-
Epic
-
Resolution: Unresolved
-
Normal
-
None
-
None
-
None
-
None
-
Hardware Infrastructure for AI cloud
-
L
-
False
-
False
-
Green
-
Targeted
-
?
-
?
-
In Progress
-
?
-
?
-
23% To Do, 8% In Progress, 69% Done
-
-
-
1. DFG:AI - RHEL AI gaps, 2. AMD + RHEL AI DOC and CI
-
2
Summary:
As per requirements coming from customers, the team would like to define an extension of validated architecture towards AI-focused workloads; this requires additional HW for testing.
Business Value:
Validation of scenarios for AI workloads brings stability and quality to commonly deployed architectures that are tracked via the VA model. Longevity and real workload within the scenario should ensure the overall quality of the product and scenario.
Goals:
- Build a plan and acquire approvals for Hardware connected to AI optimised - Validated Architecture
Scope:
- ACK from Workloads Council, CapEx Council and IT on the next plan for HW
- Not in scope: HW configuration deployment and maintenance
Breakdown of Work:
- T - Get an estimation of the cost for HW for Full support of different features
- S - Define Quality criteria for Support exception, DevPreview, TechPreview, or Full support
- S - Document HW specs, connected to Features + delivery + DELL price estimations + timeline for delivery (DELL)
- S - CapEx - Need doc summary of HW possibilities
- T - Need quotas for HW from DELL
- S - Build presentation for Workload Council
- T - Present HW requirements to the IT Labs team
- T - Build a plan with IT
Constraints or Risks:
- Not getting ACK by CapEx, WC, IT
- No suitable lab in RH
- No Cloud provider with relevant specs
Dependencies:
- RHEL AI HW Specs
- AMD, NVidia GPU release
- Build a level of quality specs for the level of support for features
Acceptance Criteria:
- Documented plan approved by leadership for AI Optimized VA
- depends on
-
OSPRH-11981 Build document for level of support vs testing support for features
-
- Closed
-
- links to