Automation is often hailed as potentially transformational for the telecoms industry. This is especially true when it comes to new technologies like 5G and OpenRan, which will impose a new level of complexity upon the industry and one which cannot be handled manually. Herein lies the problem.
Many operators have been running for decades and grown into large, complex multinationals. As a direct result of that success, they have an enormous backload of embedded technologies and processes which cannot be easily automated. This is mostly because these technologies are often siloed between disciplines and departments which artificially balkanise the data collection necessary for wholesale automation.
How siloing happens
Telecoms organisations are highly siloed across a whole number of areas. If they’re multinational – each separate country will likely have their own processes, practices and policies. If they’re big enough, each department will have their own budgets and metrics. Each part of these organisations will use their own technology stacks, databases and collect data in their own ways to serve the specific objectives of that particular department. These technologies and practices are often highly complex and they can’t share or integrate the various data streams they’re working with.
It’s this complex system of silos that prevents automation pipelines from getting off the ground. Automating processes requires the aggregate analysis and correlation of data points across the organisation – in order to train automation models. In turn, that requires automation at the most basic levels of data collection across the organisation and yet here we get unintegratable metrics, different technologies, conflicting practices and altogether separate processes. It’s technically possible to wrangle all that data together for the purposes of training an automation model, but doing so would be prohibitively time consuming, labour intensive and expensive.
This is as much of a technical problem as it is an administrative one. Operator silos create low visibility across the entire organisation and because each department is dealing with different streams of data and different metrics, they have a hard time translating that to streamlined processes and necessary organisational changes. When, for example, testing labs need a budget increase to increase test capacity, they need to justify it to management. They’ll need proof points to establish that need so they’ll need to dig through their specific tooling, manually extract the data and use those as proof points.
Testing automation
Automation is a particularly necessary development because of the rapid rise of transformational technologies like OpenRAN and 5G. These are unprecedentedly complex systems and will be deployed in such a wide array of use cases that they present a series of issues and considerations that telecommunications has never had to deal with beforehand. As a result, testing those networks is both a crucial and correspondingly complex process.
One of the places in urgent need of automation is the testing lab. It’s here where operators will test their products, services, components and technologies to see if they can serve the use cases and high standards for which they’re intended. Automation could help those labs execute hundreds or even thousands of test scenarios a day – but while testing remains manual, they are restricted to only a few test scenarios a day which dramatically slows down the process.
It’s worth noting that these problems aren’t shared amongst all 5G operators. In fact, smaller operators and start-ups are able to deploy virtualised 5G networks and OpenRAN much faster than their larger counterparts. This is largely because they don’t have the legacy backload of technology nor the silos that larger organisations are saddled with. In turn, this has allowed them to automate testing so they can test hundreds and thousands of test scenarios every day, compared with their larger, older counterparts who are stuck testing a mere handful of their potential. This faster pace in testing enables these start-up operators to launch new services faster, often leading to better user experience.
Automation is a matter of long-term survival for many operators who are now moving from the hardware based generations of mobile technology represented by 3G and 4G to the software primacy of 5G. Their ability to roll out new services and technologies effectively will have huge implications for operator’s long term profitability and market share. However, in order to go forward, they have to go back and start their automation efforts with some of their most foundational data collection processes.
By Aleksi Helakari, Business Development Manager, EMEA – Spirent
