Opinion: Inventory is Dead, Long Live Topology

Slick inventory management is key to the cost effective deployment of new services, while rapid diagnostics and repair have underpinned improved customer service in recent years. But as the shift away from physical equipment to virtual services begins to pick up pace, the concept of inventory is becoming obsolete, creating a raft of very significant new challenges.

When every operational process, from enabling new services to fault management and network optimisation today relies on physical inventory management, the successful transition to virtualisation demands a new approach that supports more intelligent automation and a way of automatically discovering what is going on in the virtual world. Brian Naughton, CTO, Accanto Systems outlines the essential role of the OSS in delivering trust and control in an automated, virtual environment.

Obsolete Model

The ability to identify the location of physical assets in the network and monitor their status has been an essential component of every telco’s business model to date. But times are changing as organisations embrace Network Function Virtualisation (NFV). From quarterly capacity planning exercises to processes for deploying new appliances and upgrading equipment every three to five years, the essential manual processes that underpin every aspect of telecoms business will become redundant in a virtual world. In an NFV environment, where network functions are decoupled from dedicated hardware devices, there are no physical resources and no physical network to monitor.

The need for network visibility is, in fact, enhanced. In this new world, telcos plan to exploit real time optimisation to automatically deploy new services, especially innovative IoT service, and scale up and scale down as required in response to demands for better customer service or power utilisation. As orchestration tools (orchestrators) take control to continually optimise and adapt network functions in line with customer demand, power utilisation and resource availability, there is a clear need for both improved network visibility and a way to attain trust in those automated processes.

Critically, organisations need a way to map, in real time, a network that is now automatically and continuously changing as the system responds to customer demand.  A virtual network should be more efficient, cost effective and service driven but where is the control in this automated and dynamic environment? How, in effect, will the telco track these continual changes and have confidence that the optimisation is truly delivering to key business targets?

Automated Topology

Fault management is a prime example of the new thinking required in a virtual world. How can a telco identify the cause of a fault, for example, when the network state is continually changing?  When a fault occurs within a traditional physical network, an alarm is triggered, and while it may not be possible to immediately identify the cause of the fault, the telco’s OSS provides a picture of the exact state of the network, enabling the required diagnostic processes.

In a virtual world, where the network state is changing from second to second, a new device may have been created and destroyed before the fault is even raised. This constant state of change makes it far harder to identify the root cause of a fault – and, critically, avoid such a fault occurring again.

When every operational system from inventory to assurance, repair to order management is designed around the concept of a static physical asset or appliance, the shift towards NFV demands a very different approach. Clearly the OSS needs to evolve fast to enable the virtual mode and ensure the telco gains the benefits of faster service deployment and continual optimisation without losing control.

Trust and Control

Telcos need new systems that can capture and analyse network state on a second by second basis and use this insight to drive better optimisation, automation and control. This real time and continuous topological analysis is very different to inventory management and demands not only different toolsets but a way of managing those toolsets to reinforce trust in this new model.

The use of orchestration tools to automatically optimise network resources and self-heal to recover from problems is clearly compelling – but the challenge for businesses is to feel 100% confident in the tools undertaking these critical actions. Essentially, can telcos trust orchestration tools to make these essential network provisioning decisions in a completely automated fashion?

The answer is yes, but with provisos. No telco could risk switching on full automation from day one – the potential fallout should an orchestration tool decide to reallocate bandwidth to the high value video watching customer from a lower value hospital running a critical patient drug administration application would be unthinkable. The key is to use human intelligence in parallel with automation to build up confidence in this model.

An OSS that is continually monitoring and analysing the entire network state enables a telco to quickly build up a picture of network activity, what works and what doesn’t work.  The analytics provides network operators with the insight required to support decision making and hence build up confidence in the process. For example, by identifying the last time this specific set of virtual network appliances was created it resulted in a fault, the self-learning process ensures this configuration is avoided in the future. Essentially this evolutionary model enables telcos to ease into the full automation enabled by NFV with full confidence in the quality of the optimisation tools.

A Dynamic World

In the new, virtual world everything is dynamic. Dynamic topology and intelligent automation are compelling, offering huge potential wins but, on the flip side, massive potential losses if things go wrong.  Automation is clearly the goal – and it enables a new business model that is key to realising the huge potential market offered by IoT. But dynamic automation must also be controlled automation. If organisations are to have full confidence in network self-healing and optimisation tools, they need trust in this model – and that means achieving an overarching view of the entire network, including every orchestration tool, and a way of managing end to end orchestration policy. Without this, can organisations confidently embrace the power of NFV?