Network functions virtualization (NFV) promises to radically change the economics and velocity of network services by deploying virtual network functions on standard, low cost, commodity off-the-shelf (COTS) servers and switches. The benefits can only be realized, however, if network functions can be seamlessly deployed and managed on distributed cloud-based infrastructure that consists of many thousands of cooperating virtual machines. TodayÍs data center architecture was built for Web and cloud native applications that are primarily stateless and feature application-level resiliency. Network functions such as subscriber access/management and networking policy databases are fundamentally different. Network functions are stateful _ they act as a single highly available managed service, while scaling in one or more dimensions: processing power, packet I/O, persistent storage, managed-visibility, and APIs. Plus, unlike the majority of web apps, network functions are real-time _ think voice and video _ and require predictable bandwidth (no jitter) and low latency. Simply porting a network function to virtualized data center architecture will not deliver meaningful change or cost savings. And trading in one expensive complexity for another simply shifts dollars to a new company or a new business unit of an existing vendor. Instead, these virtual network functions need to provide the same degree of simplicity, efficiency, and hyperscale that web applications have today. And NFV will only be successful if the industry embraces the same community-driven open development model that has transformed web based enterprise services. The RIFT.io founders believe that the best way to accomplish this is to bring the same open source, community-driven model that fuels the growth of Web-based businesses, and combine it with the scalability, security, and operational simplicity demanded by cloud-scale network services on data center architecture.