DTW (Digital Transformation World)
Context is everything for automated, Agile service operations
If you're attending Digital Transformation world, join José Domingos, BT's Chief OSS Assurance Architect, for a Campfire session on Thursday this week to discuss communications service providers’ visions for automated, Agile service operations – and most importantly, how to get there.
13 May 2019
Context is everything for automated, Agile service operations
Join José Domingos, BT's Chief OSS Assurance Architect, for a Campfire session on Thursday this week at Digital Transformation World to discuss communications service providers (CSPs') visions for automated, Agile service operations – and, most importantly, how to get there.
The advent of network functions virtualization (NFV), software-defined networking (SDN), 5G and cloud bring new opportunities and, let’s admit it, also great challenges to the telecommunications industry – CSPs and vendors, alike.
The advent of network functions virtualization (NFV), software-defined networking (SDN), 5G and cloud bring new opportunities and, let’s admit it, also great challenges to the telecommunications industry – CSPs and vendors, alike.
Business and customer drivers invariably point to a future where Agile service creation and operations are fully automated. Customers expect provisioning of services in minutes, bringing the need for a truly agile product and service creation. Rich infrastructure telemetry, artificial intelligence and machine learning provide data analysis techniques that promise truly predictive network and service assurance processes.
Service innovation can be quickly delivered through a rich software overlay layer, enabling faster deployment of new services. New resources are deployed flexibly, in software-only deployments over an increasingly intelligent, programmable and elastic infrastructure.
Finally, customers expect self-service portals and rich B2B application program interfaces (APIs) to seamlessly integrate their increasingly modernised IT infrastructure with provider network services, through digital services, APIs and data.
Today it is not uncommon to take months to provision legacy services. Service definition is largely a manual process. Customers order products using mostly legacy processes, using perhaps spreadsheets and some online forms. Services are primarily designed by one teams of people and provisioned by others. It is possible to achieve some automation, but primarily only for activation of resources. Networks expose command line interfaces (CLIs), and the dependency on inflexible physical resources coupled with traditional processes and systems makes the provisioning process slow and largely manual.
In some cases, recently, we have seen the advent of online catalogues, and automation of certain software-defined services, but this has been largely vendor specific and difficult to interoperate with legacy physical networks and other software-defined services.
So what does the future look like?
In a nutshell, customer services must take minutes to provision. Also, creation of new services must be greatly speeded up and have turnaround of days, maybe even hours.
But how can that be achieved?
Well, for a start, services and changes to those services need to be fully expressed in intent by the customer through a service catalogue and APIs. Services are assigned and designed automatically from the intent specifications in the catalogue, resulting in resources being dynamically created on demand using infrastructure that is fully intent-driven. Services need to be automatically remediated as part of their lifecycle. The network needs to be primarily composed of virtualized infrastructure that is provides the elasticity needed.
Today, it can take CSPs weeks to fix services and certainly hours to restore. Our IP networks are largely composed of distributed systems without centralized control. Some parts are slowly moving towards some centralised control, but most networks are manually repaired. Some repair and maintenance functions can be remotely performed although these are point cases. Our view of the network is imprecise and not-real-time.
With the arrival of real-time network telemetry and the vast technology leap permitted by AI systems providing the ability to perform offline and online analytics, automation and closed control loops become possible. DevOps capabilities allow automation of network repair, restore and maintain.
We can begin to envision a world where large parts of the autonomous network are fully automated with closed loop control happening based on real-time views of telemetry. Real-time insights on network performance are automatically created and provided to operations.
The end game is for services that never fail. The self-optimised, self-healing network that never fails and scales elastically. The network is fully zero-touch, managed permanently in closed loop mode. Information is analysed in real-time by AI expert systems and changes happen in split seconds automatically.
There are several TM Forum Catalyst proofs of concept that are being presented at Digital Transformation World this week which explore these concepts, such as the AI Leap and AIOps projects.
So often customers don’t have visibility of their service. They order services manually and receive paper reports. They see how their network performed yesterday, or several days or weeks ago. That may have been adequate in the past, but we have evolved. Customer portals now show some service reports and permit some changes. Largely these are still static – you can get what has been provided. Changes are difficult and APIs almost always absent.
Is this where we need to be? No!
Services need to provide real-time visibility of status and performance. Customer portals need to show 360-degree views of the customer service built on externalised APIs permitting B2B seamless integration. Customers must be able to see services and associated resources in real time but also integrate those transparently as part of their own infrastructure.
The goal of the Automated Agile Service Operations Campfire session at Digital Transformation World is to bring together CSPs and vendors to share their visions of the future. We will discuss how this vision should evolve and share some of the patterns we have developed to make it happen.
Come and join us in Nice – help shape the future and make it happen!
Service innovation can be quickly delivered through a rich software overlay layer, enabling faster deployment of new services. New resources are deployed flexibly, in software-only deployments over an increasingly intelligent, programmable and elastic infrastructure.
Finally, customers expect self-service portals and rich B2B application program interfaces (APIs) to seamlessly integrate their increasingly modernised IT infrastructure with provider network services, through digital services, APIs and data.
If cloud providers can do this (Azure, GCP and AWS to name a few), why can’t CSPs and their suppliers do it as well?
The journey towards intent-driven services
Today it is not uncommon to take months to provision legacy services. Service definition is largely a manual process. Customers order products using mostly legacy processes, using perhaps spreadsheets and some online forms. Services are primarily designed by one teams of people and provisioned by others. It is possible to achieve some automation, but primarily only for activation of resources. Networks expose command line interfaces (CLIs), and the dependency on inflexible physical resources coupled with traditional processes and systems makes the provisioning process slow and largely manual.
In some cases, recently, we have seen the advent of online catalogues, and automation of certain software-defined services, but this has been largely vendor specific and difficult to interoperate with legacy physical networks and other software-defined services.
So what does the future look like?
In a nutshell, customer services must take minutes to provision. Also, creation of new services must be greatly speeded up and have turnaround of days, maybe even hours.
But how can that be achieved?
Well, for a start, services and changes to those services need to be fully expressed in intent by the customer through a service catalogue and APIs. Services are assigned and designed automatically from the intent specifications in the catalogue, resulting in resources being dynamically created on demand using infrastructure that is fully intent-driven. Services need to be automatically remediated as part of their lifecycle. The network needs to be primarily composed of virtualized infrastructure that is provides the elasticity needed.
Is the journey towards Intent-driven services part of your vision?
The journey towards telemetry and closed-loop control
Today, it can take CSPs weeks to fix services and certainly hours to restore. Our IP networks are largely composed of distributed systems without centralized control. Some parts are slowly moving towards some centralised control, but most networks are manually repaired. Some repair and maintenance functions can be remotely performed although these are point cases. Our view of the network is imprecise and not-real-time.
With the arrival of real-time network telemetry and the vast technology leap permitted by AI systems providing the ability to perform offline and online analytics, automation and closed control loops become possible. DevOps capabilities allow automation of network repair, restore and maintain.
We can begin to envision a world where large parts of the autonomous network are fully automated with closed loop control happening based on real-time views of telemetry. Real-time insights on network performance are automatically created and provided to operations.
The end game is for services that never fail. The self-optimised, self-healing network that never fails and scales elastically. The network is fully zero-touch, managed permanently in closed loop mode. Information is analysed in real-time by AI expert systems and changes happen in split seconds automatically.
There are several TM Forum Catalyst proofs of concept that are being presented at Digital Transformation World this week which explore these concepts, such as the AI Leap and AIOps projects.
Can we deliver this vision alongside 5G? How do we make it happen as a combined industry?
The journey towards cloud-like service exposure
So often customers don’t have visibility of their service. They order services manually and receive paper reports. They see how their network performed yesterday, or several days or weeks ago. That may have been adequate in the past, but we have evolved. Customer portals now show some service reports and permit some changes. Largely these are still static – you can get what has been provided. Changes are difficult and APIs almost always absent.
Is this where we need to be? No!
Services need to provide real-time visibility of status and performance. Customer portals need to show 360-degree views of the customer service built on externalised APIs permitting B2B seamless integration. Customers must be able to see services and associated resources in real time but also integrate those transparently as part of their own infrastructure.
For this we need full API exposure and cloud-like consumption of services real-time reporting data and event data available via APIs, enabling customers to consume services as part of their automated playbooks. The journey to a cloud-like service exposure is also a journey to intent-driven services and truly automated closed-loop assurance.
From words to action
The goal of the Automated Agile Service Operations Campfire session at Digital Transformation World is to bring together CSPs and vendors to share their visions of the future. We will discuss how this vision should evolve and share some of the patterns we have developed to make it happen.
Perhaps some organisations are a little further ahead, perhaps some are behind, but without embarking together on this journey it will take a lot longer for everyone. Collaboration in targeted TM Forum standards and Catalysts will make this journey a lot shorter. What areas should we collaborate on and how do we push those standards forward? What vendors and CSPs can contribute?
Come and join us in Nice – help shape the future and make it happen!
