<< Back

Table of Contents

• Chapter 1 (Overview): Role of NFVi and current state/trends (JB – Intro)
• Introduction

  

  Figure 1-1: E2E

• Reference Architecture traceability to the CNTT
• Reference Model : explain how the two are related
• Principles

• Scope

• Chapter 2 (Architecture Requirements)

• Foundation (includes Management Plane)
  • Bare-Metal provisoning.
  • Enabler services: NAT, DHCP, DNS, IDS/IPS and Load Balancing

• Control Plane

  • (e.g. needs the ability to do…xyz for this
  • control plane provides the API endpoints, GUI and internal services for the cloud.
  • All control plane components must be configured to be highly available (HA) and deployed across multiple physical nodes.
  • API endpoints and supporting services need to handle failures of an underlying component.
  • Control plane much suppport Logging, monitoring and alerting to manage the cloud. Control plane must be Telemetry enabled to provide insight into utilisation and usage reporting

• Data Plane

  • Must allow for tenant networking within a single server and across servers.
  • Must support Service function chaining.
  • Must provide high throuput and low latency.
  • dependong on the targeted profile.
  • Must support Netowrk Acceleration (in-line and Look-Aside)
  • Must support Application specefic Acceleration (exposed to VNFs).

• Networking

  • Must allow for East/West tenant traffic within the cloud (via tunnelled encapsulation overlay - VXLAN or Geneve)
  • Must allow for management traffic of the cloud
  • Must allow folr management traffic of the servers i.e. IPMI
  • Must allow for external access to the API and GUI
  • Must allow for tenant external access from other systems or users e.g. Internet/DC networks
  • Must allow for storage access network and replication
  • Must support VXLAN/Geneve over L3
  • Must support Distributed Virtual Routing (DVR) to allow compute nodes to route traffic efficiently.

• Chapter 3 (Architectural Components)

• High Level Diagram for E2E:.

  

  Figure 1-2: High Level Ref Architecture

• More detailed Diagram (for NFVI and VIM):

  

  Figure 1-3: Focused NFVI VIM Reference Architecture

• Virtual Components (trace to Ref Model)
  • Technology choices to satisfy these requirements
  • Rationale/Explain why, how, of choices
• Physical components (trace to Ref Model)
  • Technology choices to satisfy these requirements
  • Rationale/Explain why, how, of choices
• VIM (Aspects, Impacts)
  • Technology choices to satisfy these requirements
  • Rationale/Explain why, how, of choices

• Components List Diagram:

  

  Figure 1-4: Components List

• Chapter 4 (NFVI Interactions with VIM and MANO): in here we should define our standards (JB – aligns w/Ch 7 of model)

• Chapter 5 (Leveraging NFVI for VNFs):

• Optimize: how to use this specific to this NFVI architecture to optimize VFNs behavior (e.g. performance, scalability, elasticity, resilience, etc.)

• Define how VNFs shall behave

• Chapter 6 (Design, Build, Deployment, and Operate Guidelines)

• Security
• Automation
• Lifecycle Management
• (How points downstream to Reference Implementation)