Merge pull request #7805 from johnbelamaric/wg-device-mgmt

Add WG Device Management
2024-04-29 09:27:54 -07:00 · 2024-04-29 09:27:54 -07:00 · 6046718b31
parent 2f8415788e ee3f23dda0
commit 6046718b31
12 changed files with 200 additions and 1 deletions
--- a/4
+++ b/4
@ -130,6 +130,10 @@ aliases:
  wg-data-protection-leads:
    - xing-yang
    - yuxiangqian
+  wg-device-management-leads:
+    - johnbelamaric
+    - klueska
+    - pohly
  wg-lts-leads:
    - jeremyrickard
    - liggitt
--- a/communication/slack-config/channels.yaml
+++ b/communication/slack-config/channels.yaml
@ -111,7 +111,6 @@ channels:
  - name: digitalocean-k8s
  - name: distroless
  - name: diversity
-  - name: dra
  - name: draft-dev
  - name: draft-users
  - name: druid-operator
@ -549,6 +548,8 @@ channels:
  - name: wg-component-standard-mentorship
    archived: true
  - name: wg-data-protection
+  - name: wg-device-management
+    id: C0409NGC1TK
  - name: wg-iot-edge
  - name: sig-k8s-infra
    id: CCK68P2Q2
--- a/liaisons.md
+++ b/liaisons.md
@ -57,6 +57,7 @@ members will assume one of the departing members groups.
 | [WG API Expression](wg-api-expression/README.md) | Benjamin Elder (**[@BenTheElder](https://github.com/BenTheElder)**) |
 | [WG Batch](wg-batch/README.md) | Bob Killen (**[@mrbobbytables](https://github.com/mrbobbytables)**) |
 | [WG Data Protection](wg-data-protection/README.md) | Patrick Ohly (**[@pohly](https://github.com/pohly)**) |
+| [WG Device Management](wg-device-management/README.md) | Patrick Ohly (**[@pohly](https://github.com/pohly)**) |
 | [WG LTS](wg-lts/README.md) | Nabarun Pal (**[@palnabarun](https://github.com/palnabarun)**) |
 | [WG Policy](wg-policy/README.md) | Patrick Ohly (**[@pohly](https://github.com/pohly)**) |
 | [WG Structured Logging](wg-structured-logging/README.md) | Nabarun Pal (**[@palnabarun](https://github.com/palnabarun)**) |
--- a/sig-architecture/README.md
+++ b/sig-architecture/README.md
@ -57,6 +57,7 @@ The Chairs of the SIG run operations and processes governing the SIG.

 The following [working groups][working-group-definition] are sponsored by sig-architecture:
 * [WG API Expression](/wg-api-expression)
+* [WG Device Management](/wg-device-management)
 * [WG LTS](/wg-lts)
 * [WG Policy](/wg-policy)
 * [WG Structured Logging](/wg-structured-logging)
--- a/sig-autoscaling/README.md
+++ b/sig-autoscaling/README.md
@ -47,6 +47,7 @@ The Chairs of the SIG run operations and processes governing the SIG.

 The following [working groups][working-group-definition] are sponsored by sig-autoscaling:
 * [WG Batch](/wg-batch)
+* [WG Device Management](/wg-device-management)


 ## Subprojects
--- a/sig-list.md
+++ b/sig-list.md
@ -64,6 +64,7 @@ When the need arises, a [new SIG can be created](sig-wg-lifecycle.md)
 |[API Expression](wg-api-expression/README.md)|[api-expression](https://github.com/kubernetes/kubernetes/labels/wg%2Fapi-expression)|* API Machinery<br>* Architecture<br>|* [Antoine Pelisse](https://github.com/apelisse), Google<br>* [Kevin Wiesmueller](https://github.com/kwiesmueller), Google<br>|* [Slack](https://kubernetes.slack.com/messages/wg-api-expression)<br>* [Mailing List](https://groups.google.com/forum/#!forum/kubernetes-wg-api-expression)|* Regular WG Meeting: [Tuesdays at 9:30 PT (Pacific Time) (biweekly)](https://zoom.us/j/94238112084)<br>
 |[Batch](wg-batch/README.md)|[batch](https://github.com/kubernetes/kubernetes/labels/wg%2Fbatch)|* Apps<br>* Autoscaling<br>* Node<br>* Scheduling<br>|* [Aldo Culquicondor](https://github.com/alculquicondor), Google<br>* [Marcin Wielgus](https://github.com/mwielgus), Google<br>* [Maciej Szulik](https://github.com/soltysh), Red Hat<br>* [Swati Sehgal](https://github.com/swatisehgal), Red Hat<br>|* [Slack](https://kubernetes.slack.com/messages/wg-batch)<br>* [Mailing List](https://groups.google.com/a/kubernetes.io/g/wg-batch)|* Regular Meeting ([Calendar](https://calendar.google.com/calendar/embed?src=8ulop9k0jfpuo0t7kp8d9ubtj4%40group.calendar.google.com)): [Thursdays (starting February 15th 2024)s at 3PM CET (Central European Time) (monthly)](https://zoom.us/j/98329676612?pwd=c0N2bVV1aTh2VzltckdXSitaZXBKQT09)<br>* Regular Meeting ([Calendar](https://calendar.google.com/calendar/embed?src=8ulop9k0jfpuo0t7kp8d9ubtj4%40group.calendar.google.com)): [Thursdays (starting February 1st 2024)s at 3PM PT (Pacific Time) (monthly)](https://zoom.us/j/98329676612?pwd=c0N2bVV1aTh2VzltckdXSitaZXBKQT09)<br>
 |[Data Protection](wg-data-protection/README.md)|[data-protection](https://github.com/kubernetes/kubernetes/labels/wg%2Fdata-protection)|* Apps<br>* Storage<br>|* [Xing Yang](https://github.com/xing-yang), VMware<br>* [Xiangqian Yu](https://github.com/yuxiangqian), Google<br>|* [Slack](https://kubernetes.slack.com/messages/wg-data-protection)<br>* [Mailing List](https://groups.google.com/forum/#!forum/kubernetes-data-protection)|* Regular WG Meeting: [Wednesdays at 9:00 PT (Pacific Time) (bi-weekly)](https://zoom.us/j/6933410772)<br>
+|[Device Management](wg-device-management/README.md)|[device-management](https://github.com/kubernetes/kubernetes/labels/wg%2Fdevice-management)|* Architecture<br>* Autoscaling<br>* Network<br>* Node<br>* Scheduling<br>|* [John Belamaric](https://github.com/johnbelamaric), Google<br>* [Kevin Klues](https://github.com/klueska), NVIDIA<br>* [Patrick Ohly](https://github.com/pohly), Intel<br>|* [Slack](https://kubernetes.slack.com/messages/wg-device-management)<br>* [Mailing List](https://groups.google.com/a/kubernetes.io/g/wg-device-management)|* Regular WG Meeting: [Tuesdays at 8:30 PT (Pacific Time) (biweekly)](TBD)<br>
 |[LTS](wg-lts/README.md)|[lts](https://github.com/kubernetes/kubernetes/labels/wg%2Flts)|* Architecture<br>* Cluster Lifecycle<br>* K8s Infra<br>* Release<br>* Security<br>* Testing<br>|* [Jeremy Rickard](https://github.com/jeremyrickard), Microsoft<br>* [Jordan Liggitt](https://github.com/liggitt), Google<br>* [Micah Hausler](https://github.com/micahhausler), Amazon<br>|* [Slack](https://kubernetes.slack.com/messages/wg-lts)<br>* [Mailing List](https://groups.google.com/a/kubernetes.io/g/wg-lts)|* Regular WG Meeting: [Tuesdays at 07:00 PT (Pacific Time) (biweekly)](https://zoom.us/j/92480197536?pwd=dmtSMGJRQmNYYTIyZkFlQ25JRngrdz09)<br>
 |[Policy](wg-policy/README.md)|[policy](https://github.com/kubernetes/kubernetes/labels/wg%2Fpolicy)|* Architecture<br>* Auth<br>* Multicluster<br>* Network<br>* Node<br>* Scheduling<br>* Storage<br>|* [Jim Bugwadia](https://github.com/JimBugwadia), Kyverno/Nirmata<br>* [Poonam Lamba](https://github.com/poonam-lamba), Google<br>* [Andy Suderman](https://github.com/sudermanjr), Fairwinds<br>|* [Slack](https://kubernetes.slack.com/messages/wg-policy)<br>* [Mailing List](https://groups.google.com/forum/#!forum/kubernetes-wg-policy)|* Regular WG Meeting: [Wednesdays at 8:00 PT (Pacific Time) (semimonthly)](https://zoom.us/j/7375677271)<br>
 |[Structured Logging](wg-structured-logging/README.md)|[structured-logging](https://github.com/kubernetes/kubernetes/labels/wg%2Fstructured-logging)|* API Machinery<br>* Architecture<br>* Cloud Provider<br>* Instrumentation<br>* Network<br>* Node<br>* Scheduling<br>* Storage<br>|* [Mengjiao Liu](https://github.com/mengjiao-liu), DaoCloud<br>* [Patrick Ohly](https://github.com/pohly), Intel<br>|* [Slack](https://kubernetes.slack.com/messages/wg-structured-logging)<br>* [Mailing List](https://groups.google.com/forum/#!forum/kubernetes-wg-structured-logging)|
--- a/sig-network/README.md
+++ b/sig-network/README.md
@ -72,6 +72,7 @@ subprojects, and resolve cross-subproject technical issues and decisions.
 ## Working Groups

 The following [working groups][working-group-definition] are sponsored by sig-network:
+* [WG Device Management](/wg-device-management)
 * [WG Policy](/wg-policy)
 * [WG Structured Logging](/wg-structured-logging)

--- a/sig-node/README.md
+++ b/sig-node/README.md
@ -53,6 +53,7 @@ subprojects, and resolve cross-subproject technical issues and decisions.

 The following [working groups][working-group-definition] are sponsored by sig-node:
 * [WG Batch](/wg-batch)
+* [WG Device Management](/wg-device-management)
 * [WG Policy](/wg-policy)
 * [WG Structured Logging](/wg-structured-logging)

--- a/sig-scheduling/README.md
+++ b/sig-scheduling/README.md
@ -63,6 +63,7 @@ subprojects, and resolve cross-subproject technical issues and decisions.

 The following [working groups][working-group-definition] are sponsored by sig-scheduling:
 * [WG Batch](/wg-batch)
+* [WG Device Management](/wg-device-management)
 * [WG Policy](/wg-policy)
 * [WG Structured Logging](/wg-structured-logging)

--- a/sigs.yaml
+++ b/sigs.yaml
@ -3345,6 +3345,48 @@ workinggroups:
    liaison:
      github: pohly
      name: Patrick Ohly
+- dir: wg-device-management
+  name: Device Management
+  mission_statement: >
+    Enable simple and efficient configuration, sharing, and allocation of accelerators
+    and other specialized devices.
+
+    [Additional context](https://groups.google.com/a/kubernetes.io/g/dev/c/YWXGXe07A5w/m/OqLvdQ47BQAJ)
+
+  charter_link: charter.md
+  stakeholder_sigs:
+  - Architecture
+  - Autoscaling
+  - Network
+  - Node
+  - Scheduling
+  label: device-management
+  leadership:
+    chairs:
+    - github: johnbelamaric
+      name: John Belamaric
+      company: Google
+    - github: klueska
+      name: Kevin Klues
+      company: NVIDIA
+    - github: pohly
+      name: Patrick Ohly
+      company: Intel
+  meetings:
+  - description: Regular WG Meeting
+    day: Tuesday
+    time: "8:30"
+    tz: PT (Pacific Time)
+    frequency: biweekly
+    url: TBD
+    archive_url: https://docs.google.com/document/d/1qxI87VqGtgN7EAJlqVfxx86HGKEAc2A3SKru8nJHNkQ/edit?usp=sharing
+    recordings_url: TBD
+  contact:
+    slack: wg-device-management
+    mailing_list: https://groups.google.com/a/kubernetes.io/g/wg-device-management
+    liaison:
+      github: pohly
+      name: Patrick Ohly
 - dir: wg-lts
  name: LTS
  mission_statement: >
--- a/wg-device-management/README.md
+++ b/wg-device-management/README.md
@ -0,0 +1,42 @@
+<!---
+This is an autogenerated file!
+
+Please do not edit this file directly, but instead make changes to the
+sigs.yaml file in the project root.
+
+To understand how this file is generated, see https://git.k8s.io/community/generator/README.md
+--->
+# Device Management Working Group
+
+Enable simple and efficient configuration, sharing, and allocation of accelerators and other specialized devices.
+[Additional context](https://groups.google.com/a/kubernetes.io/g/dev/c/YWXGXe07A5w/m/OqLvdQ47BQAJ)
+
+The [charter](charter.md) defines the scope and governance of the Device Management Working Group.
+
+## Stakeholder SIGs
+* [SIG Architecture](/sig-architecture)
+* [SIG Autoscaling](/sig-autoscaling)
+* [SIG Network](/sig-network)
+* [SIG Node](/sig-node)
+* [SIG Scheduling](/sig-scheduling)
+
+## Meetings
+*Joining the [mailing list](https://groups.google.com/a/kubernetes.io/g/wg-device-management) for the group will typically add invites for the following meetings to your calendar.*
+* Regular WG Meeting: [Tuesdays at 8:30 PT (Pacific Time)](TBD) (biweekly). [Convert to your timezone](http://www.thetimezoneconverter.com/?t=8:30&tz=PT%20%28Pacific%20Time%29).
+  * [Meeting notes and Agenda](https://docs.google.com/document/d/1qxI87VqGtgN7EAJlqVfxx86HGKEAc2A3SKru8nJHNkQ/edit?usp=sharing).
+  * [Meeting recordings](TBD).
+
+## Organizers
+
+* John Belamaric (**[@johnbelamaric](https://github.com/johnbelamaric)**), Google
+* Kevin Klues (**[@klueska](https://github.com/klueska)**), NVIDIA
+* Patrick Ohly (**[@pohly](https://github.com/pohly)**), Intel
+
+## Contact
+- Slack: [#wg-device-management](https://kubernetes.slack.com/messages/wg-device-management)
+- [Mailing list](https://groups.google.com/a/kubernetes.io/g/wg-device-management)
+- [Open Community Issues/PRs](https://github.com/kubernetes/community/labels/wg%2Fdevice-management)
+- Steering Committee Liaison: Patrick Ohly (**[@pohly](https://github.com/pohly)**)
+<!-- BEGIN CUSTOM CONTENT -->
+
+<!-- END CUSTOM CONTENT -->
--- a/wg-device-management/charter.md
+++ b/wg-device-management/charter.md
@ -0,0 +1,103 @@
+# WG Device Management Charter
+
+This charter adheres to the conventions described in the [Kubernetes Charter
+README] and uses the Roles and Organization Management outlined in
+[wg-governance].
+
+## Scope
+
+Enable simple and efficient configuration, sharing, and allocation of
+accelerators and other specialized devices. This working group focuses on the
+APIs, abstractions, and feature designs needed to configure, target, and share
+the necessary hardware for both batch and serving (inference) workloads.
+
+### In scope
+
+- Enable efficient utilization of specialized hardware devices. This includes
+  sharing one or more resources effectively (many workloads sharing a pool of
+  devices), as well as sharing individual devices effectively (several workloads
+  dividing up a single device for sharing).
+- Enable workload authors to specify “just enough” details about their workload
+  requirements to ensure it runs optimally, without having to understand exactly
+  how the infrastructure team has provisioned the cluster.
+- Enable the scheduler to choose the correct place to run a workload the vast
+  majority of the time (rejections should be extremely rare).
+- Enable cluster autoscalers and other node auto-provisioning components to
+  predict whether creating additional resources will satisfy workload needs,
+  before provisioning those resources.
+- Enable the shift from “pods run on nodes” to “workloads consume capacity”.
+  This allows Kubernetes to provision sets of pods on top of sets of nodes and
+  specialized hardware, while taking into account the relationships between
+  those infrastructure components.
+- Enable in-node devices as well as network-accessible devices.
+- Minimize workload disruption due to hardware failures.
+- Address fragmentation of accelerator due to fractional use.
+- Additional problems that may be identified and deemed in scope as we gather
+  use cases and requirements from WG Serving, WG Batch, and other stakeholders.
+- Address all of the above while with a simple API that is a natural extension
+  of the existing Kubernetes APIs, and avoids or minimizes any transition
+  effort.
+
+### Out of Scope
+
+- Higher-level workload controller APIs (for example, the equivalent of
+  Deployment, StatefulSet, or DaemonSet) for specific types of workloads.
+- General resource management requirements not related to devices.
+
+## Deliverables
+
+The WG will coordinate the delivery of KEPs and their implementations by the
+participating SIGs. Interim artifacts will include documents capturing use
+cases, requirements, and designs; however, all of those will eventually result
+in KEPs and code owned by SIGs.
+
+Specifically, we expect to need:
+
+- APIs for publishing resource capacity of in-node and network-accessible
+  devices, as well as sample code to ease creation of drivers to populate this
+  information.
+- APIs for specifying workload resource requirements with respect to devices.
+- APIs, algorithms, and implementations for allocating access to and resources on devices, as well as
+  persisting the results of those allocations.
+- APIs, algorithms, and implementations for allowing adminstrators to control
+  and govern access to devices.
+
+## Stakeholders
+
+- SIG Architecture
+- SIG Autoscaling
+- SIG Network
+- SIG Node
+- SIG Scheduling
+
+Additionally a broad set of end users, device vendors, cloud providers,
+Kubernetes distribution providers, and ecosystem projects (particularly
+autoscaling-related projects) have expressed interest in this effort. There are
+five primary groups of stakeholders from each of which we expect multiple participants:
+
+- Device vendors that manufacture accelerators and other specialized hardware
+  which they would like to make available to Kubernetes users.
+- Kubernetes distribution and managed offering providers that would like to make
+  specialized hardware available to their users.
+- Kubernetes ecosystem projects that help manage workloads utilizing these
+  accelerators (e.g., Karpenter, Kueue, Volcano)
+- End user workload authors that will create workloads that take advantage of
+  the specialized hardware.
+- Cluster administrators that operate and govern clusters containing the
+  specialized hardware.
+
+## Roles and Organization Management
+
+This working group adheres to the Roles and Organization Management outlined in
+[wg-governance] and opts-in to updates and modifications to [wg-governance].
+
+## Exit Criteria
+
+The working group will disband when the KEPs resulting from these discussions
+have reached a terminal state. When the core functionality for dynamic resource
+allocation (DRA) reaches GA, we will evaluate whether the working group should
+be disbanded and any remaining KEPs be left to the management of their owning
+SIGs.
+
+[wg-governance]: https://github.com/kubernetes/community/blob/master/committee-steering/governance/wg-governance.md
+[Kubernetes Charter README]: https://github.com/kubernetes/community/blob/master/committee-steering/governance/README.md