Operators’ Technology Plans for 2018: In a Word, “Murky”

We are now past the traditional fall technology planning cycle for the network operators, and I’ve heard from the ones that cooperate in providing me insight into what they expect to do next year and beyond.  There are obviously similarities between their final plans and their preliminary thinking, but they’ve also evolved their positions and offered some more details.

Before they got started, operators told me there were three issues they were considering.  First, could SDN and NFV be evolved to the point where they could actually make a business case and impact declining profit per bit?  Second, was there hope that regulatory changes would level the playing field with the OTTs?  Third, what could really be done with, and expected of, 5G?  They’ve addressed all of these, to a degree.

There’s some good news with respect to NFV, and some not-so-good.  The best news in a systemic sense is that operators have generally accepted the notion that broader service lifecycle automation could in fact make a business case.  The not-so-good news in the same area is that operators are still unconvinced that any practical service lifecycle automation strategy is being offered by anyone.  For SDN and NFV, the good news is that there is gradual acceptance of the value of both technologies in specific “low-apple” missions.  The bad news is that operators aren’t clear as to how either SDN or NFV will break out of the current limitations.

From a business and technology planning perspective, operators think they have the measure of vCPE in the context of business edge services.  They believe that an agile edge device could provide enough benefits to justify vCPE deployment, though most admit that the ROIs are sketchy.  They also believe that the agile-edge approach is a reasonable way to jump off to cloud-edge hosting of the same functions, though most say that their initiatives in this area are really field trials.  That’s because virtually no operators have edge-cloud deployments to exploit yet.

It’s interesting to me that SDN and NFV haven’t introduced a new set of vendors, at least not yet.  About two-thirds of the operators say that the vendors they’re looking hardest at for SDN and NFV are vendors who are incumbents in their current infrastructure.  The biggest place that’s not true is in “white-box” switching, and in that space, operators are showing more interest in rolling their own based on designs from some open computing and networking group than in buying from a legacy or new vendor.

In the NFV space, computer vendors are not showing any major gains in strategic influence, which is interesting given that hosting is what separates NFV from device-based networking.  The reason seems to be that “carrier cloud” is where servers deploy, and so far, NFV is confined to agile CPE and doesn’t contribute much to proactive carrier-cloud (particularly edge-cloud) deployment.  Somewhat to my own surprise, I didn’t see much push behind “carrier cloud” in the planning cycle.  I think that’s attributable to a lack of strategic focus among the computer vendors, and lack of a single decisive driver.

The lack of a decisive driver is reflected in my own modeling of the market opportunity for carrier cloud.  Up to 2020, the only clear opportunity driver is video and advertising, and operators have both regulatory and competitive concerns in both these areas.  Video on demand and video streaming are both slowly reshaping content delivery models, but there seems little awareness of the opportunity to use operator CDNs as a carrier-cloud on-ramp, and frankly I’m disappointed to see this.  I hope something changes in 2018.

On the regulatory side, note my blog on Monday on the FCC’s move.  Operators are both hopeful and resigned on the proposed change, which is somewhat as I’d feared.  They recognize that unless the FCC were to impose Title II regulation on operators, they have little chance of imposing restrictions on settlement and paid prioritization.  They also believe that whatever happens in the US on “net neutrality” up to 2020, it’s not going to reimpose Title II.  Thus, their primary concern is that a change in administration could result in a reversal of the Title II ruling in 2020.  That limits the extent to which they’d make an aggressive bet on paid prioritization and settlement.

The US operators I’ve talked with are cautious about even moving on settlement, fearing that a decision to charge video providers (in particular) for delivery would result in consumer price hikes and potential backlash on the whole regulatory scheme.  Thus, they seem more interested in the paid prioritization approach, offering at least content providers (and in some very limited cases, consumers) an opportunity to pay extra for special handling.

Outside the US, operators believe that if the US applies paid prioritization and settlement to the Internet, many or even most other major markets would follow suit.  However, they don’t think it would happen overnight, and that makes the confidence that US operators feel in the longevity of the regulatory shift very important.

For 2018 and probably 2019, I don’t see any signs that regulatory changes will have a major impact on technology choices.  SDN could be facilitated by paid prioritization, but current plans don’t include SDN because the shift couldn’t be easily reversed if policies changed.  Fast lanes may come, but they won’t drive near-term technology changes.

Any hopes of changes, at least technology changes, come down to 5G.  In areas where mobile services are highly competitive (including the US and EU), 5G deployment may be mandatory for competitive reasons alone.  In the US and some other markets, 5G/FTTN combinations offer the best hope of delivering “wireline” broadband at high speeds to homes and small business/office locations.  All of this adds up to the likelihood that 5G investment is baked into business plans, and that’s what I’ve been told.

Baking it into technology plans is a given at one level (it’s mandated) but difficult at another (what do you bake in?)  Like almost everything else in tech, 5G has been mercilessly overhyped, and associated with a bunch of stuff whose 5G connection is tenuous at best.  Let me give you some numbers to illustrate this.

Of the operators who’ve talked to me on the topic, 47% say that there’s a credible link between 5G and NFV, and 53% say 5G doesn’t require NFV.  On SDN, 36% say there’s a credible link and 64% say there isn’t.  In carrier cloud 68% say there’s a credible link to 5G and 32% say “No!”  Fully 85% say that you could do 5G credibly with nothing more than changes to the radio access network (RAN).  So where does this leave 5G technology planning?

5G core specifications won’t be ratified for almost a year, and it’s not clear to operators how much of the 5G capabilities that are then finalized in standards form will be translated into deployed capabilities, or when.  Much of 5G core deals with feature/service composability, and some operators argued that this sort of capability hasn’t been proved in the higher-value wireline business services market.

Where this has left operators this fall is a position of management support for 5G deployment but only limited technical planning to prepare for it.  The sense I get is that operators are prepared to respond to competitive 5G pressure and do what the market demands, but they really hope (and perhaps believe) that 5G won’t be a technical planning issue before 2020 or even 2021.

Across the board, that same confusion seems to hold.  In fact, this year’s planning cycle is less decisive than any I can recall in 30 years, though some of the winning technologies of prior cycles never really made any impact (ATM comes to mind).  Might the lack of direction, the emphasis on response rather than on tactical planning, be a good sign?  Perhaps the market can pick better technologies than the planners, and it appears that for 2018 at least, the planners are looking to the market for direction.