In December, Ciena created a new business division, Ciena Agility, specifically targeted at the growing opportunities of Software-Defined Networking and Network Functions Virtualization. With the introduction of itsAgility Matrix software and VNF marketplace, Ciena is looking to help network operators unlock new business models serving the enterprise and enable a broader ecosystem of virtual network function vendors. With us today to talk about Ciena’s efforts in NFV and SDN in general is Chris Janz, VP and CTO of Ciena Agility.
TR: Ciena’s Agility portfolio covers both SDN and NFV. Do you see these technologies as tightly interwoven or more as separate silos?
CJ: I wouldn’t say they’re totally separate silos. We’ve got a pretty heavy focus on provider managed services, especially touching enterprise and that whole interesting area around the virtualization of IT, which enables IT-as-a-service and leverages cloud connectivity, driving new requirements for dynamic networking. Things we did with our V-WAN product early on the SDN side of our portfolio really had that focus. What we’ve done with Agility Matrix is a nice complement and really in a similar wheelhouse. With Agility Matrix, we’re focused on how to enable the on-demand aspects of managed services beyond connectivity with VNF functions that we can drop into the service path. Matrix is distinct from the SDN pieces of our portfolio in that it can be deployed alone and, in fact, “box-free”. It can be delivered outside the framework of a connectivity managed service, or it can be bundled with one and coordinated with the SDN side of the picture. . These examples show: SDN and NFV are related in many ways, but they’re not the same thing.
TR: How far off do you think we are from broad adoption of NFV in service provider networks?
CJ: Adoption has already begun, but it depends by what you mean by broad adoption. It’s definitely going to be a process over time to get to the point where we’ll be able to say that 50 percent of the network or more is really an SDN/NFV architecture and operations stack, versus existing software operations and network infrastructure. That’s not a one- or two- year thing. But what we are seeing in 2015 is the deployment of some NFV capability in specific service frameworks as well as early SDN-related functionality deployments. We’re seeing RFX activity from a broad selection of service providers pointing at specific projects touching different parts of the network and software, and looking for different benefits. 2015 is probably not the year of very significant operationalization, but it is very clearly shaping up to be the year where we knock down a lot of the preambles to that. It’s very real and getting a lot of serious effort from across our customer base.
TR: Are their particular VNFs that the sector is especially excited about in the short term?
CJ: There are, absolutely. When you look at the managed services to enterprise context, virtualization of the edge router is an obvious focal point. More broadly, the kind of appliances that tend to need to be present for dealing with critical traffic like encryption engines and WAN optimization/acceleration,—we’re definitely seeing interest in that as well.
TR: Why choose the role of coordinator rather than develop your own VNFs?
CJ: When we looked at the picture, we asked ourselves, “How can we best contribute?” Every one of these applications is a specialized engineering problem, and the spirit of the whole movement is to get to a software-enable environment where the power of an ecosystem can be readily brought to bear. What we’re doing is enabling the operationalization of services on a mechanical level, as well as on an economic and business model level, to connect service providers to a community of application vendors that specialize, to a large degree, in what they are producing. True – we are not building the VNFs, but we are facilitating the availability of a best-of-breed library for service providers and provide the glue to make the model work for everyone concerned.
TR: Is it challenging to be at the center of an ecosystem where you have to coordinate the delivery of software from so many vendors with different agendas and schedules?
CJ: It’s not as challenging for Ciena as you might think. The problem of making an appliance, i.e. creating the software version we can deploy on a VM or container, is a problem that is tackled by the vendors, and most have already addressed the challenge or are on the road there. None of them are coming to the virtual delivery of their product cold and just rolling up their sleeves now—far from it. Of course, there are some things related to platforms and interfaces that we have been working with the lead set of partners on, and some of these things have involved real invention. What has excited this community is the prospect of a whole new business model and opportunity to sell to a new set of customers with a low barrier to entry and cost.
TR: How does the revenue model work for carriers and vendors with Ciena’s Agility Matrix? Who gets paid when?
CJ: That’s where I think some of the more interesting thinking and possibilities are focused. If you take Agility Matrix out of the picture, the normal method of offering these types of services is for a provider to offer a prepaid service model to its customer. To do this, the service provider must select an appliance, and then sign a contract for a number of months or years for a price. The appliance vendor then ships the appliance to the service provider, and together they’ll endure a lengthy testing and qualification process, and slow individual instance installations. Even when physical appliances are replaced with virtual inventory in the NFV equivalent, the service provider is still looking at buying blocks of licenses up front from the vendors. That may result in a lower cost of investment, , but the provider is still in the game of trying to guess which appliances they’ll need, and at what rate of speed the services are going to be of interest to their customers, and then make that upfront investment. Investment risk remains. With Agility Matrix, the prepaid model is still available of course, and it needs to be there because a known way of working is the path out of the gate in lots of places. But Agility Matrix also offers another model: the possibility of a postpaid, or pay-as-you-earn model. You can almost think of it as a consignment model, where the provider is really not making any up front investments in any particular volume of instances of VNFs. At the point where there is actual consumption by an actual customer, that’s where the provider starts owing money to the vendor. And the vendor benefits from being able to offer their product at low cost to a volume of service providers with a single point of contact and a fairly low barrier to entering the market. And Ciena doesn’t compete with them on what they do. Overall, the postpaid model and pay-as-you-earn aspect are what has really excited both the providers and the appliance vendors we are working with.
TR: Do you see any major hurdles left technologically before Ciena’s Agility Matrix is ready to fulfill expectations?
CJ: We have a clear product set that is moving to release, a partner framework behind it, and a clear operational model, along with a number of engaged customers. There’s a lot of lifting, but I couldn’t point to anything as a particular technology heavy nut remaining to crack.
TR: So much of what NFV seems to be offering is about performing existing tasks in a new framework. Do you think there are entirely new applications that NFV will bringto the table?
CJ: That’s a really interesting question, and for me it’s useful to separate it into two parts. In the enterprise services framework, the appliances do a lot of things in that sphere right now and it’s hard for me to think of some brand new functional piece that will arise with a need behind it – though addressing security is likely to see some novel approaches. Once we go through the virtualization and on-demand delivery model, functions that exist are going to carry us a lot of the way. But as you look into the virtualization of hardware more broadly in the network, there I think things get a little different. When you think about virtualizing routers and switches as well as control plane processes, what’s the potential scale of impact of that on the network? I don’t think we know yet. I don’t think we’re headed for a world where networks become nothing but generic servers with software running on and above them. But clearly there is scope for more to be done by software in the network at large. Just how far virtualization and SDN will go is where we will see some real innovation.
TR: What is the next step for Ciena’s Agility portfolio?
CJ: We’ve got a lot of focus on getting through the initial adoption phase. That involves lots of fine tuning on the path to operating deployment. We have a fairly broad suite for SDN and NFV, and it’s a focal point to make sure that is fleshing out and building returns. Of course there are extensions on a portfolio level in the works that I don’t want to be specific about now.
TR: Thank you for talking with Telecom Ramblings!
If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!Categories: Industry Spotlight · SDN · Telecom Equipment
The best example of a service provider who has truly transformed customer service delivery is Pacnet and its software transformation of network functional and defined networking via its PEN portal. My understanding is the underlying team that worked this was an outside firm who helped to dramatically change the capabilities of traditional network service offerings by using software to orchestrate network element and service provisioning. The environment supports multiple vendors and enables customers to procure and deliver services across all of the Pacnet Datacenter footprints as well as network locations.
The Pacnet PEN stuff is definitely slick. But it is ultimately limited to their DC’s and a couple services. I think everyone is waiting for that moment when the same thing interconnects all the majors worldwide at once!
Federating network and service flow through provisioning has been proven to be possible using software but it isn’t in place inter-network providers. Its a question of control. If you allow someone else to allocate and presumably provision resources that you would potentially require for yourself then how do you ensure QoS, CoS and network “call admission control”? That is why Pacnet focused entirely on its own network and service capabilities that it can control.