Network Virtualisation? It’s all about Automation
Published on : Monday 30-11--0001
Communications service providers (CSPs) are facing highly disruptive challenges. Expanding volumes of data and video, mobile workload volatility, a greater number of connections and demand for low latency are driving CSPs to develop transformative strategies. And it is not only consumers who are forcing CSPs to reinvent their networks. As enterprises continue to digitise operations, their demand for bandwidth, more connections and lower latency substantially increases as well. This growth is due to innovations in areas like cloud computing and Internet of Things (IoT), as well as video-centric applications, such as those for training and video surveillance.
Expanding bandwidth in the traditional way – by adding new hardware appliances – is simply unrealistic. It requires significant funding and cannot keep pace with exponentially growing demands. The imperative for action has never been stronger. As a key enabler of the coming 5G infrastructure, network function virtualization (NFV) is the next logical step in the network evolution.
With 5G, network virtualisation and cloudification are fundamental to realising the required network services delivery. In particular, network virtualisation will enable 5G networks in which various virtual networks run on top of a single, physical infrastructure using technology that allows the network to be shared dynamically, or “sliced.” Virtual networks then can be rapidly customised to meet the needs of operators, consumers, and enterprise applications and services, such as those for remote healthcare and connected cars.
CSP executives’ view on network cloudification
In an effort to understand the industry’s progress in network virtualisation, identify leaders, comprehend their vision and derive learnings to share with their peers, we conducted extensive research, including interviews with 200 CSP executives – of which 71 CxOs – across the globe (see Re-envisioning the CSP network).
Our research revealed that network virtualisation is already helping a number of CSPs drive efficiency and agility in their networks to create new value. We applied cluster analysis to identify segments among the interviewees based on their approach to network virtualisation and how they execute on that vision. Three archetypes emerged, which we have named CSP Innovators, Evaluators and Laggards (see Figure 1).
Figure 1
Most CSP Innovators are already implementing virtualisation and cloudification technologies to support current and/or new services. They understand the importance of automation and say that artificial intelligence (AI) is a key underlying technology for effectively automating network operations. Innovators (25 per cent of the group we interviewed) are the standouts. They report they have outperformed their peers in both revenue growth and profitability in the past three years and also lead in innovation.
Evaluators are either conducting operational trials or testing the technology in a lab environment. They intend to integrate AI into their automation plans when starting their network virtualisation journey. Laggards are falling behind the Innovators and Evaluators. They are still in the consideration/evaluation phase, and automation for network operations are not yet on their radar.
Adaptable and automated networks to pave the way for 5G
Virtualisation at scale requires a network cloud infrastructure that allows optimisation through rapid scaling and descaling, resource sharing, agility and availability through lifecycle management of network services and applications. This provides scalability, business agility, fast- service innovation and delivery, identified by the Innovators as the top 3 drivers for network virtualisation.
Network virtualisation and cloudification go hand-in-hand with automatic operations. For the Innovators, the ability to more easily scale networks and meet service expectations are key drivers for automation. The vast majority of Innovators have already automated in some form more than 30 per cent of network functions, far more than their peers. Approximately four-out-of-five expect to have automated 50 to 70 per cent of their network functions within two-to-three years, with an ultimate goal of 70 to 90 per cent of network functions.
Key performance indicators (KPIs) are essential for assessing the success of network automation initiatives (see Figure 2). For Innovators, time-to-market (TTM) of new services is by far the most important KPI; for the others, it is OpEx. Innovators also rank customer experience measures as a key KPI. Indeed, automation can significantly reduce unpleasant experiences customers or enterprises regularly have with the network.
Figure 2
A “thinking” network for change
CSPs increasingly recognise AI’s central role in automating the network. Twenty-six per cent of our interviewees – and 96 per cent of Innovators – are using, or plan to use, some form of AI for automation. This enables faster decision making by capturing and processing network data and performance of key services in real time and by automating network functions. Organisations can train AI systems to look for patterns – detect, predict and localise irregularities in the network – and to take proactive steps to fix them before they impact customers or enterprises. AI can be combined with automation to solve problems and then apply the right resolution to reduce the complexity of operations and the number of operators needed.
CSPs can also use real-time analytics and AI to calculate future states based on various conditions and business policies. This enables zero-touch automatic provisioning of network resources in an optimal way to help improve service. Intent-based systems can combine this capability with orchestration automation to provide closed-loop control of the service lifecycle. AI-enabled operations can then drive machine-enabled service automation to move from a current state of service to the future desired state – enabling automation in real time without having to program every option operationally.
Innovators understand that proactively performing predictive maintenance on the network infrastructure and assets can yield a significant return in OpEx and customer satisfaction (see Figure 3). Other key areas they mentioned for using AI applications include self-diagnosing (such as troubleshooting and automatic detection of network problems), improving traffic management and self-healing of networks.
Figure 3
You’ve reached the end of this network!
The need for radical change is immediate and the price of staying on a rapidly disintegrating foundation – or burning platform – too high. No matter how much bandwidth is created from any amount of financial investment in networks, demand for bandwidth will always be greater. CSPs can choose to constantly fall behind and fail – or they can breathe new life into a network built on predictive, automated operations using cognitive capabilities that outthink demand and outperform expectations.
Captions
Figure 1: Cluster analysis of survey data.
Figure 2: Most important KPIs for success of automation initiatives.
Figure 3 Areas where Innovators are using or plan to use AI to support processes and personnel.