Vital services require automation

Broadband is now considered a vital service because other lifeline services, like power and supply chains, rely on it for their own function. Similarly, people are more dependent today on broadband to participate in society, including work, education, healthcare and communication. “Internet is a basic human right, like you and I have the right to drink water,” said Tareq Amin, Group Chief Technology Officer, Rakuten during his closing keynote for FutureNet World on April 21. But as telecom services take on this new level of responsibility, they are also transforming to far more complex and dynamic technology platforms, like software-define networks and cloud-native IT. This sharp rise in complexity necessitates substantial network and operations automation.

Read the latest report from TM Forum Orchestrating broadband as a vital utility.

Automation needed to conquer complexity and support criticality

“The future is automation,” said Neil McRae, Managing Director and Chief Architect, BT to the FutureNet World audience. “If you run into advanced applications and life or death services, we can’t have the network react by human intervention – it’s got to be instantaneous,” McRae says. Automation is necessary to overcome the complexity inherent in orchestrating and assuring critical services over dynamically scalable, virtualized networks that provide as much raw capability at the edge as they do in the core. Network operations will need greater automation to fill a range of needs inherent to critical services:

• To operate the infrastructure remotely: Automation is key to operating infrastructure in remote areas without needing to source, train, and work hard to find staff. “If we can replace untrained manpower with trained machines, then we will,” said Randeep Singh Sekhon, Chief Technology Officer, Airtel. To achieve that level of automation, Singh added, a CSP needs visibility across all aspects of the network – core, RAN, backhaul, fronthaul – and to “correlate across them all.”
• To design for failure: Automation enables a CSP to “design for failure,” says Maria Eugenia Armijo Marchant, Integration Specialist Solution Architect, Telecom Argentina, so that “no matter how hard you try to harm an architecture or platform, it will still serve you at some level,” she explains. Armijo Marchant says this degree of resiliency goes beyond the network, especially in software-defined, cloud-native settings, to “architecture resiliency” where the entire, software-based architecture is designed for failure and to eliminate single points of failure.
• To provide critical resiliency: Armijo Marchant says being able to detach applications from a fixed runtime environment, in IT and software-defined networks, is crucial to achieving a high degree of automated resiliency as is ensuring that microservices pods will continue to run without “a person watching alarms on each virtual machine.” Similarly, she says, infrastructure can be detached from underlying hardware, which in turn allows a CSP to pre-design strategies for off-loading, traffic management, and crisis response.
• To achieve zero downtime: Automation is required for “reducing and diminishing network outages and incidents and downtimes to zero,” Michael Frankle, EVP & CTO, TDC told the FutureNet World audience. With critical services, minimizing outages may not be enough. CSPs want to gain predictive capabilities to focus on “managing congestion, correlation, and sniffing out troubles before they happen,” explains Frankle.
• To drive continuous improvement: Automation helps to drive continuous improvement, says BT’s McRae. A fast-emerging best practice is to use historical use cases to define and refine automated response plans. CSPs are training AIs and machine learning (ML) engines to make automated problem resolutions predictive and preventive by expanding the number of known use cases and automated actions, and improving them incrementally but continuously over time.

Network automation becoming predictive

While a culture that embraces failure may sound counterintuitive, proponents of AIOps, ML, and continuous improvement (CI) find failure informative in building predictive models. Service assurance engineers may use anything out of the ordinary that happens in the network as a new use case, monitor for it, and aim to respond preventively to it in a zero-touch fashion. Over time, use cases and fixes are refined to improve a CSPs ability to predict and prevent service degradations before they occur and provide critical services the 100% uptime and resiliency they require. For a closer look at how CSPs are automating network and service operations to support critical B2B services, read the latest report from TM Forum Orchestrating broadband as a vital utility.