Having invested big bucks to become a software company, Calix seems poised for a major payday and a chance to box with the big boys in access networking.

June 28, 2017

17 Min Read
Calix Gambles on a Software-Defined Reboot

In early June, at the historic St. Francis Hotel in San Francisco, Calix did something its executive team says it hasn't done in the company's 18-year history: gather an array of industry analysts and trade media to hear a company presentation. The occasion? What they believe is groundbreaking technology and business change within the company.

The event underscored two important aspects of what's happening at Calix Inc. (NYSE: CALX): First, that Calix's move to software-defined access has been a long time in the making but is now fueling freaky-fast new product introductions and expanding the company's potential market. And second, that Calix needs more people to understand and acknowledge the significance of what it's done in order to capitalize on its apparent market advantage because, to date, it remains the smallest player in market capitalization, by a wide margin, in an ever-shrinking field of access vendors.

By CEO and President Carl Russo's measure, Calix has spent the last decade, and particularly the last six-and-a-half years, transforming itself from a wireline access systems company to "a communications software and services company that, oh by the way, ships systems as well." The significant investment in time and talent -- Calix now has top executives it has plucked from Juniper, Commscope and Salesforce.com, among others, and a core group of developers it acquired with Occam -- has not made the company a Wall Street darling, however, and may have had the opposite effect.

Figure 1: Calix President and CEO Carl Russo Talks to Analysts

"Wall Street is waiting to see the value in terms of customer wins and revenues," says Julie Kunstler, Ovum Ltd. senior analyst. As a technology analyst, she considers Calix's new software-defined access platform to be "highly differentiated."

"There was a monumental effort. It was very well designed and will serve their customers very well," Kunstler comments.

Independent analyst Matt Davis heard what Calix executives had to say, but also saw it underscored by open praise from large customers or potential customers -- namely Verizon Communications Inc. (NYSE: VZ), via a Calix press release, and CenturyLink Inc. (NYSE: CTL), in the person of Frank Miller, vice president of architecture, speaking to the media/analyst event.

"I am normally very wary of words like first, newest, or unique," he comments. "What makes me more confident that they do have a time advantage over their competition is listening to their customers. It is often hard to get a top network executive from a company like Verizon to be quoted in your press releases and be on stage at your customer events, like CenturyLink was. To have them come out and talk about the differences in terms of speed-to-market this thing can enable gives me confidence that this wasn't just a line of bull."

Kudos from customers and analysts are great, as are significant revenue bumps in the last two quarters and a projected 10% revenue bump this year over 2016. But Calix continues to post losses, based on its higher expenses, and expects that trend to continue through 2017. Its market cap of just under $350 million is about a third of Adtran Inc. (Nasdaq: ADTN)'s at $974 million and a much smaller fraction of the bigger players in the field including Huawei Technologies Co. Ltd. and Nokia Corp. (NYSE: NOK).

So the company that survived the telecom bust of 2001, outlasting literally dozens of other access companies, must now prove its bold gamble on software-defined access can pay off and it can not only survive but thrive going forward.

Next page: The making of AXOS

The making of AXOS
Calix's AXOS, or Access eXtensible Operating System, made its debut last October as the vendor's software-defined access platform, and is the core system that will enable Calix "to introduce more products this year than in the course of our 17-year history combined," says Geoff Burke, the company's senior director of marketing. (See Calix First to Launch Software-Defined Access.)

Depending on who you talk to at Calix, AXOS has been four, six-to-seven, or even more years in the making but everyone agrees it is the key element in the company's strategy, along with the Calix Cloud, which already has 400 customers for its analytics and services, and the E-series, the systems that deliver copper-based and fiber-based services options.

At its core, AXOS is a software-defined access platform that separates hardware from software, uses Netconf/YANG to communicate with any operations and support systems, has a modular architecture that allows new service modules to be rapidly added, and can support any physical-layer access technology -- fiber, copper, coax or wireless -- and any SDN controller. Because it uses Netconf/YANG, AXOS can communicate with any orchestrator, so that it enables key pieces of a network operator's access sector to be interchanged without complete disruption.

Figure 2: AXOS Architecture

AXOS is strategically positioned to support the carrier move to consolidated access, putting business, residential and backhaul services onto one fiber infrastructure, using different wavelengths, while still supporting copper, coax or wireless access for final connections to customers.

As Russo told the press/analyst meeting, AXOS grew out of thinking that for him started in 2007. Not long after the iPhone was introduced, Russo said he realized that networks and business models would change dramatically going forward, driven by "two irresistible forces," namely the amassing of compute and storage within the data center to deliver content and applications, and the increasingly device-driven personalization of applications that supplanted switch control of services. The network that sits between them, regardless of the physical medium, is being pulled into one simple unified access infrastructure, according to Russo.

"And that infrastructure will be all services over IP over Ethernet over fiber over wireless," he said. Setting aside for the minute that some customers will be so big that they need direct fiber connections and some so sparsely spread about that fiber will never reach them, this paradigm works for most of the planet, Russo said.

This new network paradigm essentially looks like a global client-server network where the server is the cloud and the client is the consumer device, he added. Focusing Calix on the converged network access that would connect the two, the metaphor became clear: "What we needed was a great outdoor access LAN, which conveniently became our GOAL," he said. "We wanted to figure out how to build that great outdoor access LAN that allowed all the devices and the subscribers to connect to all the content and applications in the cloud."

Key attributes for that GOAL would be that it is always-on, i.e, hitless, self-aware, simple to use and easy to modify, Russo explained. Calix's first attempt to build that architecture was its Ethernet eXtensible Architecture (EXA) for its E-series products, and involved the critical step of moving its gear off custom ASICs and onto merchant silicon. But it didn't go far enough, he said.

"We have been a merchant silicon company for the last decade," Russo said. "We haven't really talked about it. If you test that against others you know, you'd find that's quite unique -- it brought a lot of modularity, but we still had a real-time complex asynchronous communications system."

Then in 2010, Calix bought competitor Occam Communications, known for being the only access vendor that was Internet Protocol-based. Through IP, Occam had figured out how to limit the variety of use cases and simplify the system so you could "plug it in and walk away," Russo said. It was a restart for the team designing GOAL, using five engineers from Occam and one from Calix beginning in 2011. And thus AXOS was developed -- carefully so, from a fresh sheet of paper, he added, separate from the Calix product lines of the time.

Next page: But wait, everyone else has software-defined platforms, too!

> But wait, everyone else has software-defined platforms, too!
In creating AXOS, Calix started from the ground-up to change the way it operated as a company, as well as the way it developed products and worked with its customers.

"We actually have become a software company," says Michael Weening, the former Salesforce.com Inc. and Microsoft Corp. (Nasdaq: MSFT) executive who joined Calix as executive vice president of sales and marketing. "We are bringing capabilities that currently exist in other places in the network, like in the data center and at the edge, and we are bringing them into an intelligent access network in ways that have never been done before."

Hardware design and engineering became part of its manufacturing process. "We essentially white-boxed ourselves," Russo commented.

Calix executives insist that others, such as their arch-rival Adtran Inc. (Nasdaq: ADTN), are creating middleware as an abstraction, or "a shim," to achieve what they are calling software-defined access and not building a system from the ground up and rebuilding the way their companies operate as well. Weening is quite passionate on the subject.

"What does it mean to build an access system from the ground up -- all the components parts?" he said. "It's not middleware. We are not taking an old system and putting something on top of it that acts like a shim and hides complexity and allows you to take a step without making a change. Where the rest of the industry has gone down the middleware path, we have built AXOS from the ground up. And it's the significant investment that has cost us hundreds of millions of dollars."

Ovum analyst Kunstler agrees that what Calix is doing is "either extremely evolutionary or revolutionary, somewhere between there," with the proof being the pace at which the company is now rolling out new capabilities, such as the 10G-EPON option for cable that it added to AXOS in a 120-day development period. (See Calix Crafts 10G EPON Backing for Cable.)

In fact, the product pipeline has been burning up a bit lately at Calix, including the first G-fast product certifications from the Broadband Forum , NG-PON 2 channel bonding, a new Subscriber Management Module, Layer 3 routing via the AXOS RPm module, and a commercially hardened version of Residential CORD (central office re-architected as a data center) done with Radisys Corp. (Nasdaq: RSYS) using the AXOS OFx (OpenFlow) Connector it developed with ON.Lab in less than four weeks. (See Calix Claims G.fast Certification First, Calix Combines Edge, Access on AXOS, Making R-CORD Production Ready, Calix Touts Channel Bonding for NG-PON2, Calix & Radisys Bring R-CORD to Market and Calix Connects With CORD, ONOS Networks.)

Figure 3: Calix Product Flood

Teresa Mastrangelo, BroadbandTrends analyst, also sees Calix's rapid product ramp as one proof point that the company has done something different from its competitors, including Adtran, Huawei and Nokia.

"Everyone else has open APIs and basically another layer to add onto their OSS infrastructure to provide the hooks so they can sort of do SDA but not without all the features that Calix can enable," she says.

Those vendors are likely to say, however, that they are responding to what their customers want, Mastrangelo adds. "They will argue that customers want a hybrid system because they want to move slowly."

Because Calix was building AXOS from the ground up, it chose to build an operating system that was agnostic to the physical medium underneath, able to work with any orchestrator and also open to anyone's SDN controller, said Shane Eleniak, vice president of systems products. All of that is based on understanding that its customers want more automation in the access, but also more control over the customer experience and the ability to make changes in the network without disturbing services or having to do massive regression testing and re-integration.

"The hard part of changing anything in the access is getting it integrated into the back-office systems," Eleniak said. AXOS and its AnyPhy approach allows the services to stay the same, even as the underlying access technology evolves, because there's always something better around the corner.

Figure 4: The Strategy in One PowerPoint Slide

Another advantage is the ability to do real-time analytics -- something the Calix Cloud supports -- to know what's happening in the network at a granular level, which will help operators support a better quality of customer experience but also open the door to potential new service options, he adds.

That becomes more important, Mastrangelo notes, as broadband penetration rates hit a level at which adding lots of new customers becomes harder. Differentiating services and being more easily able to try new options will be a bigger priority for network operators going forward, she notes.

"This market is not about growth in subscribers -- it is already pretty highly penetrated -- so now it's about better service, better applications, better things," Mastrangelo says. "What Calix is doing is giving operators a tool to play around with stuff and turn up a service quickly, and if it's failed, you pull it down, no harm done."

CenturyLink has used Calix's AXOS system as part of its G.fast service rollout in Minnesota, where it is also using Platform CenturyLink, its multi-domain orchestration platform that, like AXOS, using Netconf/YANG models and open APIs, to automate service processes, Frank Miller, vice president of access at CenturyLink told the Calix media/analyst gathering. The longer-term goal is to give customers control through a portal of service changes, he said.

"Our number one goal is customer experience, because we live in an Amazon universe right now -- you press the button and something happens," Miller said. "We should also be able to expose our APIs to the largest consumers of the world, so they can get instant gratification."

Figure 5: Talking Turkey Calix's Geoff Burke (left) questions CenturyLink's Frank Miller at analyst conference. Calix's Geoff Burke (left) questions CenturyLink's Frank Miller at analyst conference.

One other impressive aspect of the AXOS strategy, Mastrangelo says, is the AXOS Sandbox -- downloadable software that runs on a PC and can be used by operators to test and use software in advance of the hardware delivery so they can begin the integration process with their individual operations.

Analyst Davis also sees advantages to AXOS when it comes to being able to make changes or additions to the network without the concern of impacting overall performance or negatively impacting another part of the network. "If you look at the maintenance windows these companies have labored under forever, this could eliminate the potential of them taking down other applications or the network when changes or upgrades are made."

Next page: Where does Calix go from here?

Where does Calix go from here?
One thing Calix executives made clear at the St. Francis meeting: They are expecting big things to happen for the company now that the AXOS platform is in place.

"They've gone all-in on software-defined access," says analyst Mastrangelo. "This is their future."

Immediately, it seems to position Calix well to compete for the one major North American piece of business -- Verizon's NG-PON2 initiative -- once that gets going, the three analysts agree.

"There is every indication that Verizon is going ahead with NG-PON2, and I think Calix will get a piece of it, but how big a piece compared to Adtran, it's hard to say, or even whether Nokia will come in and take some of it," says Kunstler.

Both she and fellow analyst Davis think the industry's movement toward converged access architectures -- combining business, residential and wireless backhaul services onto one network -- also plays to Calix's strengths.

"As bandwidth demands continue to grow -- for either of the three -- why not do it on a single network and why not make that a PON network? And XGS-PON, which is 10Gbit/s symmetrical, can support all of it," Kunstler says. "It cannot support a data center that needs hundreds and hundreds of gigs, but most businesses don't need hundreds of gigs; they need 2 gigs or 5 gigs."

And while PONs weren't initially considered feasible technically for mobile backhaul, that is now a generally accepted practice, and with the rise of small cells and 5G, those backhaul demands will continually grow, she notes.

And then there's the "copper gang," as Russo refers to them, a set of North American operators just below the top two, including CenturyLink, Windstream Communications Inc. (Nasdaq: WIN), Frontier Communications Corp. (NYSE: FTR), TDS Telecom , Consolidated Communications Inc. and Cincinnati Bell Inc. (NYSE: CBB)

"They have the biggest challenge of the wireline service providers," Russo says. "They are publicly traded, they have lots of copper, and they have a heavy regulatory regime."

Calix is working to convince this group of customers that they can "invest tactically and make it work strategically" by putting AXOS into place because, as a non-blocking modular system that lacks a big backplane and shared switch, the platform makes it possible to "buy what you need when you need it, " Russo says. "Not only is our hardware built that way, our supply chain is built that way."

Davis believes the Calix story will be compelling to this group of customers as they work to push fiber deeper into their networks even if that last "pigtail" remains copper. In addition, he notes, the fact that Verizon is pushing for a converged access network -- one network to handle business, residential and backhaul services -- plays to the strengths of a platform like AXOS, and is something other operators are also looking to do.

The question then becomes: Which other markets, outside the US where Calix has been strong in the rural market, might AXOS enable? China is clearly out -- hello, Huawei and ZTE -- but Europe, other parts of North America and the Caribbean are all potential growth spots.

"Calix has a foothold in Canada, the Caribbean and Europe and I wouldn't be surprised to see them have a chance with some of the large PTTs in Europe," Davis says.

Asked about expansion outside North America, Russo didn't sound wildly enthusiastic, saying his company has "a mountain of opportunity" in North America to address before getting too concerned about further expansion. As Kunstler points out, Calix bet early on PON and didn't have the strong DSL chops of its rival Adtran, so to some extent, it has been waiting on the North American PON market to get moving, which is it finally starting to do.

There is also the prospect the company could be bought while its market cap is still relatively low, but some of the bigger players who might have been considered acquirers don't seem to be hungry to venture into -- or back into -- the sometimes messy access market.

And from all outward signs, Russo seems ready to finally capitalize on the investments his company has been making for at least all of the current decade. He sees Calix as as a comparatively small company, able to grow significantly by moving to software and services, and not on of the telecom behemoths struggling in this new paradigm.

"We're a disruptor at $500 million; we are not a disruptee at $20 billion," he said. "We only have to find another $500 million in business and we are double the size we are [now]."

And he clearly thinks that is where Calix is headed.

— Carol Wilson, Editor-at-Large, Light Reading

Subscribe and receive the latest news from the industry.
Join 62,000+ members. Yes it's completely free.

You May Also Like