cfn/ubiquitous-computing-scheduling
Code Issues Proposed changes
e94826ce29
ubiquitous-computing-schedu.../UCS-Introduction.md
liuhaijun 8f126a978f note 
Change-Id: I179d06bba8754722e9021af65443c28acc5ff681
2023-08-10 02:00:36 +00:00

5.1 KiB
Raw Blame History

Ubiquitous Computing Scheduling Sub-group Introduction

Ubiquitous Computing Scheduling is to build a unified scheduling platform for computing resources including center cloud, edge, and terminal computing..

Project Facts

Project Creation Date: 2022/11/14

Primary Contact: *Xiaoqiao Geng & gengxiaoqiao@inspur.com *

Project Lead: Andre Hou & hougx@inspur.com

Committers:

Mailing List: ubiquitous-computing-scheduling@lists.opendev.org

Meetings: No sub-group meeting time. Use bi-weekly meeting of CFN WG.

Repository: https://opendev.org/cfn/ubiquitous-computing-scheduling.git

StoryBoard: https://storyboard.openstack.org/#!/project/cfn/ubiquitous-computing-scheduling

Open Bugs: N/A

Introduction

With cloud native technology as the basic technology stack, unified management and perception of central cloud, edge cloud, non-cloud computing power and other types of computing power are realized, as well as cross-domain task scheduling, collaborative scheduling and real-time dynamic adjustment of applications are realized, and the basic platform of universal computing scheduling is built. Main Functions:

  1. Strategy management and intelligent decision-making • Receive input of candidate computing node information for scheduling resource requirements or meeting resource requirements. • Output preferred network resource node information based on intelligent algorithms and historical data. • Responsible for the full life cycle management of scheduling policies.
  2. Computing power perception • Receive queries on the type, capacity and performance of resources and resource combinations required by choreography and scheduling, and feedback information of computing nodes that meet the requirements of resources and resource combinations. • Realize the perception and recording of metadata information such as the quantity, distribution and occupation of resource object information such as computing power, network and storage. • Realize real-time awareness of the status information of computing, network, storage, application, service and other objects, and provide awareness for intelligent decision making, computing scheduling, dynamic capacity expansion and shrinkage, and disaster recovery.
  3. Computational force scheduling • Based on the request of choreography parameters, component capabilities such as computing power enablement, network enablement, application enablement, data enablement and service enablement are invoked to complete collaborative scheduling at the edge of cloud network. • Based on application availability requirements, combine computing power perception module. and intelligent decision module to complete disaster recovery and automatic capacity expansion and shrinkage of computing power.
  4. API Gateway: • As a northbound interface, it provides a standard external interface for communication with modules such as network choreography.

Documentation & Training

N/A

Release Planning & Release Notes

1.Computing network perception: Receive orchestration and scheduling queries on the type, capacity, performance and other requirements of resources and resource combinations, and feedback the computing power node information that meets the resources and resource combinations;

  • a).Realize the perception and recording of metadata information such as computing power, network, storage and other resource object information such as quantity, distribution, and occupancy;
  • b).Realize real-time perception of object state information such as computing power, network, storage, application, and service, and provide perception capabilities for intelligent decision-making, computing power scheduling, dynamic expansion and contraction, and disaster recovery;

2.Policy management and intelligent decision-making:

  • a). Receive the information input of scheduling resource requirements or candidate computing power nodes that meet resource requirements;
  • b). Based on intelligent algorithms and historical data, output optimal computing network resource node information;
  • c). Responsible for the full life cycle management of the scheduling strategy;

3.Computing power scheduling:

  • a). Based on the orchestration parameter request, call the computing power enablement, network enablement, application enablement, data enablement, service enablement and other component capabilities to complete the collaborative scheduling of the cloud network edge;
  • b). Based on the availability requirements of the application, the joint computing network perception module and intelligent decision-making module complete the disaster recovery and automatic expansion and contraction of computing power;

Previous Releases

N/A