October has been the busiest month for software defined networking (SDN) events, where literally every week has been crammed with multiple conferences, industry, and standards meetings.
In October alone, I had the honor of chairing the SDN Workshop at the Telecommunications Industry Association’s TIA2013 event (Oct. 7-8, Washington D.C.), the SDN/ONF Workshop at the SDN & OpenFlow World Congress (Oct. 15, Bad Homburg, Germany) and the SDN & OpenFlow Workshop at the Broadband World Forum (Oct. 22, Amsterdam).
Speaker after speaker during all three workshops extolled the benefits of virtualization and revealed common expectations for improved agility and faster time to new services, while reducing their opex. What is different is that as recently as six short months ago, at least for some major operators, network functions virtualization (NFV) was considered the present, while SDN was perceived as the future. Now, SDN was cast as an NFV enabler, and some consider the two inextricably linked.
The Past and Present State of NFV
My first blog post for this series provided an introduction to NFV and the European Telecommunications Standards Institute (ETSI) Industry Specification Group (ISG). The ISG’s charter is to analyze the requirements, leveraging the collective wisdom of the diverse network operators and vendor community alike.
Part 2, on the business case for SDN, was motivated by at least some major operators indicating that while NFV was front and center in their plans, SDN was perceived as a future initiative. Such a proclamation is not unfounded, as NFV offers the potential for significant benefits on both the revenue and cost size of the ledger.
At the 2013 SDN & OpenFlow World Congress, one year after 13 operators released the NFV white paper that catalyzed the ETSI NFV ISG, a revision was introduced. Contributors to the ‘White Paper 2.0’ doubled and included the world’s largest fixed and mobile operators. What is particularly significant is that both white papers are not formal standards deliverables, but represent a collaborative and informal effort to clarify the objectives and status of NFV.
Neither the white paper, nor the set of end-to-end documents recently approved, elaborated on the relationship between NFV and SDN. The White Paper 2.0 authors did remove the controversial Venn diagram that implied a very modest intersection between NFV and SDN, another indicator that the thinking that has clearly progressed.
NFV Myths
Even though NFV is increasingly sharing the stage with SDN, confusion abounds. In my admittedly limited survey of three SDN Workshops, I observed the following prevalent myths.
The ETSI NFV ISG is a standards body. The White Paper 2.0 explicitly clarified the mission of the ISG:
Although ETSI is a Standards Development Organisation (SDO), the objective of the NFV ISG is not to produce standards. The key objectives are to achieve industry consensus on business and technical requirements for NFV, and to agree common approaches to meeting these requirements. The outputs are openly published and shared with relevant standards bodies, industry fora and consortia to encourage a wider collaborative effort. The NFV ISG will collaborate with other SDOs if any standardisation is necessary to meet the requirements.
NFV equates to “The Cloud.” While NFV is expected to realize the many benefits of public and private cloud services, there are differences:
- Generic cloud services tend to be compute-centric, whereas NFV is inherently connect-centric.
- Generic cloud services are optimized to enable multiple tenants to share compute, storage, and connect resources in a highly cost- and energy-efficient manner. NFV must scale network functions to serve millions and even tens of millions of subscribers. In fact, many large-scale enterprise applications are not virtualized to avoid the overhead.
- Generic cloud services are typically characterized by many relatively small VMs to optimize the utilization of the system resources; NFV deployments for large operators will consist of fewer but much larger VMs to accommodate the vast scale of large operators.
- Generic cloud services seek to decouple the virtual and physical domains. Carrier networks demonstrate what some operators refer to as “shape,” which consists of definitive segments and service boundaries with orderly handoffs.
NFV is about capex. While the ISG refers to high-volume, standard servers, switches and storage, during the recent SDN workshops (and others before) there was a strong consensus that the expected value proposition is to improve agility and operational efficiency. Capex reduction is sought after, but as a secondary goal.
NFV winds down in January 2015. Although the formal charter of the ETSI NFV ISG expires then, the efforts initiated by the NFV ISG will most certainly resume, albeit in an existing standards body or industry group.
We are on the verge of adding another: NFV is not dependent upon SDN. While workshop participants paraded to the podium one after another to acknowledge that SDN is a key enabler of NFV, few asserted that SDN is ready for prime time. In fact one analysts referred to SDN as in its infancy, which stimulated debate about how far SDN has progressed.
Up Next: ETSI NFV Meeting No. 4
On the verge of the fourth meeting for NFV (Oct. 30-Nov. 1 in Sunnyvale, Calif., hosted by Juniper Networks), armed with a solid and aligned functional baseline, attention will shift to the working groups. There are over 300 individuals registered to attend, which is attributable partly to the venue being Silicon Valley, where many technology firms are located, and partly to the principal factor: NFV remains hot.
SDN will play an increasingly important role in NFV; look to each of the NFV ISG working groups and expert groups, with the Infrastructure (INF) and Management and Automation (MANO) teams taking the lead. Those are the two working groups that will address the network, while others are more focused on the compute domain.
Now that the NFV ISG has finalized the proof-of-concept (PoC) framework document, we expect to see additional PoCs in 2014, which in turn will help refine both NFV and SDN. Standards enhancements and open-source software will not be far behind.
All indicators reveal a definite shift underway. Last year we were asking, “What is SDN?” and this year, two questions emerged: “Why?” and “When?” Next year, I predict the beginning of another shift, to “How.”