CORD : Satifying-Cactus Roadmap

An overview of the Roadmap was also presented at CORD Build (November 7-9, 2017). See the attached presentation.

 

Priorities

CORD 4.0 stabilized service developer interfaces. The next goal is to build out CORD’s service portfolio:

  • Upgrade all current R/E/M-CORD services to the 4.0 API
  • Integrate latest access peripherals – VOLTHA and xRAN
  • Expand service portfolio to include micro-services – Kubernetes
  • On-Board other VNFs into CORD – ONAP
  • Streamline service on-boarding based on experience

CORD 4.0 refactored build system to improve developer workflow. The next goal is to exploit this flexibility to improve operator workflow:

  • Automate the build-and-install process for physical PODs, including discovery and configuration of the POD switching fabric
  • Make it easy to specify (and change) service profiles independent from configuring the underlying platform
  • Improve lifecycle management capabilities to include in-service-software-upgrade of the CORD control plane
  • Demonstrate how CORD can leverage available infrastructure rather than require that a POD be build on top of bare metal

All this work is driven by the following deliverables:

  • Multi-Access Edge Cloud
    • R/E/M-CORD Services running on the same platform
    • Includes VOLTHA and xRAN access peripherals
  • Managed XGS-PON Peripheral
    • Includes VOLTHA
    • Light-and-Right R-CORD (Kubernetes-based)
    • Includes OSAM (Open Source Access Manager)
  • Integrate CORD and ONAP
    • Use Case 1: OSAM
    • Use Case 2: E-CORD / MSO
    • Use Case 3: A-CORD / DCAE

Platform Roadmap

 Build System

  • Improved Development Workflow
    • Easier debugging -- ElasticStack & structured logging 
    • Easier to develop use cases à Decouple profiles from platform
  •  Improved Operator Workflow
    •  Fast, foolproof install process
      •  Install all containers from Docker Hub onto K8S
      • Install a generic CORD system, use dynamic service onboarding 
    •  Modular, flexible CORD
      • Use existing infrastructure (e.g., OpenStack, provisioned nodes)
      • Exchange pieces of the system (e.g., K8S for OpenStack)
    •  Discover and configure the fabric
      • PoC script to bootstrap the fabric at install time (QA) 
      • Fabric configuration based on XOS models
    •  Support DB migration

Container Orchestration

  • Deploy CORD platform components using k8s
    • OpenStack
    • XOS
    • ONOS/ONOS Apps
  • Support Container-based VNFs using k8s
    • Common overlay network between OpenStack VMs and Docker Containers
    • Hybrid VM-Container service platform
  •  Demonstrate Light-and-Right CORD configuration
    •  Run a configuration with k8s but no OpenStack

XOS

  • Improved Support for Developers
    •  Static Analysis for services and manifests
    •  Simple Synchronizer template, with most code auto-generated
    •  Unit test framework for new Model Policies & Sync Steps
    •  Dynamic service on-boarding
  •  Improved Synchronizer and API performance
    •  Auto-generated test coverage to include end-to-end tests
    •  Improved Support  for Deployment Engineers
    •  Better contextual tracing and debugging support
    •  Dynamic service on-boarding
    •  Enforce Interface models between Service Instances
    •  Cleaner visualization of the service graph

QA

  • Expand test coverage to include M-CORD and E-CORD
  •  Extend automated tests to more fully exercise the platform
    •  Functional regression tests – Black box tests to make sure base components have not regressed
    •  End to end CI/CD tests – To make sure a system can be built from scratch, deployed, and can pass a baseline of tests for both control and for traffic.
    •  Performance tests – So can track performance over time
    •  Build out performance automation framework
    • Populate framework with a few baseline performance tests