This is a guest blog by Caroline Chappell, Principal Analyst, Cloud and NFV at Heavy Reading.
Over the past year, the topic of network functions virtualization (NFV) management and orchestration (MANO) has moved from the margins of NFV conversations to center stage, and from theoretical speculation about what orchestration is and does to detailed discussions about the functionality required within the NFV MANO stack.
This reflects the way in which the entire telecom industry shifted in 2014, with network operators collectively confirming that NFV is the direction most of them intend to take. A number of operators are now making aggressive promises to roll virtual network functions (VNFs) into production networks in a relatively short time.
A key driver for the rapid introduction of NFV is the need to control operational costs. While not a complete answer to telcos’ operational cost woes, NFV should enable the programmatic orchestration of a software-based network and therefore high levels of management automation. As we pointed out in our report 2020 Vision: The Heavy Reading Perspective, first movers to an NFV management model expect initial investment to be high because of the organizational changes involved, but opex to be significantly lower once it is in place.
Aggressive timescales are concentrating minds on how to operationalize NFV — how VNFs will interwork with a discrete pool of IT cloud-based infrastructure resources, and how VNFs will themselves be managed in this dynamic, new environment.
NFV MANO functionality — which hasn’t been needed to operate networks before and is therefore additional to existing OSS — fundamentally needs to be defined and understood before operators can introduce it into their operations organizations.
MANO concepts may be rooted in IT cloud management systems, but that’s not the whole story, as Chris Bilton, Director of Research and Technology at BT, pointed out at Light Reading’s November 2014 event, OSS in the Era of SDN & NFV: Evolution vs. Revolution.
IT cloud management approaches need to be infused with network management understanding to result in what Bilton called the “third way” of managing next-generation, virtualized networks. The industry is still working through what that “third way” looks like and its consequences, not only for VNF and NFV Infrastructure management, but for the traditional OSS stack, too.
In Heavy Reading’s opinion, there are three milestones for the delivery of the “third way” virtual network management that need to be reached in 2015:
1. Make NFV real. If NFV is to progress, proofs of concepts (PoC) and trials need to come out of the lab and into live networks. Managing and orchestrating a single VNF in a siloed PoC is a different matter from managing it as part of a production network. Real experience with NFV MANO will be gained when it is tested in a live environment, and operators have to face and overcome OSS integration and cultural, organizational and process change issues. New services such as VoLTE are likely to be the earliest sources of NFV implementations. Nokia points out that operators don’t have to roll out VoLTE as a fully virtualized service from day one. They can start with a relatively simple VNF, such as the telecom application server, to iron out MANO/OSS issues, and then follow with virtualized IMS CSCF, HSS, SBC and EPC functions as the “third way” of network management beds down.
2. Define and stabilize the NFV infrastructure management domain of the NFV MANO. The virtualized infrastructure management (VIM) and inter-domain resource management (NFV Orchestrator) functionality required for NFV are the least controversial and most fundamental part of the MANO architecture. They are also the best understood, as discussed at the OSS in the Era of SDN & NFV: Evolution vs. Revolution event. There is still work to do in extending operators’ favored VIM candidate, OpenStack, with carrier-grade and NFV-specific management features, but a reference implementation from the OPNFV initiative will go a long way toward providing the industry with the stable, standardized cloud management platform it needs. OPNFV continues to attract new supporters to join founding members, including AT&T, HP, Intel, Nokia and RedHat.
3. Address NFV security in conjunction with the NFV MANO. In a virtualized, software-driven world, security cannot afford to be an afterthought because of the raised potential for cyber-attack — especially where an entity as critical as the network is concerned. Yet security has largely been absent from MANO conversations, partly because in the physical network, security operations belong to a separate organizational silo. NFV represents a large opportunity to break down the artificial barrier between security and other FCAPS management operations, and to introduce stronger and verifiable levels of governance through policy-driven automation. Heavy Reading expects that NFV security issues will rise up the agenda in 2015 as operators start to design and implement the end-to-end management processes that span OSS and the NFV MANO stack.
This blog is also featured here.