I've been meaning to write this article for a while now, but I kept putting it off. According to Alexander Rozental, a researcher and a clinical psychologist at the Karolinska Institutet in Sweden, procrastination has four variants: expectancy, value, time and impulsivity. People put things off because they don't think they are worthwhile (e.g. flossing teeth), because they don’t think they will realize the value from a particular task (e.g. learning Chinese is worthwhile but I have left it too late), because the value is too far away in time (e.g. I might have changed jobs or be dead by the time I've figured out what Kubernetes actually is), or because they are impulsive (which sounds like a contradiction in terms).

When it comes to mobile technology generations, telecom operators are highly procrastinational, having been burned by excessive impulsivity back in the 3G spectrum auctions. Some operator CXOs and boards are unconvinced that 5G is worth rushing to invest in. They will be fast followers and let others flatten the learning curve. Others think there is value in 5G but their particular organization might not be best placed to exploit. And the rest think that the benefits from 5G are so far in the future there is simply no need to rush.

As always, there is an element of chicken-and-egg: who wants to build a 5G network if there are no 5G-compatible user devices (not to mention a lack of spectrum in certain countries)? And who wants to launch a smartphone if there are no networks that take advantage of its expensive 5G capability? In Dan Jones's 5G Device Timeline for 2018 & Beyond he notes that Verizon Communications will have 5G CPE from Samsung in 2H18 for residential broadband, and AT&T will have its mobile "puck" in late 2018. Full-blown mobile 5G smartphones are expected to be launched in 1H19.

Procrastination avoidance strategies
There are a number of strategies for dealing with procrastination as outlined in this HBR article. You might visualize how great things will be once 5G is rolled out: remote telesurgery, flying cars, mobile ARPU growth. Alternatively one can pre-commit publicly like the operators in the US, China and South Korea who are all engaged in PR one-upmanship. Another approach favored by the vendor community is to scare operators with the downside of inaction: "OTT will eat your lunch."

Where to start with 5G?
Personally, I find that breaking large tasks down into smaller steps is the key to avoiding procrastination. For telecom operators embarking on 5G one of the first steps they should consider is densifying the network. This involves the addition of small cells and the extension of fiber closer to more cell sites (fronthaul), as outlined in this "how-to" guide from Amdocs: Accelerating RAN densification and 5G deployment. In areas of high demand, operators will need dense networks comprised of macro and small cells transmitting in multiple frequency bands. The need to coordinate radios to minimize interference is driving the development of new Centralized RAN (C-RAN) architectures -- sometimes described as vRAN, Elastic RAN, and Cloud RAN. On a more prosaic level, all these new transceivers will need to be located on new sites leased from landlords or local authorities and supplied with power and fiber cables. Negotiating leases and rights of way for cabling is a lengthy process so operators need to start planning now even if deployments are years off.

NFV drives need for dynamic inventory
As I wrote in the previous article in this series, NFV deployment goes hand in hand with 5G and is key to delivering capital efficiency by deploying capacity as and when it is needed, when it is needed (not just based on long-term planning). Network virtualization will present new operational challenges for service providers. The 5G cloud RAN (C-RAN) will coexist with non-virtualized network infrastructure creating a complex hybrid RAN. Physical network resources such as eNodeB's will need to be maintained in an inventory along with their related logical resources such as link, circuit and port numbers. Virtual network resources such as vEPCs will need to be tracked and managed alongside the physical network in an active inventory, together with the IT cloud infrastructure needed to support them.

To be ready for 5G, the inventory system needs to be as dynamic as the self-configuring network it describes. The virtualized areas of the network will self-configure to respond to traffic demands, network faults, or new service requests. The inventory system needs to be capable of rapid updates in response to these autonomous network changes. The current state of the virtual and logical network layers, and their interrelationships with the physical network underlay are critical for service fulfillment and service assurance. Virtual test agents, element managers and domain controllers operating in isolation will not be able to provide operations staff with the full, end-to-end picture needed for root-cause analysis and fault resolution. Only an inventory system with end-to-end network coverage can provide that picture.

This blog is sponsored by Amdocs.

— James Crawshaw, Senior Analyst, Heavy Reading