近日，Gartner Research在网络，数据中心的资深研究人员Andrew Lerner发表了一篇文章关于Gartner是否会对SDN和NFV单独做市场象限图。其观点是：Gartner不会。原因是：SDN和NFV不是一个单独的市场YET。
弯曲科技并不完全认同Gartner的观点。但倾向于同意其在SDN方面的洞察。在NFV方面，由于主要是运营商在博弈，同时许多Service Chain都还在POC阶段，NFV的演化目前来看还略微早。但网络中的服务节点（Service Node）会全面受到冲击是没有任何怀疑的，例如，ADC系统，防火墙系统，UTM系统，IDS系统，to name a few。
弯曲科技认为，SDN对于云服务而言，其最大的特征是为网络虚拟化提供动态配置功能。例如，2层和3层网络，安全策略。从这个角度出发，可以基于SDN的方法论，作出一些工具产品来提供给CSP，云服务提供商。SDN对于运营商而言,最大的特征是策略迅速收敛,例如,路由策略,安全策略等等.
下图是弯曲科技对SDN,NFV和Clouds之间关系的理解概念图:
从上图可见, SDN和NFV本身作为一个市场还有研讨和商榷的空间. 当然, 作为对网络,数据中心,运营商的各种接入网的改造,SDN和NFV将扮演重要的角色. 这个价值是没有问题的.
下面是Andrew Lerner的文章原文:
Is SDN a Market?
by Andrew Lerner | July 20, 2015 | Submit a Comment
This is a Guest Blog from Joe Skorupa
A topic that comes up periodically from clients surrounds defining SDN (and NFV) as a market. The conversation usually goes something like this:
“You guys wrote some of the first research on SDN and NFV and publish forecasts for everything IT. So, when will you publish a forecast on SDN and NFV?”
And another variation of this is
“When will you publish an SDN Magic Quadrant?”
The simple answer is: We won’t, because SDN and NFV aren’t markets. They are an architectural approach and a deployment option, respectively. On a related note: We’ve had similar questions regarding Big Data, which is described in Andrew White’s blog entry: The Elusive Magic Quadrant for Big Data – is Big Data a Market?
SDN is an architectural approach
We do realize that other firms have published reports on the 186 quadrillion bitcoin market for SDN and NFV. However, market forecasts predict the revenues generated by the sale of goods and services and SDN is an architectural approach. The capabilities to build a network based upon that architecture (such as OpenFlow) can appear in Ethernet switches, but that doesn’t make them part of the SDN market any more than VLAN tagging makes those switches part of the VLAN market. The market segment is Ethernet switches and they have features that may be included at no extra charge, or that may carry a premium, at least for a while. So, there is no SDN market, just products that can be used to deliver that architecture, and many of those products also support legacy network architectures.
NFV is a deployment option
NFV is a fancy TLA for a disaggregated network service (software separate from hardware) such as an application delivery controller, session border controller or firewall. When acquiring these services your only option used to be an integrated appliance – software bundled with hardware. Today, virtualization, containers, and performance improvements in x86 processors make other deployment models possible and even desirable. So, for example, we forecast and track the total market and may break out integrated appliances, perpetual license software purchase, yearly lease and monthly lease of software, be it bare metal, VM or container-based. At some point we might even track support-only revenues for open source offerings. But whatever the product category, NFV is simply a set of deployment options. NFV may have an impact on the market size because the hardware revenue often goes to a server vendor and is tracked as part of the server market and only the software revenues will appear in the network equipment market forecast and market share.
So, in the end, the adoption of SDN as an architecture may or may not increase revenues in multiple product categories just as the adoption of NFV may result in a decrease in the size of the ADC or SBC market. We will do our best to explain what is happening in each of the markets rather than publish eye-catching numbers that distort the picture of what is going on in the market.
转载自：弯曲评论
