Nova LoadBalancer Overview

Registered by Alexander Chadin on 2015-10-27

OpenStack Load Balancer is an automated system of workload distribution among physical servers in a cluster. It is designed to optimize resource usage, reduce response time and allow horizontal scaling of a cluster (the ability to add or remove physical servers in case of shortage or excess of resources at any time after that the load balancing system automatically distributes workload in the cluster).

Blueprint information

Status:
Not started
Approver:
Alexander Chadin
Priority:
Medium
Drafter:
Alexander Chadin
Direction:
Needs approval
Assignee:
None
Definition:
Drafting
Series goal:
None
Implementation:
Unknown
Milestone target:
None

Related branches

Sprints

Whiteboard

These are the results that the system helps to achieve:
• Reducing the possibility of servers overloads
• Even resource usage
• Flexibility
• Energy costs reduction.
Load Balancer allows working with the following set of algorithms:
• Threshold algorithm (the decision to migrate is based on threshold crossing)
• Standard deviation algorithm (the decision to migrate is based on hosts overload ratio)
• minRAM maxCPU algorithm (instances are migrated to hosts with the most loaded CPU and the least loaded RAM)
• Standart deviation minimization algorithm (The couple for migration is chosen based on maximum minimization of standart deviation after simulating their migration)
• Underload algorithm (finding the average host load and comparing it to the threshold value)

(?)

Work Items

This blueprint contains Public information 
Everyone can see this information.

Subscribers

No subscribers.