SYM – Title Slide

OCP Global Summit

October 18, 2023 | San Jose, CA

Life Cycle Management @ Scale

Nirav Shah, Cloud Software Architect, Intel Corporation

Scott Ramsey, Technologist, Dell Inc.

Acknowledgements: Anil Agrawal (Meta), George Kola (Google), Changho Choi (Samsung), Chukwunenye Nnebe (Microsoft), John Leung (Intel), Murugasamy Nachimuthu (Intel), Panos Christeas (Meta)

SYM - Content

Data Center @ scale

100s of thousands of geo located and/or distributed servers!

Ocean!

Ocean!

Rack 1

Rack N (geo-located)

Rack Z(distributed)

Hyperscale Cloud?

• Heterogeneous configurations.
• Economy of scale (cost savings)
• A Hierarchy of data
• AI enabled processing
• Security

Components

Systems

Racks

Data

Center

Telemetry Data Sharing 🡪 Recap

Tiers of system aggregation

Component/Resource aggregation

Privilege Panel

Hyperscale Cloud Life-Cycle 🡪 Opportunities

• Real Estate
• Utilities
• Network
• Systems sourcing
• Configuration
• Conformance/Certification
• Validation
• Orchestration/Ramp
• Utilization
• Fault Detection/Repair
• Servicing
• Inventory management
• Backup & Reset
• Decommission
• Recycle

• Ecological & Sustainable (Sustainability)
• Source (Modular Data Center)
• Consistent( Ready™ Facility Recognition)
• Setup Networking, Security
• Configure Security, Storage, Server, Power, F/W
• Conform and/or Certify
• Orchestrate/Ramp
• Hardware Management
• Fault Management
• Software management
• Inventory management
• Backup & Reset
• Decommission
• Recycle

STAGES

FACTORS

STANDARDS

Data Ocean problem 🡪 A case study

Example Device/System data tree

​

Device or System Resource

Type

identifier

Hardware failure

Hardware failure

Utilization

Health Status

identifier

Unique ID

Version

Version

Date

Status

Last Checked

Uncorrectable error count

uncorrectable error threshold

Corrected error count

Corrected error threshold

Corrected error policy

uncorrected error policy

type

identifier

Signature

Version

raw data

compliance and warranty

Type

uptime

utilization

last utilization reset

last updatime reset

Hyperscale Data flow

System

System

System

Device A

Device Z

……………

System

System

Device/System Data Model

Device A

Device Z

……………..…

IB Agent

OOB Agent

Vendor AI/ML Aggregation Logic

Hyperscale Data Model

Summary

• Data centers sizes growing at an exponential scale
• Define a hyperscale data
• Applying the general life cycle to a hyperscale data center
• Identified unique opportunities for standardization
• A case study projecting advantages of standardization vis-à-vis aspects of the life cycle.

​

Call to Action

• Engage with OCP Hardware Management Project “OCP - Cloud Service Model” workstream to participate in defining “Life Cycle Management @ Scale” requirements and interfaces
• https://ocp-all.groups.io/g/Cloud-Service-Model/calendar
• https://www.opencompute.org/projects/hardware-management

BO - Content

Image

Replace this image with your full-screen photo

SYM - End

This should be the final slide

Click here to review OCP Trademark Guidelines

Please remove all yellow notes before submitting your presentation.

OCP Global Summit | October 18, 2023 | San Jose, CA

Life Cycle of a hyperscale Cloud

• Orchestration
• Budget Power/Cost balance
• TCO
• Real Estate
• Sustainability Supply Chain requirement

​

Device/System 🡪 Example Data Model

Device/System Resource

• Type
• Identifier
• Hardware failures
• Software failures
• Utilization
• Compliance/Warrantee/Certification
• Health Status

​

​

​

Bug/Fault Resource

• Type
• Global unique ID
• Telemetry Signature
• Impacted version
• Status

SYM - Content

Health Status Resource

• Unrecovered failures #
• Unrecovered failure threshold #
• Recovered failures #
• Recovery action status

Utilization Resource

• Type
• Uptime
• Avg Utilization
• Last reset Uptime
• Last reset Avg Utilization

Certificate/Compliance Status

• Last checked
• Status
• Due date

Identifier Resource

• Global unique ID
• Software/Firmware version
• Hardware Version

Hyperscale Opportunities

• Economy of scale (cost savings)
• A Hierarchy of structured data
• 100Ks of racks
• w/ heterogenous system configs.

​

Life Cycle Manageability @ Scale Overview

Goal

• A call for a standard data representation model for life cycle manageability @ scale.
• Focus areas:
• Health status thresholds for Maintenance/Repairs/Decommission
• Software/Firmware quality thresholds for Diagnostics/Serviceability
• Telemetry for overall security/regulatory compliance.
• Hyperscale management of Inventory thresholds
• Flexibility to allow for Vendor specific AI/ML aggregation

​

SYM - Content

What Goes Through the lifecycle?

Proposed Redfish Hyperscale data tree

Hyperscale Resource

• Compliance Status

​