The Road Ahead for Network Operations: Facing the Reality of the Journey to Automation

Automation has long been the goal of every network operator and seen as a required capability to support the network evolution and technology advancement. Many of the service innovations that are being promoted (white-box, Network Function Virtualization (NFV), cloud, edge-computing, etc.) all require automation as a standard part of their offering to realize the justification for investment.

However, the industry has struggled to remove the operational costs from network operations due to the limited capabilities of SNMP (Simple Network Management Protocol) & CLI (Command Line Interfaces) which, in turn, have inhibited innovation and automation. Today’s reality is that we still maintain largely physical networks with legacy interfaces.

There needs to be a revolution in the way operators work, with the focus on flexible, scalable interfaces supported by a ‘models-based approach’ to make automation a reality.

Physical Networks

Many network operations teams have a list of automation needs, yet struggle with prioritizing where and how to automate. Given the limited benefits of CLI and scripting, new technologies such as YANG and TOSCA provide modeling capabilities to capture this complexity, allowing for advanced logic and simplified programmability.

Operators should prioritize automation problem domains that can be expressed as models. Models can replace the exponential complexity that exists in templates and scripts today that are stalling network automation.

The benefits from such use-cases will provide not only support to the current physical network but will provide a platform from which to integrate with virtual network capabilities.

Virtual Networks

Most network operators are deploying initial virtualization use-cases as well as evaluating more extensive virtual capabilities. Virtualization is a concept and capability that will place the burden of managing even more complex networks and services on network operations teams.

But the initial journey into virtualization has been a bumpy one. Since most virtualization technologies are fairly new, most provide advanced API and YANG capabilities but also add complexity that traditional physical networks don’t. Given automation is a necessity for virtualization, many of the same network management processes can be extended across each technology domain.

A Fork in the Road – Converged Operations

The diverging management practices between physical and virtual networking domains are creating an ‘Operations Gap’. The benefits touted by adopting new virtualization and cloud concepts are many, but this tooling does not extend to existing networks.

This operations gap will grow as operations team are re-trained for new services while continuing to support existing infrastructure, placing a heavier burden on the ops staff. It is critical that similar automation strategies are applied to the physical network to drive towards converged operations.

Although similar concerns have existed with previous network technologies, the complexity of virtualization will be exacerbated by the lack of network management and interface innovation. New services will consist of a component architecture made up of containers, VMs and commoditized infrastructure.

The complexity that existed within a single network element will now be distributed among a variety of platforms. MANO (Management and Orchestration) promises to capture much of this complexity but is both in its infancy as well as significantly changing the way networks are built and managed.

MANO will drive new automation requirements that must be extended to support all varieties of virtual technologies being deployed across Enterprise and Service Provider networks.

The Destination – The Modern Network

We are converging toward the modern network and operations teams must consider a unified automation platform, built on a model-based approach, will be key to the adoption of new networking technologies. Existing business practices will not allow operations teams to treat these technologies in a silo given their potential impact on customers and services.

As networks become more complex, the need to simplify and automate is the only path operators can take to scale their business, increase service velocity and drive cost from operations.