Release Notes#

Revision History - 24.06#

Profile

Summary

Profile 1 : VA Enablement Node K8s Profile

1. Installed RKE2 agent

Profile 2 : VA Enablement Node Profile

1. Installed Media Tools - Opencv, FFMPEG, MTL, Mesa Driver, oneVPL, MCM, Intel XPU Manager, Intel Media Driver
2. Installer Observability Stack for Baremetal OS - Platform-Observability-agent, Prometheus, Grafana, cAdvisor
3. KVM and QEMU: Supported for bare-metal virtualization.
4. Installed AI BOX Components - VA Base, VA Base-Dev, VA DLStreamer, VA OpenCV FFMPEG
5. Inband Management: Support for Turtle Creek has been added.
6. iGPU support enabled for core platforms.

Profile 3 : RT Enablement Node Profile

1. Installed RT kernel
2. Installed Real-Time Performance Measurement (RTPM)
3. Installed ECI Packages

Revision History - 24.08#

Profile

Summary

Profile 1 : VA Enablement Node K8s Profile

1. Inband Management: Support for Turtle Creek has been added.
2. Observability Stack for Kubernetes Cluster:
    - Platform-Observability Agent Installed directly on the OS, covering OpenTelemetry (Otel), FluentBit, and Telegraf.
    - cAdvisor: Installed for monitoring container resource usage.
    - Prometheus: Installed using the prometheus-community/prometheus Helm chart.
    - Grafana: Installed using the grafana/grafana Helm chart.
3. Network SRIOV Plugins: Added for enhanced networking capabilities.
4. Network Feature Discovery: Implemented using Docker.
5. Akri: Deployed with Docker for edge device discovery.
6. IPSec Stack: Introduced for enhanced network security.
7. GPU Plugins: Available through Helm charts for improved GPU management.
8. KVM and QEMU: Supported for bare-metal virtualization.
9. Istio Operator: Installed to manage service mesh architecture.
10. OpenVINO Runtime and OpenCV: Added for AI inference and computer vision tasks.
11. GPU Drivers: Integrated to enhance GPU support.
12. Kubernetes Virtualization: Improved virtualization capabilities within the Kubernetes environment.

Profile 2 : VA Enablement Node Profile

1. Enabled GPU drivers for Xeon platforms
2. Introduced separate sets of AI Box Docker files and Docker validators for each supported platform, incorporating the necessary GPU/CPU enablement in the Release service.
Included support for the Alder lake platform in AI Box Docker image build and validation.
3. Enhanced the installer to set “6.6-intel” as the default kernel, which remains the default even after a reboot.
4. Resolved DL-Streamer Docker build failures by generating separate installation scripts for each platform, addressing the required GPU/CPU enablement.
5. Implemented functionality to handle the expiration of the External Release service token and automatically re-authenticate using a refresh token.
6. Added dynamic geo-location detection and automatic GitHub mirror updates specifically for the PRC region.

Profile 3 : RT Enablement Node Profile

1. Enabled Observability Stack: Integrated the Observability Stack along with the ECI packages.
2. Installed the observability components (Platform-Observability-agent (covers, Otel, fluentbit, telegraf) Prometheus, Grafana, cAdvisor).
3. Installed Docker Images: Installed Docker images for Grafana, Loki, and Promtail.

Revision History - 24.11#

Feature :-#

  • Added Uninstallation feature for all the profiles

SBOM :-#

Profile

Summary

Profile 1 : VA Enablement Node K8s Profile

1. Upgraded to Ubuntu 24.04
2. enabled multi-node configuration
3. Integrated Media Tools and DPDK
4. Removed the helm chart installation steps from profile 1 script
5. Created a new script for helm chart installation to support the multi-edge worker nodes
6. Removed platform observability , fluent bit, telegraf, from baremetal
7. Enabled SRIOV plugin support

Profile 2: VA Enablement Node Profile

1. Removed Release service token dependency
2. Enabled NPU Support
3. Added NTP Server prompt. User can directly enter their preferred NTP server