Project Description:

The purpose of this project is to enable dynamic composition, deployment and optimization of NFV infrastructures and resources, using DMTF Redfish APIs for CDI management, as part of the Multi-Cloud layer in ONAP, and to enable use of CDI-composed resources for VNF instantiation.

Background

Composable Disaggregated Infrastructure (CDI) enables dynamic composition of computer systems (a.k.a compute resources)  from pools of disaggregated compute, storage and network components. The composition process is model driven, based on declarative templates that define what the composed system should look like in terms of its components and hardware capabilities. The disaggregated resource pools are contained within a set of racks, managed by the "infrastructure manager" software. The "infrastructure manager" software exposes set of northbound interfaces for consumption and administration of CDI. These interfaces are based on industry standard DTMF Redfish APIs. The following diagram illustrates basic components of CDI.


Use Cases

The  use cases to be enabled by this project are as follows:

IDDescription
1Dynamic composition and deployment of NFV virtual infrastructure instances in-band, as part of network service instantiation, or out-of-band, as part of NFV infrastructure management. Examples of virtual infrastructures include OpenStack and Kubernetes instances.
2Dynamic composition and rightsizing of virtual infrastructure resources for use during VNF homing and placement.
3On-demand scaling of virtual infrastructure capacity.
4On-demand scaling of virtual infrastructure resources, within a given virtual infrastructure instance.
5Dynamic resource sharing, balancing and reuse between multiple and different virtual infrastructure instances.

Benefits

The benefits of using composable disaggregated NFV infrastructure (NFV-I) are as follows:

  • On-demand optimization of VNF resources, in line with VNFD specified resource requirements.
  • Reduction of orchestration (instantiation/operation/remediation) failures due to lack of required infrastructures and/or resources.
  • On-demand dynamic capacity optimization across multiple virtual infrastructure environments
  • Unobstructed access to infrastructure telemetry
  • Industry standard interfaces models and interfaces
  • TCO (CAPEX and OPEX) reduction as consequence of dynamic resource sharing/optimization, and infrastructure management automation.

Scope:

The following functionality will provided:


Deliverables

Dublin
Release

1

DMTF Redfish® API adapter (management/composition)

Yes

2

Redfish® API based automated resource composition and instantiation

Yes

3

Redfish® API based automated virtual infrastructure (IaaS/PaaS) composition and instantiation.

Yes

4

TOSCA based resource, automated virtual infrastructure (IaaS/PaaS) composition and instantiation

-

5Dynamic virtual infrastructure scaling-
6Dynamic resource scaling-
7CLIs/SDKs for programmatic access-
8HPA support -
9

AAI integration (inventory/capacity information)

Yes

10

DCAE integration (telemetry)

Stretch

11

Orchestration integration (composition/scaling)

-

12

OOF integration (optimization)

-


The initial focus of this project is on composition and optimization of NFV-I resources across OpenStack and K8S environments. Composition and optimization of resources consumed by ONAP components is out of scope.

Architecture Alignment:

The following diagram illustrates how project functionality fits in with the rest of ONAP components. The "CDI Infrastructure Management Adapter" component, shown in the diagram, contains the bulk of the implementation and consists of "glue code" required to interaction between ONAP components' native interfaces, DMTF Redfish APIs and TOSCA templates.


CDI Infrastructure Manager...

CDI Infrastructure Manager


CDI Flows...


CDI Flows


Project Dependencies

The project depends on the following ONAP components:


ComponentDependencyRelease
1AAIInventory and capacity data integrationDublin
2DCAETelemetry integrationDublin
(stretch) 
3SOResource and/or virtual infrastructure environment optimization and scaling during instantiation-
4APPCResource and/or virtual infrastructure environment optimization and scaling during operation and remediation-
5VF-CResource and/or virtual infrastructure environment optimization and scaling during instantiation, operation and remediation-
6OOFResource optimization-


This project aligns with the following standards and information models:

    • DMTF Redfish Information Model
    • DMTF Redfish REST APIs
    • OASIS TOSCA Simple YAML Profile

The project has no dependencies on other open source projects.

Other Information:

  • Link to seed code  - None
  • Vendor Neutral - Yes
  • Meets Board policy (including IPR) - Yes

Use the above information to create a key project facts section on your project page

Key Project Facts:

Facts

Info

PTL (first and last name)Bin Yang
Jira Project NameTBD
Jira KeyTBD
Project IDTBD
Link to Wiki SpaceTBD

Release Components Name:

Note: refer to existing project for details on how to fill out this table

Components Name

Components Repository name

Maven Group ID

Components Description



org.onap.




Resources committed to the Release:

Note 1: No more than 5 committers per project. Balance the committers list and avoid members representing only one company. Ensure there is at least 3 companies supporting your proposal.

Note 2: It is critical to complete all the information requested, that will help to fast forward the onboarding process.

Role

First Name Last Name

Linux Foundation ID

Email Address

Location

PTL



Committers


















Contributors













  • No labels

4 Comments

  1. is it sub-project of MultiCloud?

    does it means design vim resource as CDI and them handle the CDI resource only? How to group the VIM resource in design time, if need to switch from 1 group to another group shall we change the design or how to do it?

    1. Yes, this is a sub-project of the Multi-Cloud project.

  2. Is this pertinent to private clouds only (Do Azure and AWS expose APIs to do this type of thing ) ? If its only private clouds , how does this relate to Akraino ?

    1. Brian Freeman - To my knowledge, none of the public cloud providers support Redfish composition APIs. However, this does not mean that they won't in the future. For now, however, it's for private clouds only...

      I am not participating in Akraino, so I can't comment on applicability of CDI with respect to it.