It took some time for the industry to realize that to operationalize NFV, the Management and Orchestration (MANO) stack is not enough. But great strides have been made over the past year in understanding the role of the MANO, and where it needs to be augmented with OSS functionality.
The third ETSI NFV white paper, published in October 2014, talked about the relationship between NFV and OSS for the first time, thanks to a handful of dedicated network operations experts who had been pushing the initiative on this point. And there has since been further clarification around, first, the role of the MANO in orchestrating NFV infrastructure, and deploying and monitoring individual virtual network functions (VNFs); and second, the need for OSS to manage the customer-facing service context in which those VNFs sit -- alongside physical network functions (PNFs) for the foreseeable future.
Nevertheless, as the white paper made clear, NFV will have a major impact on operational processes and systems and is a catalyst for their transformation. The paper urges operators to identify the parts of their OSS that need to be redesigned to take into account the highly dynamic changes in network topology and connectivity associated with NFV. Keeping track of these changes -- and their impact on service composition and performance -- is a key challenge for future OSS.
Operators that are already making significant commercial advances with NFV, such as Pacnet, which is now owned by Telstra, tell Heavy Reading that NFV’s introduction has gone hand-in-hand with OSS transformation as a result of serendipitous timing. However, few operators -- even as they acknowledge that their OSSs need to evolve to accommodate NFV -- are in such a fortunate position. Yet they don’t want to hold up NFV adoption until such time as they are ready to overhaul their OSS.
Heavy Reading research shows that the more advanced an operator is in implementing NFV, the more confident it is that it doesn’t need to wait until the "perfect" OSS solution comes along. However, such operators also admit that there are plenty of unresolved operational questions to keep them awake at night.
The keys to driving NFV into the network before these questions are completely answered (which may take several years) appear to be a willingness to:
- Bring NFV R&D/engineering and IT/network operations teams together in a meaningful way to start understanding where operational detail is going to live, at which layer of the MANO/OSS stack it will be managed and what the operational responsibilities of each layer will be. There will inevitably be architectural compromises that need to be made when implementing NFV at such an early stage of the market, but advanced operators have an idea of the future combined MANO/OSS architecture and the abstraction and automation principles they are striving for.
- Use a temporary OSS solution and manual processes to get started. Few operators today are ready to flex the full might of the MANO and scale VNFs in a highly dynamic way. So although automated elasticity is an end goal of NFV, it’s perfectly possible -- as operators are proving -- to begin with a simpler subset of MANO functions that sit more easily with existing OSS and manual workflows, and to achieve a basic working relationship between the two. Some operators are drafting in new OSSs to gain experience of how a future, combined solution might work. This allows them to build up the knowledge they need to put together an informed RFP for a more permanent and transformational OSS/MANO integration later on.
Operators starting to plan for the operational transformation needed for NFV are in a strong position to influence the APIs needed to hook OSS and the MANO together, the scope of which is still under discussion. The relationship between OSS and MANO is an evolving area that badly needs input from the bleeding edge to flourish. If it remains theoretical and on paper, NFV will be a long time coming.
— Caroline Chappell, Principal Analyst, Cloud & NFV, Heavy Reading
This blog is sponsored by Alcatel-Lucent.
This is the third in a three-part series on "Accelerating NFV Implementation." For parts one and two, see: