Private networks: All about coverage and capacity, for now, and LTE fits the bill
HANOVER, Germany—For all of the hype around tailoring industrial networks to meet the most demanding applications and connecting all sorts of automated and intelligent “things”, private networks are, at this point in their development, largely being driven by companies wanting dedicated coverage and exclusive capacity, and to connect human workers, says Jagadeesh Dantuluri, GM of private and dedicated networks at Keysight Technologies.
In most cases, private LTE works just fine for meeting current needs, he continues, speaking during a session at Hannover Messe. That’s not to say that there is no network tailoring going on, but from Dantuluri’s description, it tends to be taken care of with some tweaking of LTE, perhaps along the lines of optimizing the uplink to cope with on-site data generation (public cellular networks are always optimized for far faster downlink speeds than uplink speeds).
Dantuluri says that the most common use cases Keysight sees for private networks, accounting for something like 90% of them, involve connecting people whether that means tablets for workers or things like IoT wearables that support safer working conditions. Physical security, such as connecting security cameras, also does get enabled by private networks currently, while he categorizes things such as process automation for industrial manufacturing, remote control of machines and AR/VR as future use cases, rather than actually being used in a practical way by enterprises at this point in time.
The challenges that are in the way of more widespread adoption? Integration with legacy systems is the first and biggest, he explains, followed by cost.
“Just because 5G has come, doesn’t mean people will leave everything and come to 5G,” Dantuluri says. It takes time, effort and money to deploy private 5G, when most companies already have a combination of Wi-Fi and wired connectivity that works. “All those systems will not go away, even with 5G,” he adds. “They will co-exist.” And if companies really do want to get into cellular, well, 4G is a lot more mature and less expensive.
To those challenges, Dantuluri adds: A lack of industrial-grade 5G chipsets and devices, lack of cellular system familiarity/expertise in enterprise IT departments, and depending on the region and/or company, lack of spectrum resources. He also points out that a number of practical questions have yet to be answered: For instance, who guarantees the potentially stringent Service Level Agreements that private networks promise to deliver: An operator, perhaps—but what if, as in Germany, enterprises don’t need an operator because there’s a dedicated spectrum allocation? Will it then be an equipment provider, or a managed solution provider, in the case of networks-as-a-service?
Fundamentally, Dantuluri explains, a private network’s traffic is going to look radically different than a public cellular network and will need to be configured, tested and monitored accordingly. An automated mobile robot (AMR) on a factory floor, for example, is going to be constantly sending and receiving large amounts of data in real-time or near-real-time—which is not how a typical smartphone operates, Dantuluri points out. Someone will need to emulate and test in order to figure out things like, how many AMRs and other device-types that an enterprise’s network can reasonably support, and map the necessary network conditions to KPIs for SLAs and monitoring.
But at an even more basic level, Dantuluri says, for many verticals, private 5G is still an unproven technology. They may have heard of it, they may be interested in learning about it, but they don’t know whether it will actually solve the problems they want it to solve, and how it will work compared to their current connectivity set-ups. Partnerships, education, demonstrations and overall “confidence building” will help chip away at that, he says, but it will take time for industry verticals to understand and buy in to 5G as a solution for their particular needs.
“The most hyped-up use cases still require some time to sort out all these issues,” Dantuluri says.