Project “OSCAR” provides a deployment platform for automatically installing and configuring the software components of OPNFV systems. It is intended to be used by operators who are deploying OPNFV in integration labs and in production networks, and is intended to be part of, and integrated with, the OSS environment. The input to OSCAR will be the set of images and packages (with associated metadata) created by the Octopus CI environment (or other sources) as shown in the diagram below.
A primary goal of OSCAR is to support the installation of a variety of OPNFV software stacks that are based upon on different underlying components, since each operator may have their own preferences based on their own unique business and technical requirements. These variations could, for example, leverage different SDN controllers or approaches to compute infrastructure (e.g. hypervisor vs container).
The OSCAR project is described in detail in this presentation: oscar_overview_v8.pptx
In Release 1 of OSCAR, it is proposed to support two OPNFV stacks, in order to demonstrate operation with multiple stacks. The stacks are described in the following table:
OPNFV Component | Stack1 | Stack2 |
---|---|---|
Virtual Infrastructure Manager (VIM) | OpenStack (Juno) | OpenStack (Juno) |
Network Controller | OpenDaylight | OpenContrail |
Compute OS | Ubuntu, Debian, Fedora | Centos, Ubuntu |
Virtualization | KVM/QEMU | KVM/QEMU, Docker |
Virtual Networking | Open vSwitch | OpenContrail vRouter |
Preloaded VNFs | As per BSG + TBD | As per BSG + TBD |
Installation and Orchestration | TBD | TBD |
OSCAR will validate each installation once it is completed, and reports will be available to show the configuration and status of each installation under OSCAR’s control.
Later releases of OSCAR will implement the OPNFV platform lifecycle for each installation under management, including component upgrades, scaling and site migration. Images and packages from Collaborative Development projects will be included in OPNFV stacks managed by OSCAR as they become released through Octopus. Additionally, it will be desirable to support configuration of network infrastructure to provide connectivity from physical networks to the virtual infrastructure supporting the VNFs.
OSCAR will be based on a set of open source management tools, selected according to best-fit for specific tasks. Available tools include: Cobbler, Puppet, Chef, Ansible and several others. The overall workflow for deployment is likely to be based on TOSCA.
Documentation will be provided to enable users to edit the components of an OPNFV stack or to create new stacks. This will include specification of component images and packages, scaling rules, high availability configurations, and health check metrics and methods.
The primary focus of OSCAR is managing pre-production and production deployments in operator labs and networks. The intent is to develop a system which provides “recipes” or templates that can be used to create OPNFV stacks whose configurations match desired levels of scalability and availability as required by individual operators and for each specific deployment.
Related Projects
OSCAR will provide automated tools for installation of various components as well as health check for individual components. Testing will include:
Release 1
Release 2+
Documentation will be provided that describes the OSCAR architecture, components, supported stacks, installation procedure, user interface, APIs, and configuration files.
The OSCAR project relies on the following open source projects:
Project Creation Date:
Project Category: Integration & Testing
Lifecycle State:
Primary Contact:
Project Lead:
Jira Project Name: OPNFV System Configuration And Reporting
Jira Project Prefix: OSCAR
Mailing list tag [oscar]