Operators and vendors have so far mainly focused on the capabilities and benefits of software-defined networking (SDN) and network functions virtualization (NFV) with little thought about how they will be "operationalized."
Yet a large part of the reason SDN and NFV are disruptive is that they challenge the status quo where network management processes and systems are concerned. As my recently published report, "Managing the Virtualized Network: How SDN and NFV Will Change OSS" points out, SDN changes the demarcation between OSS and the network by providing an alternative, single, centralized, real-time touchpoint to it, while NFV requires the implementation of a completely new cloud management stack. (See SDN & NFV to Shake Up Operator OSS Market, Heavy Reading Finds.)
Both technologies will potentially sweep away the dozens, or even hundreds of vendor-specific, technology-specific and/or service-specific network management tools that service providers use to configure, monitor, and troubleshoot the multiplicity of physical devices at different layers of their networks.
Getting rid of these tools and systems is one thing -- it's an objective that telcos have had in mind for years. But as usual with new and disruptive technologies, the sticking point is the organization -- or multiple organizational silos in the case of large operators -- that have grown up around these management tools. Telcos have enormous investments in people and processes that are very hard to unravel. Many incumbent operators in Europe, for example, have ageing workforces that are difficult and expensive to renew, retrain, or simply shed. Since SDN and NFV promise high levels of management automation, the brutal (and largely unspoken) truth is that large numbers of operational staff may no longer be needed if and when they are deployed.
So given the organizational disruption involved, will operators bring SDN and NFV into their legacy core networks? At the recent Global SDP Summit in Rome, Antonio Manzalini, of Telecom Italia (TIM) 's Future Centers/Innovative Architectures group, made the distinction between the "in track" (legacy network) and new, "disruptive" network application of these technologies. His conclusion was that it will be too costly and time-consuming to bring SDN and NFV into the legacy network, which should be managed as-is until end of life.
But SDN and NFV will be key to new network architectures of the future, such as networks supporting user clouds at the extreme edge. Operators are starting to work on such networks so that they're ready to support new business/services opportunities with appropriate operational structures and economics as legacy networks die out. One of the reasons SDN and NFV are resonating so well with telco cloud providers is that cloud datacenters represent another more-or-less "greenfield" area that lends itself to the adoption of new technologies, especially as SDN and NFV have such large synergies with compute cloud.
The assumption so far is that SDN and NFV will transform existing networks: But how possible and desirable is this from an operator point of view? Or should SDN and NFV be used to underpin a next-generation network architecture that will take over from legacy at some future point? At Light Reading's recent Ethernet and SDN Expo these issues were high on the agenda, with the thorny issue of operationalizing SDN and NFV, and whether and how their management approaches can co-exist with today's OSS, a real focus of debate, as they have been again this week at the SDN & OpenFlow World Congress.
And this is only the beginning: Expect to hear a lot more about these issues in the coming years.
— Caroline Chappell, Senior Analyst, Heavy Reading