5G development should take to heart a number of lessons that the industry has learned from previous generations of technology, former CTO of the Federal Communications Commission and Columbia University professor Henning Schulzrinne told the audience at last week’s Brooklyn 5G Summit.
Schulzrinne, who currently serves as a technical advisor to the FCC, pointed out that the telecom industry doesn’t have a great track record of being able to accurately predict which use cases will become major drivers for a new generation of technology. Previous generations of mobile technology were typically built with some general expectations for what would be the primary jump forward, Schulzrinne said — and the industry ends up being surprised by what actually ends up spurring usage and adoption. In 2G, carriers touted the improved quality of digital voice — but the surprise was the quick adoption of SMS. For 3G, he said, operators looked to WAP to boost usage, and found that users really wanted access to the open Internet rather than walled gardens. For LTE, operators wanted to leverage carrier-centric IMS, but instead usage has been dominated by over-the-top apps such as YouTube, WhatsApp and others.
Schulzrinne said that if he had to make a prediction, he thought that for 5G, the major driver won’t necessarily be the quality of service-focused paradigms that industry is currently focused on, such as very low latency and high reliability for mission critical use cases. Instead, he thinks that the ability to deliver high capacity at low cost is more likely to drive the technology shift.
Among his advice for pitfalls that 5G technology should avoid:
- Avoid complexity. Voice over LTE and IMS have taken far longer to roll out thatn expected, he noted, because IMS is “immensely complex” — in part because it ties the applications to the underlying infrastructure, with the goal that only incumbent carriers could offer IMS services. Those hooks to infrastructure ultimately slowed operators down with difficult, expensive roll-outs, while OTT services leaped ahead and drove adoption. “If you want to kill a good network idea,” Schulzrinne said, make it so complex “that only one or two vendors can understand it.”
- Plan inter-carrier interfaces, to avoid roaming issues that have impacted previous technology generations.
- Keep the network application-neutral.
- Take a cue from fiber-to-the-X installations and reuse backhaul where you can find it.
Schulzrinne also said that a carrier-centric, vertically integrated network model where a single entity owns the spectrum and the network infrastructure, has stores and branded devices and authenticates users — the model for 1G to 4G — is not necessarily the most efficient or effective model for 5G but still seems to be the default assumption as the technology is being laid out. Every layer of 5G could potentially be operated by a different player, he said, and “5G should work just as well as if it was integrated. We’re not there, but I’m afraid we’re missing that chance with 5G.”
