Selfies go mainstream – why IoE networks need to self-organize

Once the realm of science fiction, the internet of everything (IoE) is fast becoming a reality. For communications service providers, this requires a new, automated way of organizing networks, and they’re using ideas first floated in the 3GPP’s specifications for self-organizing networks (SON) to do it. Dr. Phil Marshall takes a look at how SON concepts are being applied in the IoE.

With its promised billions of interconnected devices and people, the IoE demands service agility that cannot be supported by conventional appliance-based networks and siloed operational models. As AT&T’s Chief Strategy Officer, John Donovan, told attendees at the Open Networking Summit a year ago, “The traditional hardware approach is too slow. The only way to stay ahead of the demand curve is software, and we’re all in, in virtualizing our network.”

Web platform providers like Airbnb, Facebook and Google have raised the bar for customers’ expectations. They create new services in minutes, not months or years, with real-time and on-demand capabilities for service curation, management and orchestration. Network operators know they must respond by enabling complex digital ecosystems supported by a wide range of service level agreements (SLA), and this requires automation.

Reframing the experience

Vodafone Ireland Chief Marketing Officer, Brian Corish, described how digital natives are impacting his business, saying, “Our challenges are not set by our competition. The challenges are set by customer expectations. When looking at user experience, our competitors are Google and Airbnb. We have to re-frame our digital experience to match that.”

To do so, operators are transforming their business processes, networks and operations, adopting new technologies like network functions virtualization (NFV), software-defined networking (SDN), cloud-based applications and 5G.

The IoE needs automation because there simply isn’t time for manual processes with the number of changes that will need to happen every second in a network that has billions of nodes running millions of applications.

As the World Economic Forum pointed out recently in its Digital Transformation Initiative Telecommunications Industry white paper, “A network servicing 10 million endpoints and 10,000 nodes could see these numbers increase by up to five times by 2020. In terms of incidents per hour, this would lead to a 25 times increase from 400 incidents to as many as 10,000 per hour – an amount that is impossible to handle manually.”

Operators are looking to SON concepts to manage the IoE, but the scale really demands something akin to SON ‘on steroids’. When the 3GPP initially laid out the SON specs for radio access networks nearly a decade ago, it divided solutions into three categories: self-configuration, self-optimization and self-healing. The IoE needs all three.

“SON was first an idea for automation, but implementation was not done in a way that was very beneficial,” Klaus Martiny, Senior Program Manager, Deutsche Telekom told attendees of a 5G workshop at TM Forum Action Week in February. “Now is the time to talk about it more in detail, looking at what we learned from SON and how we can apply it.”

Intent-based policies

To achieve automation in networks made up of virtualized and physical components, network operators are using modular autonomic computing platforms that provide vertically partitioned and horizontal layered network services. These platforms are directed by intent-based policies that drive the internal network service implementations of closed control loop processes of fulfillment, assurance and billing.

Closing the loop means collecting and analyzing performance data of network resources to figure out how a network service can be optimized and then applying policy through orchestration to make the changes in an automated way. Intent-based policies contrast with traditional approaches by focusing on the policy goals as opposed to directing low-level configuration of network resources.

Operators, standards-development organizations and open source groups are embracing this approach because it enables standardized policies to cope with complex, multi-vendor ecosystem layers (for example, business, service and network). Closed loop automated functions are implemented in self-contained modular ecosystem platform domains. Ideally, standardized interfaces are implemented between the domains, so that they can be combined to create autonomic solutions.

With modular, Lego-like structures, service level orchestration and management capabilities can be rapidly developed and enable agile and versatile system-level automation. This approach builds on techniques developed in other areas, such as military operations.

Proceeding with caution

Service providers recognize the important role of automation in the IoE. However, many also view automation as risky and they stress the importance of rigorous trials and modeling to ensure that the automated functions perform as intended.

“For us, the network is a cost center – we need to manage networks with [fewer]humans and more automation,” says an executive at a Tier 1 operator. “But we need to be careful. Sometimes automation doesn’t work and can be more costly than doing things manually.”

Service providers have made some progress with operational automation in areas where new and complex technologies are deployed, or to eliminate mundane processes. For example, SON automation has been implemented by many mobile service providers for LTE and femto-cell radio network configuration management, and in some cases for optimization and self-healing.

“The network is a cost center – we need to manage networks with [fewer]humans and more automation…but we need to be careful.”

A TM Forum Catalyst project called Recover first, resolve next applied closed loop functionality, with the SON principles of self-configure, self-optimize and self-heal, to quickly recover from an SLA breach using a virtualized evolved packet core (vEPC) operating in a hybrid network environment made up of virtualized and physical components.

The breach was initially addressed by automatically recovering, or healing, from the fault or degraded performance scenario (recover first). Post-recovery, further self-healing, configuring and optimizing was used to resolve any issues in the service infrastructure (resolve next).

With this approach, the SLA breach can be efficiently resolved, with reduced impact on service performance. Orange was one of the sponsors of the Catalyst, and from the company’s perspective the primary benefits were identifying opportunities to decrease OpEx through automation and learning how to manage SLAs in hybrid networks, according to Michel Valette, Test/Diagnostics Domain Manager.

Using SON principles is not just confined to service performance. Another Catalyst called Maximizing profitability with NFV orchestration demonstrated how closed-loop self-configuration and self-optimization can be used to orchestrate network resources based on policies that simultaneously consider both technical and financial metrics. And one called Closing the loop – Customer experience, policy & virtualization used closed-loop self-healing and self-optimizing functionality in a converged and hybrid network, running both appliance-based and virtualized infrastructure.

Both network and service KPIs were used to detect inefficiencies, churn and revenue risk as a basis for core and RAN optimization.

TM Forum’s Zero-touch Orchestration, Operations and Management (ZOOM) project has created two information guides based on learnings from these Catalysts and others. Dynamic control architecture for managing a virtualized ecosystem (IG1128) provides an in-depth look at six closed loop Catalysts, while Business rationale and technical overview for orchestration and autonomic control loops (IG1139) provides a model for orchestrating a hierarchy of closed control loops.

Team members are working on a new Catalyst project called 5G service operations, which will be demonstrated at TM Forum Live! 2017 in May, that aims to show how the scale, volume and velocity of 5G workloads can only be achieved using these techniques.

This article appears in Perspectives 2017: Digital transformation opens new markets. Download the full report free here.


    About The Author

    Phil leads Tolaga's software architecture and development, and directs its thought leadership for the mobile industry. Before founding Tolaga, Dr Marshall was an Executive at Yankee Group for nine years technologies and telecommunication regulation.

    Leave A Reply

    Back to top