Amid a growing shortage of workers, the wireless industry is expected to deploy hundreds of thousands of new wireless towers, small cells and distributed antenna systems over the next several years to meet the demands of 5G deployment and exploding data usage. How will the industry achieve more with less? Automation may be one answer.
Emerging technologies such as cloud software, mobility, drones, artificial Intelligence and in field-test automation offer significant opportunities to streamline and automate indoor and outdoor deployment programs. Panelists at a Connect (X): All Access virtual event discussed the current state of automation in the tower industry, why it is important and how the ecosystem can move forward. The session, “Automation Tools for Faster, More Profitable Infrastructure Deployment at Scale” was hosted by Tim Downs, Executive Producer at the Wireless Infrastructure Association, and included insights from Subbu Meiyappan, CEO, and Co-Founder of automation software company EC Site, Piyush Raj, Managing Director of SBA Communications, and Kishore Raja, VP of Engineering at Boingo.
To view this panel in its entirety as well as other content on demand, visit https://connectxallaccess.vfairs.com/.
Panelists agreed that the wireless industry could benefit from deploying automation tools in the site setup and network optimization processes but that the industry generally is behind other industries in deploying such tools. The telecom ecosystem is large and includes a wide variety of stakeholders, making adoption of end-to-end automation tools challenging, and until now, there has not been a return on investment to justify deploying automation tools. But that is changing with 5G.
Automation can mean many things, from streamlining input of site data into software that can then be shared across systems, to using computer vision to automatically identify components at a site. Ten years ago, employing electronic downtilt tools for antenna deployments and network optimization saved the industry untold time and money by allowing antenna repositioning without a truck roll. With virtualization, artificial intelligence and machine learning, the potential benefits of automation are even greater today.
“There are certain things you have to do manually – physical installation of servers, physical cabling, deploying small cell eNodeBs, towers and antennas,” said Raja. “But once you have those absolutely necessary items so many other things that go in the system can be virtualized. With the advent of autonomy and virtualization, the networks will adapt to changing conditions, and adding AI and machine learning will further prepare the networks to be proactive rather than be reactive so you will have a very consistent customer experience, create new opportunities and new revenue models, and allow different types of vendors to participate in the ecosystem. It’s basically a boon for the American economy.”
While automation sounds like a great idea, widespread deployment of tools won’t happen until players in the telecom ecosystem adopt them.
“Who wants to pay for it? That’s the million–dollar questions,” said Meiyappan. “Whoever feels the most pain will pay for it.”
Pain points that can be solved by automation include deployment and maintenance costs, time to market and a consistent high-quality customer experience, the panelists said. Raj said he sees adoption starting from the top of the food chain – at the wireless carrier level – because that is where the capex spend comes from. Wireless carriers are also feeling pressure on their margins as the revenue per gigabyte is going down, which could drive adoption of automation, he said.
“Obviously when you do things manually, you need more people and you need more time,” said Raj. “Every carrier has a site database, and in the past that database needs to be entered manually by each employee or whoever the network engineer is once the site is built. Now everything has an API, and that means you can have one software talk to another software in the same language and now you can take the software and update the whole database within a very short period of time. You save on the cost because you need less people. You save on the time. And you save on operating expenses because now you have more accurate information that came from the tool rather than somebody making an error and somebody correcting it later.”
Eliminating the human element can also free people up to perform more important tasks.
“Humans tend to get fatigued. A machine on the other hand … can repeat the same set of instructions thousands to millions of times once programmed by a human,” said Raja. “It can repeat that instruction any number of times exactly the same way, point to point, without having the human element of fatigue, which makes automation so much more efficient. And why do you want to have a human bogged down with manual tasks and mundane tasks which otherwise can be done by machines? … Once the system is set up and installed you can do things like benchmarks, performance metrics, KPIs, and SLAs. These can be repeated, automated any number of times so you have a consistent system which has constant checks on it. You don’t need to have humans going to check on it all the time.”
Meiyappan pointed to a use case at a small–cell vendor where technicians would collect data at a site, bring it back to the office, review it, find missing data and then have to get permits to go back to the site to collect the missing data. When working within a 90-day time frame for completion of a project, doing this work manually can be burdensome and questions persist about accuracy of the data collected.
“Typically, in the software industry they say you can have two of: faster, cheaper or better,” said Meiyappan. “In the telecom industry as you go to implementations, I think you can have all three. It can be faster. It will be cheaper if you do automation and it will definitely be better because you get really accurate information at your fingertips throughout the entire system. You collect data only once. It’s never entered again. … Automation where you collect the data with the device in the field and it goes all the way to the cloud to the operator database, that would be the Holy Grail.”