When it comes to adoption of SDN, every network operator has its own approach, even though all are ultimately headed in the same direction. Level 3 Communications' CTO Jack Waters underscores this reality in discussing his view of SDN, which leans heavily on the advantages it brings to network provisioning.

In a recent conversation following a meeting in Chicago with Level 3 Communications Inc. (NYSE: LVLT) customers, Waters shared his views on SDN, the market in general and the upcoming challenge of integrating tw telecom inc. (Nasdaq: TWTC)'s network (if the acquisition announced in June goes through as planned).

He sees SDN as presenting two possibilities: First the separation of the control and data planes for big iron within the network, the switches which live at the heart. That's "interesting," says Waters, but more appealing to big data center operators in the near term. The second SDN use case, transforming the provisioning process using Netconf/Yang, which was embraced by the Open Networking Foundation for configuration of OpenFlow enabled devices, has much greater appeal. (See Netconf & Yang Go Mainstream.)

"That piece of it has more relevance for us," Waters says, because it enables a common data model that can be abstracted from vendor-specific provisioning processes and allow the network operator to have a common way of provisioning across its network based on service definitions that specific network configurations.

Figure 1: Level 3 CTO Jack Waters

"That is really interesting because today, the provisioning process is not as abstracted, it's more specific to each vendor's equipment," he says. "If that part takes hold, it changes our business model."

Greater customer control
When NFV is added to that more abstracted provisioning process, the network operator is in a position to give its customers the ability to turn up specific functions -- such as firewalls, for instance -- in an automated, on-demand service model. At the very least, the network operator's own provisioning processes become much more automated and flexible, he says.

"Customers don't get enough control and they are asking for it," Waters says. "I don't think the economic model of how they pay for bandwidth or for these virtualized network functions has been worked out yet. But we know that is where we are headed."

Our "NFV and the Data Center" event digs deeper into how telecom service providers are evolving their data center strategies. You can check it out here on Light Reading.

The biggest challenge to getting there is having the hardware vendor support of the software interfaces needed to support the more abstracted model, and on that front, Waters admits some mild frustration with vendors who purport to be creating a more open interface to their equipment but wind up adding proprietary extensions or other "enhancements" that become roadblocks to the larger process of creating that abstraction layer.

"We'd love to see more open interfaces -- at least for the stuff we want to use," he says. "Things tend to remain open as long as there is a pretty level playing field. But once there is a clear winner, things start to bend in that direction."

Vendors still tend to develop proprietary extensions, even to open and widely used standards such as BGP, Waters notes. "We try not to corner ourselves and have any proprietary extensions force us to have to buy from a specific vendor," he comments.

In general, he sees virtualization efforts by vendors as still in their very early stages and "not fully baked yet," but is more intrigued by some of the pure software plays he is seeing than in hardware-software combinations.

Next page: Merger mania Merging networks
After nine network acquisitions by Level 3 in nine years, Waters is an old hand at the challenge of merging network operations, as he is likely to soon be doing when the acquisition of tw telecom is complete. The back office integration remains the toughest challenge and Level 3 has taken its knocks in the past for not managing integration as fast as its investors had hoped to see synergies driving up revenues and profits. (See Level 3: 'Size Matters' Isn't Enough, Level 3 to Acquire tw telecom for $5.7B and Level With Me: When Rumors Get Real.)

"Every company [Level 3 has bought] is different, but we are experts now at interconnecting networks and finding the network synergies. We've learned every time, and we have a good idea of what works and what doesn't."

At this point, there are even written playbooks, assembled in past integrations, on different aspects of the integration process. But that doesn't mean anyone thinks it's going to be easy. The management team has been announced and one of the first steps once the merger is complete will be deciding on a set of products to take forward to the market and prepare the sales staff to sell. (See Level 3 Unveils Post-Merger Management Team.)

Interestingly, both Level 3 and tw telecom have endeavored in recent years to develop flexible approaches to delivering bandwidth, trying to match on-demand cloud computing with on-demand networking. Level 3 offers Cloud Connect, a pre-provisioned connection to popular clouds such as Microsoft Corp. (Nasdaq: MSFT) Azure or Amazon Web Services Inc. or to interconnection sites such as Equinix Inc. (Nasdaq: EQIX) or Digital Realty Trust Inc. , via links that can be dynamically configured and are billed on a usage basis. (See Level 3 Offers Private Links to Cloud and Level 3 Cloud Connects Digital Realty.)

The tw telecom approach is more about bandwidth on demand, which is just a different part of the same spectrum of services, Waters says. "As we get to know more about their offer, we'll decide how to move forward, but we don't think there is anything substantially different." (See tw telecom Changes CoS Dynamics, tw telecom Delivers On-Demand Ethernet and ESDN: tw telecom Stresses Customer Control .)

Waters doesn't gloss over the potential challenges ahead, even though as a "pretty mature operating environment," tw telecom may not pose some of the problems of previous integrations. The hard thing is always the merger of back office systems, which never happens overnight, he says. And until the merger is complete and legal, sharing of information is limited so he doesn't yet know that real challenges may still lie ahead.

Stronger enterprise focus
Level 3 shifted its focus to the enterprise market about 18 months before Jeff Storey took over as CEO in early 2013, Waters says, and currently gets 65% of its business from that segment, a switch from its earlier days when wholesale was the bigger nut. It was a slow process to change the company's emphasis, but that is expected to accelerate further with the tw telecom buy, given that company's tight focus and substantial success in the enterprise space.

Waters sees enterprises in the midst of a technology transfer right now, with more of them shifting to cloud services and to SIP trunks or MPLS connections to an IP-PBX or even a hosted voice service. Level 3 is well established in delivering SIP-based services and has been heavily focused on connecting data centers with fiber to support cloud offerings. In addition to the four companies already on Cloud Connect, the carrier is pursuing other cloud connections as well as it expands.

"We are more interested in getting it right -- Cloud Connect requires more tightly integrated connections so that the set-up on the customer's end is more automated," he says. Because its network has global reach, Level 3 can offer cloud service delivery to multinationals -- something tw telecom couldn't offer on its own network -- and this is an area where the combined entity could really thrive, Waters says.

"TW has a great enterprise business but they don't have the global footprint to offer to that existing customer base and we do," he says.

Oh, and that nasty business with Verizon…
No conversation with Level 3's CTO would be complete without some discussion about the finger-pointing going on over where the network congestion really occurs when it comes to delivering Netflix Inc. (Nasdaq: NFLX) traffic over the access lines of certain broadband ISPs. Level 3 is one of the major transit networks for Netflix, along with Cogent Communications Holdings Inc. (Nasdaq: CCOI), and has long maintained the problem lies at the interconnection points between transit and local ISP networks, which need to be upgraded on the local ISP's side to prevent slowing of Netflix streaming video.

Waters sticks to that song -- Level 3 believes it would take a small investment on Verizon Communications Inc. (NYSE: VZ)'s part to unclog the interconnection points and allow the Netflix videos to stream more effectively over the FiOS network. (See Blog Wars: Verizon, Level 3 Duke it Out.)

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