It’s Time for a Reality Check on Network Automation

Why it’s so hard – and 3 steps that will help ensure success.

The telecom industry has been talking about network automation for the past decade, with promises of zero-touch automated networks with virtualized, cloud-based OSS systems managing everything seamlessly behind the scenes.  If you were to go on what the the industry pundits and analysts have to say- every operator’s network is nearly fully automated and OSS vendors have virtualized solutions that are already broadly deployed. This leaves most operators believing that their networks are woefully behind the times. But… is that the reality on the ground? Does ‘everyone else’ already have a fully automated network that hums along powered by rainbows and unicorns?

The answer is no – at least not yet.  According to TM Forum’s Digital Transformation Tracker, only 15% of communications service providers (CSPs) surveyed characterize their company as ‘innovative and risk-taking’.  This is a percentage that has remained constant for years. At Enghouse Networks, we engage with carriers all over the world, and we’ve found a common theme – automation isn’t easy, and it’s not just because most CSPs aren’t risk takers.

Four reasons why network automation is so challenging:

• New skillsets are needed for technologies like IoT, SD-WAN, and 5G – and CSPs are having trouble filling these roles. In addition, organizations are siloed, with knowledge unevenly distributed. Sharing information across organizations and keeping everyone and everything aligned has become a real challenge.
• SLAs are evolving and getting more stringent as businesses increasingly view connectivity as a mission-critical resource. In the past, best-effort was good enough. Now, tracking and delivering on these promises can be daunting.
• Customer expectations are changing. With hyperscalers like AWS and Microsoft provisioning services in real time, subscribers expect the same on-demand service from telcos – but they were never designed this way. Can CSPs compete?
• Network complexity is overwhelming. Amid system upgrades, network operations teams are tasked with getting newly virtualized systems and legacy tools and technologies to work together to keep critical services running. This is about as easy as building a plane while it’s flying.

Despite the difficulties, no one is saying automation projects aren’t worth the effort. Networks are getting larger and more complex. Subscriber demand for data is unstoppable. The manual tasks that were easily managed yesterday have grown to such a scale that there’s no other way around it – automation is the only answer. But before projects like these can be successful, it’s important to start with a strong foundation, and that includes understanding what the current network consists of.

CSPs must have an accurate inventory of both their physical and virtual assets before undertaking big transformation projects like network automation. But often, inventory systems are outdated and siloed. This can lead CSPs to make decisions based on wrong information – creating costly, business-impacting mistakes. The kind that keep CTOs and CFOs up at night.

Observe, Organize and Overcome

Here are three steps every CSP should take on their path towards automation:

Step 1: Observe. Recognize that real networks are messy – made up of multiple vendors, technologies, and ‘vintages’ – eras of software solutions that were designed to meet the challenges of their time, not today’s. Understanding this reality is the first step – and helps sets the stage for what comes next.

Step 2: Organize. It’s time to embody what Albert Einstein once said, “Out of complexity, find simplicity.” First, eliminate unnecessary complexity from your network, including superfluous systems and tools such as various ‘band-aid’ solutions that can accumulate over time. This can lead to replication of misinformation and silos of inaccurate data. Next – and this is so critical – clean up your inventory. This effort should include both network and services – so critical decisions can be based upon accurate information.

Step 3: Overcome. Automation can be scary. Prioritize efforts by focusing on simple, repeatable tasks first, and put processes in place that allow for manual checks if these are required by the organization. Next, leverage advanced technology, including ML/AI, so systems can continuously learn and improve over time.

Here’s an example of a phased approach for automating service assurance:

1. Automatically raise a trouble ticket. Systems can be trained to detect when trouble tickets need to be created and can then be designed to automatically raise them – without manual interventions. This saves time, money, and resources.
2. Automatically inform internal personnel and customers about service-impacting network faults. Don’t wait for customers to complain- be proactive. This reduces repair times and improves customer satisfaction.
3. Automatically take basic actions to fix faults. Is this a common, recurring problem with a common, recurring solution? Then automate it. Let your network engineering team focus on the harder problems.