Exploring how autonomous networks enable novel, disruptive services

The Catalyst project, Autonomous networks hyperloops for disruptive digital services Phase II, imagines new disruptive services that are completely autonomous, designed for zero-fault operations, with auto-scaling to meet customers’ high expectations. The project explores how autonomous networks (AN) can enable novel, value-generating services.

The project looks at how AN platforms and infrastructure at the mobile edge enable communications service providers (CSPs) to provide connectivity and deliver applications to vertical sectors. Leveraging AN, they can ensure business service continuity, provide on-demand services and create new service experiences for customers.

This builds on work from the Catalyst’sfirst phase in 2020, which was conceived as a multi-phase project from the start in recognition of AN’s critical importance. The goal was to show how autonomous networks can facilitate and accelerate time to market for next-generation services, and in particular, to develop and demonstrate quick wins.

Contiguous, closed loops

The first phase drew on TM Forum’sAutonomous Networks Project, which launched in 2019. The team amassed assets and coordinated its efforts withETSI for the technical architecture. The term ‘hyperloops’ describes the suite of contiguous, closed loops the team is developing, each with the right characteristics for its domain, from an IoT monitoring device, for example, through the network, to the edge and the cloud.

The champions of this second phase are NTT and Orange, supported by participants Beyond by BearingPoint, Cognizant, Futurewei and UBiqube. The team is usingTM Forum’s Open APIs and theOpen Digital Architecture (ODA) for its solution.

Andreas Polz, SVP Technology, Innovation, Standards, and Chief Evangelist at Beyond by BearingPoint, is also Co-chair of Open APIs at TM Forum and a Forum Distinguished Fellow. He notes, “[The Business Process Framework] eTOM is always in the background to orient ourselves and we are finding gaps and possible additions to it. We are discovering a lot in this AN project as we are moving from the theoretical discussion in the project to testing and validating, and now we need to add elements from a Self-X perspective, to the Business Process Framework to capture all the different domains and resources that are involved in AN.”

Takayuki Nakamura, Manager at NTT, adds, “To build a complete AN stack we need to define the way to incorporate across different domains of components specified by TM Forum’sOpen Digital Framework (ODF)”.

Cross-domain, cross-layer linkages

This is because, as Polz says, “After phase one we are discovering ever more that when Network, IoT and IT systems

operate in an autonomous way the historic distribution across the domains from the Business Process, andInformation

(SID), andApplication Frameworks (TAM) suddenly need links that were not previously obvious. AN requires many

more cross-domain and cross-layer linkages than was previously envisaged.

“A process that in the past was considered to live in the resource or service domain needs much more interaction and more data exchange with a customer domain and product domain than the previously envisioned in classic IT systems.”

The team believes that automation only carried out in separate domains could result in an operator having automation for its own sake, rather than gaining the full business benefits from constant data exchange across the domains as needed. This is automated by the hyperloops orchestrating activity within the various domains’ closed loops, automatically and constantly aligning the resource intent with the business intent that lives in the customer domain.

In addition, Polz says, “We are experimenting and innovating with the APIs, and think that we will probably need to develop one for capping budgets in one of our use cases”. For example, in the Catalyst’s scenario, the end customer is a farmer and a number of use cases are demonstrated via an ecosystem of specialist services providers in addition to the CSP, or other related IT solutions. This is very important to the operator champions which see their future in B2B2X

One of the use cases demonstrated involves normal operations run at the edge, such as monitoring temperatures in greenhouses. If an abnormal temperature is detected, additional monitoring is triggered, such as a video feed, so that people in the central monitoring hub can see the state of the plants. Bringing additional monitoring into play also triggers more network and IT resources, and maybe more human resources.

This increases the cost to the service provider as well as to the customer, so the AI in the system needs to calculate if there is enough left in the customer’s agreed budget to implement more resources by analysing the farmer’s business intent. It might be that the extra resources are activated or it could be that the system simply sends an alert. This is all achieved through the feedback loop.

This is key: AI can undertake analysis to take informed decisions but it is imperative that boundaries are set regarding those decisions, rather than handing too much control to the AI. The team is proposing a Limits and Caps Management API to add to the Forum’s Open API suite to manage the limits of what autonomous operations and AI based activity are allowed to do.

Watchthis video to see the team talking about the Catalyst.

Visit the Catalyst space for more information from the team.