NFV: I/O (PCIe) Based NUMA Scheduling

Registered by Adrian Hoban on 2014-06-04

The NUMA locality of I/O devices is another important characteristic to consider when configuring a high performance, low latency system for NFV workloads.

This blueprint aims at combining NUMA based PCIe device information with the CPU/NUMA related topology info included in the juno release.

This optimises Openstack guest placement by ensuring that a guest bound to a PCI device is scheduled to run on a NUMA node that is associated with the guests pCPU and memory allocation.

Blueprint information

Status:
Complete
Approver:
John Garbutt
Priority:
Low
Drafter:
Adrian Hoban
Direction:
Approved
Assignee:
James Chapman
Definition:
Approved
Series goal:
Accepted for kilo
Implementation:
Implemented
Milestone target:
milestone icon 2015.1.0
Started by
James Chapman on 2014-12-29
Completed by
James Chapman on 2015-02-05

Related branches

Sprints

Whiteboard

Gerrit topic: https://review.openstack.org/#q,topic:bp/input-output-based-numa-scheduling,n,z

Addressed by: https://review.openstack.org/128344
    I/O (PCIe) based NUMA scheduling

Spec merged, approved blueprint --johnthetubaguy 17th October 2014

Bumping to k-2 -- mikal 9 Dec 2014

Addressed by: https://review.openstack.org/140289
    Add numa_node to PCIDevice

Addressed by: https://review.openstack.org/140290
    PCI NUMA filtering

Addressed by: https://review.openstack.org/152915
    Add PciNumaInfoWeigher
The last patch is optional, previous two patches implement this BP

(?)

Work Items