The Driving Technologies for Network Operators in 2018

If you’re a tech analyst, you just have to do a blog on what to expect in the coming year, no matter how overdone the topic might be.  OK, here’s mine.  What I want to do is look at the most significant trends and issues, the ones that will shape the market for years to come and also establish vendor primacy in key product areas.  I’ll also make note of a couple things that I don’t expect to happen.  Obviously, this is a prediction, but it’s based on modeling and information from the buyers and sellers.

The top issue on my list is SD-WAN.  It’s important because it’s the first broad initiative that separates “services” from “infrastructure”, and that’s critical for transformation of the operator business model.  While the term doesn’t have quite that vague but positive meaning that we’ve come to know and love from other tech developments, there are significant differences in the meaning, application, and implementation of the concept.  In 2018, the pressures of a developing market are likely to start narrowing the field in all respects.

SD-WAN is an “overlay” technology whether or not the specific implementation uses traditional overlay tunnel technology.  You build an SD-WAN over top of some other connection technologies, most often VLANs, VPNs, and the Internet, so it’s not strictly a “transformation” technology with respect to infrastructure.  What an SD-WAN network offers is an independent service layer, a totally controllable address space on top of any supported combination of connections, in any geography.  Because you can support most SD-WAN technologies with a software agent, you can add cloud hosts to the SD-WAN without trying to convince cloud providers to build your boxes into their infrastructure.

The concept of overlay tunnels has been around for a long time, of course, so it’s not so much technology that’s going to make a difference in 2018, it’s application.  Business services are an easier target for competing access and service providers, because you can sell to businesses easier.  Try selling home broadband door to door and you’ll see what I mean.  Managed Service Providers have already gotten the message, but in the last quarter of 2017 it’s become clear that the big news is going to be competitive access providers, including cable companies.  SD-WAN, for them, can generate both an in-area business service model without having to redeploy infrastructure, and a global service footprint exploiting someone else’s access.  This is an irresistible combination.

SD-WAN isn’t just for business services, either.  You can use overlay technology for any connectivity mission, for video delivery, for a form of slicing of infrastructure, and as the basis for cloud multi-tenancy.  At least a couple SD-WAN vendors are already seeing that broader set of opportunities, and I think that’s the tip of the iceberg.

One reason is competitive pressure.  SD-WAN is a sure bet for cable companies or any operator who has national/international service aspirations and a limited access network geography.  We can also already see signs that SD-WAN will be a greater part of telco service plans.  For the telcos, it also offers the opportunity to simplify infrastructure, lower business service costs, and exploit consumer-level broadband access for smaller branch locations, particularly where there’s not a lot of business customers and justifying carrier Ethernet is harder.  By creating an architectural separation between services and infrastructure, SD-WAN makes both more agile, and facilitates a lot of the other market-shaping technologies for 2018.  If even one significant group of operators get the point, everyone else will follow.

Despite the critical value of the technology, winning in the SD-WAN space in 2018 may not be as easy as just tossing a product out there and waiting for someone to notice.  Operators have a different agenda for SD-WAN.  They might want to integrate it with NFV and virtual CPE, for example.  They certainly want to have as much management automation as possible.  They’ll need to be able to link it to current business services, perhaps MPLS, perhaps VLAN, perhaps both.  They will probably want to look at having “interior” network elements that work with edge elements, because that offers them a differentiator.  They may also want to avoid products that have focused on selling direct to enterprises, since these would obviously not offer operators much opportunity.

The next market-shaper in 2018 is zero-touch automation of service and infrastructure processes.  We have been dabbling around the edges of this since 2012, but it’s become clear (again, mostly in the last quarter) that things are finally getting serious.  The TMF has worked on the issue for over a decade, and they have good engagement with the CIOs in operators, but they apparently haven’t been able to move the ball as much as operators want.  If you read the white paper that was issued by the ETSI Zero-touch network and Service Management ISG (ZSM ISG), you’ll see that it overlaps a lot of the TMF ZOOM stuff, and it creates a kind of functional overlay on the NFV ISG.

Technically, zero-touch automation is a collision of a need to support diverse goals and the need to do it with a single technology, a single architecture.  We have operations people who focus on the infrastructure, OSS/BSS people who focus on services and administration, and CTO people who do the standards and propose future technology options.  We somehow have to blend the personalities and the areas they champion, into a single model.  Since we’ve been gradually developing bottom-up stuff in all these areas for years, you can understand how that blending might pose a challenge.

In fact, the biggest challenge the ZSM ISG will have to surmount is standards.  If this is another standards process, it will create media buzz, attract support, spend forever getting tiny details straight, miss the big picture, and eventually lose the audience.  On the other hand, if this body starts looking like a software engineering case study that loses touch with the problem set, it will end up disconnected from the goals that have driven operators to create the group in the first place.  It’s a delicate balance, one that no standards body in two decades has been able to strike.  I can’t promise it will be struck by the ZSM ISG, but by the end of the year we’ll know whether this initiative has failed.  If it does fail, then I think we’ve seen the end of relevant standards for software-centric services.

This is another challenging space for vendors.  Operators have a growing preference for open-source tools in the service lifecycle automation space, which limits commercial opportunity.  They also want full-spectrum solutions rather than components, so it might be wise for any player in the space to look at how they might integrate with ONAP/ECOMP.  That could avoid having to develop a flood of add-on tools and elements, maintain compatibility with vendor offerings, support SDN and NFV…you get the picture.

And speaking of open-source, guess what our next market-shaper is?  Operators have been divided for some time on just how they advance their own cause.  Standards bodies end up dominated by vendors because there are more of them, and because vendors build the products that have to build the networks.  Operators are generally unable, for anti-trust reasons, to form operator-only bodies or even bodies where operators dominate.  There’s been operator interest in open-source software for service management for at least ten years that I’m aware of (I was a member of a TMF open-source group that was launched to address operator interest in the topic back in 2008).  While open-source is a market-shaper, the real insight behind this is AT&T’s “ah-ha!” moment.

AT&T, I believe, recognized that even open-source wasn’t going to do the job, because vendors would dominate open-source projects as easily as standards activities.  Their insight, which we can see in how their ECOMP service management software developed and how their “white-box OS” is currently developing, was to do a big software project internally, then release it to open-source when it’s largely done.  Vendors are then faced with either spending years trying to pervert it, or jumping on board and reaping some near-term benefits.  You can guess what’s going to happen in 2018.

This isn’t going to be a play for the small vendors, unless you want to build into an architecture like ONAP/ECOMP.  The buy-in to participate in the essential industry forums and attend all the trade shows and events is considerable in itself, and then you have to be able to sustain the momentum of your activity.  Historically, most open-source has been driven by vendors who didn’t want to try to sustain critical mass in proprietary software, but recently there has been growing operator interest.  They want to build something internally, then open-source it, which means that it limits software opportunity in the entire space that operators might target with open-source.  Watch this one; it could make you or kill you.

The final market-shaper for 2018 is 5G-and-FTTN broadband services.  While we’ve had a number of technical advances in the last five years that raise the speed of copper/DSL, we can’t deliver hundred-meg broadband reliably from current remote nodes, even fed by fiber.  If there’s going to be a re-architecting of the outside plant for “wireline” broadband, it has to be based on something with better competitive performance.  That’s what makes 5G/FTTN look good.  Early trials show it can deliver 100-meg-or-better in many areas, and probably could deliver at least a half-a-gig with some repositioning or adding of nodes.  It puts telcos on a level playing field with respect to cable CATV technology, even with recent generations of DOCSIS.  Competition, remember?

The important thing about the 5G/FTTN hybrid is that it might be the technical step that spells the end of linear-RF TV delivery.  Cable operators have been challenged for years trying to address how to allocate CATV spectrum between video RF and IP broadband.  5G/FTTN raises the stakes in that trade-off by giving telcos even more broadband access capacity to play with, and if we see significant movement in the space in 2018, then we should expect to see streaming supplant linear RF for TV.

The downside for 5G/FTTN may be the rest of 5G.  Operators I’ve talked with rate the 5G/FTTN-millimeter wave stuff their top priority, followed by the 5G New Radio (NR) advancements.  There’s a lot of other 5G stuff, including federating services beyond connection, network slicing, and so forth.  None of these get nearly the backing in the executive suites of operators, though of course the standards types love them all.  Will the sheer mass of stuff included in 5G standards weigh down the whole process?  It seems to me that the success of any piece of 5G in 2018 will depend in part on how easily it’s separated from the whole.

How do you play 5G?  In 2018, anyone who thinks they can make a bundle on anything other than 5G/FTTN is probably going to be seriously disappointed, but other than the major telco equipment vendors in both the RAN/NR and fiber-node space, vendors will be well advised to look for adjunct opportunities created by 5G.  Video could be revolutionized, and so could business services, and 5G/FTTN could be a major driver for SD-WAN too.  A symbiotic ecosystem might evolve here, in which case that ecosystem could create most of the 2018 and even 2019 opportunity.

Now for a few things that will get a lot of attention in 2018 but won’t qualify as market-shapers.  I’ll give these less space, and we may revisit some of them in 2019.

The first is carrier cloud.  I’m personally disappointed in this one, but I have to call things as they seem to be going.  My model never predicted a land-rush carrier cloud opportunity in 2018; it said we’d add no more than about 1,900 data centers, largely due to the fact that the main drivers of deployment would not have hit.  Up to 2020, the opportunity is driven mostly by video and ad caching, and the big growth in that won’t happen until 5G/FTTN starts to deploy in 2019.  We will see an uptick in data centers, but probably we’ll barely hit my model forecast.  Check back next year!

Next is net neutrality, which the FCC decided it would not play a significant role in enforcing.  There is talk about having the courts reverse the FCC, or Congress changing the legislative framework that the FCC operates under, restoring neutrality.  Possible?  Only very dimly.  The courts have affirmed the FCC’s right to decide which “Title” of the Communications Act applies to ISPs, so that will likely happen here too.  Without Title II, the same courts ruled the FCC lacks the authority to impose the neutrality rules.  Congress has never wanted to take a role in setting telecom policy, and in any event the same party controls Congress as controls the FCC.  The order will likely stand, at least until a change in administration.  How operators will react to it is also likely to be a yawn in 2018; they’ll wait to see whether there’s any real momentum to change things back, and won’t risk adding to it.

Another miss in 2018 is massive SDN/NFV deployment.  Yes, we have some of both today, and yes, there will be more of both in 2018, but not the massive shift in infrastructure that proponents had hoped for.  Operators will not get enough from either SDN or NFV to boost profit-per-bit significantly.  Other market forces could help both SDN and NFV in 2019 and 2020, though.  We’ll get to that at the end of next year, of course.  The fact is that neither SDN nor NFV were likely to bring about massive transformational changes; the limited scope ensures that.  Operators are already looking elsewhere, as I note earlier in this blog.  Success of either SDN or NFV depends on growth in the carrier cloud, and 2018 is too early to expect much in that area.

Were we to see rapid movement in all our market-leading technologies, we could expect 2018 to be a transformation land-rush.  Even just two of the three would likely result in a boom in the industry for at least some of the players, and a single one would be enough to change the game noticeably and set things up for 2019.  In my own view, we should look at 2018 in just those terms—we’re teeing up technology for full exploitation in 2019 and 2020.  Don’t let that futurism lull you into delay, though.  The winners in 2019 and 2020 will almost surely be decided next year, and you’re either in that group by year’s end, or you’re at enormous risk.

I wish all of you a Happy and prosperous New Year.