OPNFV Summit 2015

This presentation will outline key community activities and technologies under development in Brahmaputra that provide value to the OPNFV project and community. The presentation will highlight key trends and activities in the community as they emerge as OPNFV transitions from a constrained Arno development cycle to an inclusive Brahmaputra release project. The presentation will cover interactions and touch points with other communities and the importance of these relationships.

The presentation will provide a glimpse of where we are as a community, what to expect of the Brahmaputra release and where the community is heading for future releases.

Most VNF use cases are designed and presented around VMs, but what about Docker containers? Many VNFs can run as applications in user space, as though run in Docker containers. This provides resource reduction, cost reduction, easier manageability, along with the ability to place or co-locate VNFs of the same tenant on one or several VM instances. This presentation will bring a real world example of an IMS system, based on the open source Clearwater from MetaSwitch, which will demonstrate a way to deploy, monitor, auto-heal and auto-scale Clearwater on Docker containers.

OPNFV (Open Platform for NFV) has defined a common infrastructure for Virtualized Network Functions (VNFs) which, as this talk will describe, can be successfully used for a wide range of demanding enterprise applications as well, primarily in highly regulated and strong SLA focused segments. In this talk we will describe the vision and trigger the discussion on how OPNFV and other open source initiatives can jointly benefit all workloads, telco and beyond.

While little discussed, there is a solid case in placing some VNFs at the edge of the network, but this notion implies a distributed model that VNFs were not originally created to address. One of the problems in placing VNFs at the network edge is that the edge equipment may be very cost sensitive and in many cases, consumer grade.

These consumer-grade devices cannot support a virtual container such as a KVM because of the huge memory requirements of an additional OS. A Docker container has been proposed and implemented with great success.

Testing and performance characterization are key capabilities being developed by the OPNFV community towards deployable NFV solutions. In this presentation, representatives from Orange, Ericsson and Intel will review test strategy for the OPNFV B-release describing scope and work flow of test projects. Experiences from the initial release as well as existing infrastructure and aspects common to all test projects will be discussed, including tooling and relation to continuous integration. In addition, details of how the overall OPNFV test framework shall be enhanced and extended, will also be covered. The presentation will also explain how new OPNFV developers can approach testing new features and interact with other test projects.

While the construct of Network Functions Virtualization and Software Defined Networking has graduated from a conceptual to a viable technology phase, real world deployments are still proving to be a challenge. Telecom Equipment Manufacture’s (TEM’s) have to find a reliable, deterministic manner to benchmark their wares to specific virtual network functions. Benchmarks which are interoperable, repeatable, based on open standards supporting plug & play modular architectures, to accommodate multi-vendor VNF support will be needed.
Intel has a deep experience in stemming from the successful adoption of virtualization in the data centers; Learnings from the virtualization of the data center journey are directly applicable to standardizing and accelerating the benchmarking for NFV, all with the aim to vastly accelerate the adoption curve of virtualization in networking.

In this panel discussion, we explore a critical aspect of the OPNFV ecosystem: Interoperability. Is it just a dream, or is it something the community can build it around as a differentiating feature of this ecosystem? What are the goals of interoperability, how to balance these goals and what roles different organizations and OPNFV in particular may play? Panelists from open source communities, standard organizations, service operators, software developers, and conformance and certification specialists come together to discuss how to best achieve interoperability.

The objective of the OPNFV Pharos project is to develop a federated test-bed capability across geo's to be hosted by OPNFV community members. In this presentation, representatives from Intel, Orange and Ericsson will review progress, explain the vision and give an update on plans for B-release. Topics will include a description of existing community and Linux Foundation infrastructures as well as governance and policy, lab specifications and tooling. OPNFV certification and compliance will also be discussed.

Testing is a key challenge for OPNFV. Many projects are dealing with test and generating results on different testbeds relying on different hardware. All these differences lead to a need of a management framework. The goal is to collect, display, analyze and compare heterogeneous sources and provide ultimately a unified user experience.

This presentation will detail this architecture within OPNFV continuous integration environment. An early version of the test dashboard will be shared with the community.