOC must also evolve to ensure the delivery of consistent experiences for all users and devices. What could the evolved SOC look like?
The SOC is fundamental to connecting experience to network performance
We’ve already witnessed the start of the shift from the NOC – network operations centre – towards a view that embraces services – the service-oriented operations centre, or SOC. The essential difference can be summed up by:
“The NOC tells me what things are working; the SOC tells me how this affects my services.”
While we won’t go into technical details here, the SOC achieves this by correlating and aggregating data from multiple systems into a complete, end-to-end view, so that all data that relates to the systems involved in, say service A, can be brought together.
This essentially eliminates the legacy, siloed view in which systems were managed separately – without the consolidated view necessary to support services. We’ve written about this on numerous occasions and it’s clear that the service-oriented view is now starting to take precedence, delivering insights above and beyond the end-to-end view.
The gap between problem detection and resolution needs to close
The SOC is a great step forward but there remain gaps to close. To explore these, it’s helpful to review an existing definition of SOC capabilities that can be found in work completed by the Telecom Infra Project:
“Service monitoring is the most vital function that is supported by the SOC. SOC staff monitors service trends, dashboards or reports made available to them…Once a service degradation is detected, SOC staff conducts preliminary filtering…and handles the case escalation according to defined procedures.”
And therein lies the first problem that needs to be confronted and which means that the SOC is set for another wave of evolution. Put simply, the gap between problem detection and problem resolution needs to close.
Second, the impact of this lag will soon be even more evident, when 5G ushers in a much broader range of services, covering both consumers and new B2B opportunities. This will generate a huge amount of new data and there will also be greater diversity in the type of data. That’s because new B2B services, enabled by URLLC and mIoT may have very different QoS performance criteria and offers a range of user experiences for people and, increasingly devices.
As such, new KPIs will be created for services that are delivered to customers from industry verticals (this is an exciting area – it’s well-worth checking out a new paper from 5GPP on this emerging topic), all of which will require management in order to ensure that SLAs are met.
Finally – and adding to the puzzle – is recognition that experience is a huge factor in determining customer loyalty. It’s well-known that individual customer experiences can differ greatly – and these differences are not necessarily clear from the network data available. Put simply, people expect more from new 5G services and speeds, so delivering the best experience will become ever more critical.
As such, service-oriented solutions that can understand and manage the way in which services are consumed are now of great interest – attention has shifted from simply maintaining the network to ensuring that it delivers what customers expect.
To summarise, we have three key problems to consider for the evolution of the SOC:
- Closing the gap between detection and resolution
- Supporting entirely new classes of services
- Maintaining customer experience, not just network performance
Automation is essential as we move to new service models and partnerships
Automation is key to solving these problems, because the unprecedented volume and variety of data goes far beyond the manual capabilities to which the above definition refers.
Manual methods cannot help us to deliver the services and experiences demanded.
Instead, service monitoring systems must be connected to solutions that process network data automatically and, in turn, to the APIs that drive network and service operations and delivery. So, when an alarm is triggered (e.g. when a service deviates from the required performance targets and KPIs), the severity, cause and appropriate resolution steps can be identified without manual oversight, and the necessary action taken in a fully automated closed loop.
But, while that sounds great, how does it help deal with customer experience, which, as we’ve already stated, may be negatively impacted without necessarily triggering a network alarm?
Well, the trick is to ensure that machine learning and artificial intelligence can be applied so that SOC systems can learn to extrapolate from the processed data and to predict when even a minor change could lead to something that impairs performance – and to learn with machine learning (ML) how to correlate KPIs with customer experience in a new way. It will be impossible to achieve this for a network that supports millions of users and, potentially, hundreds of millions of connected devices without using such ML and AI-based processing.
The evolved SOC – blending automated action with new experience measures
As we noted, here, new techniques will enable correlation of real customer experience reports with the network conditions responsible, going beyond far the traditional focus on alarms and triggered reports. If we can realise this vision – automating problem detection and resolution and integrating real reports of customer experience – then the SOC really will prove to be the key to future service delivery.
Of course, we are some way from realising this vision. But, it’s important to be clear about what we think the SOC should deliver in the future. If we really want to capitalise on the unparalleled opportunities for service innovation that 5G brings, we need to work to make this happen. We’ve successfully automated the NOC. We have brought new insights that have driven early SOC deployments. It’s only a matter of time before the SOC reaches this level – and we’re providing innovations to make this happen.
 “Service performance measurement methods over 5G experimental networks” 5GPP, White Paper – ICT-19 performance KPIs