NFV represents the most significant change to global communications networks in the last 30 years. At this critical juncture, it’s essential to adopt an NFV strategy that maximizes flexibility, network agility, and interoperability, while harnessing the rapid innovation in progress across the NFV ecosystem. When considering an investment in a particular commercial or in-house built NFV platform, it’s important to think through the following considerations.
CONSIDERATIONS FOR OPEN NFV
Open is more than a buzzword. In the early days of a new market, custom, proprietary solutions can come quickly to market to meet acute short-term needs. Utilizing these solutions are often sensible to jump start evolution, but in the long term, it can lead to re-creating vendor lock-in and overly customized solutions.
Open interfaces enable interoperability between solutions, in-house or vendor built, so that end users are able to leverage best of breed options in the ecosystem, avoid vendor lock-in, and future-proof their networks.
The shared technology investment that is at the heart of open source provides core set of capabilities that will be built on by an entire ecosystem. Over time, this investment will only increase, surpassing narrow developments within the silo of a narrow corporate boundary. It provides anyone a significant acceleration of any developer hoping to impact NFV.
OPNFV is an open source, community-built platform to accelerate NFV products and services. By bringing together a whole community across vendors, users, and geographies, it brings resilience to the ecosystem and a development scale than cannot be replicated within the boundary of one entity.
NFV represents an unprecedented change in networking. An NFV Infrastructure platform will typically consist of bare metal hardware, VMs and containers, a cloud orchestration system such as OpenStack to manage those resources, Software-Defined Networking (SDN) controllers to manage the control plane of the network, forwarding plane solutions, data plane acceleration stacks to enable high packet processing throughput, and management and orchestration functions. Ensuring that the necessary pieces from diverse groups work together is a non-trivial exercise.
As an integration project, OPNFV focuses the common configuration and common deployment aspects across the entire end-to-end platform. OPNFV scenarios are pre-integrated and tested, giving anyone looking to commercialize an NFV solution an important head-start over starting from scratch. We are also open to anyone who is willing to integrate their project into the platform.
Solutions need to take account of traditional metrics of network robustness: scalability, throughput, fault tolerance, and security.
NFV offers enhanced options in this area. As a cloud-based architecture, NFV supports scaling various pieces of the software stack independently of others, dynamically scaling up and scaling down resources based on application demand or consumer usage patterns, and leveraging easily available COTS of open source hardware to simplify supply chains.
OPNFV works to ensure that all components developed in the ecosystem meet network services needs, whether related to fault tolerance, high availability, throughput performance, multi-site capabilities, or upgradeability.
The promise of NFV is that moving to software-based networks will enable end users to take a far more agile approach to offering network-based services. As opposed to traditional services, some of which would require an entire forklift network upgrade to deliver, NFV enables operators to “fail fast, fail often” when bringing new services to market.
DevOps Continuous Integration and Continuous Deployment methodologies are key to this transition and the backbone of OPNFV. On a nightly basis we build 30+ scenarios and deploy them in an automated fashion to Pharos labs across the entire globe on multiple hardware platforms. In the context of a particular release, we create and tear down thousands of OpenStack data centers integrated with multiple NFV components and perform thousands of lines of automated test code against them. This level of built-in testing and automation enables a new paradigm of network provisioning, speed, and technical diversity.