NFV Evolution through Machine Learning And Artificial Intelligence

 

NFV might be considered the evolution of virtualization, the introduction of highly advanced orchestration models is evolving the NFV to a whole new paradigm. Chaining VNF’s (Virtual Network Functions) is becoming a powerful base to insert artificial intelligence and Machine Learning into network operations. Service Providers can enable disable services and instances based on QOS, QOE or even from their BSS insights.

Service efficiency and profitability will sky rocket when you allow algorithms to decide the number of instances to be available, which voice services to deploy, how many hardware resources to use and when should they be enabled. In a practical example we can use one of the hottest topics today which is the Unified Communications.

Imagine a self-service portal where the enterprises sign up and download a provisioned client. On the backend an artificial intelligence algorithm reading through the provisioning data will decide which public cloud to use to better serve the customer and which API’s need to be configured to interconnect with the customers’ requirements, while at the same time generating customized documentation that teaches the customer on how to best use the service and how to efficiently troubleshoot it. All of this with zero intervention from the service provider.

Machine Learning can then analyze the customer’s usage and QOS data and take decisions to improve the service, like seamlessly migrate the VNF’s to a new cloud provider without any service disruption to improve the KPI’s. Same can be applicable in Wholesale carriers, in which they prefer to use their own private cloud. Once they register for the service, Artificial Intelligence will deploy the full service based on the initial data provided by the customer and decide the best Orchestration model. It will manage the voice services automatically by constantly studying several metrics such as Minutes of Usage, ASR, QOS or business KPI’s. Machine Learning can collect knowledge and spawn new services automatically by launching applications from a repository, these applications would be selected based on network trends and prediction models created based on past data.

If carefully analyzed several applications were mentioned previously such as voices switches, Billing platforms, Big Data, Business Intelligence, Machine Learning and Artificial Intelligence. All these applications can know seamlessly interact and provide actionable key points to the service providers. Entire ecosystems can be created, generating new opportunities for any new software vendor which translates into quicker innovation. Operation costs are heavily reduced improving the profit margins and reducing the friction to launch any new service. Subscribers have a new mindset they want the services Now, Everywhere and Always. The merging of NFV and Orchestration with Machine Learning and Artificial Intelligence will address those demands, by allowing the service providers to deploy new voice services instantly, in any part of the globe and able to scale them elastically based on quality, performance and availability requirements.

 

 


miguellopes_webinarphotoAuthor:

Miguel Lopes
Vice President, Product Line Management