JVD Validation Goals and Scope

Tests Objectives

The primary objectives of the JVD testing can be summarized as:

Qualification of the complete AI fabric design functionality including the Frontend, GPU Backend, and Storage Backend fabrics, and connectivity between AMD GPUs and Vast Storage.
Qualification of the deployment steps based on Juniper Apstra.
Ensure the design is well-documented and will produce a reliable, predictable deployment for the customer.
Validation of Thor2 5m DAC operation with Juniper’s TH5 chip.
Validation of QFX and PTX platforms in different roles in the fabric.

The qualification objectives included validating:

Validation of blueprint deployment, device upgrade, incremental configuration pushes/provisioning, Telemetry/Analytics checking, failure mode analysis, congestion avoidance and mitigation, and verification of host, storage, and GPU traffic.

The AI JVD testing for the described network included the following:

Design and blueprint deployment through Apstra of three distinct fabrics
Fabric operation and monitoring through Apstra analytics and telemetry dashboard
Congestion management with PFC and ECN, including failure scenarios
End-to-end traffic flow, with Dynamic Load Balancing (DLB)
System health, ARP, ND, MAC, BGP (route, next hop), interface traffic counters, and so on
Software operation verification (no anomalies, or issues found)
Validation of Thor2 5m DAC operation with Juniper’s TH5 chip.
AI fabric with Juniper Apstra successfully performing under the following required scenarios (must):
- Node failure (reboot)
- Interface failures (interface down/up, Laser on/off)

Under these scenarios the following were evaluated/validated:

BCM957608 THOR2 NICs
Pollara NICs
Mellanox Connect-X NIC card default settings.
DSCP and CNP configuration on the NICs
Connectivity between fabric-connected hosts created by Apstra towards NSX-managed hosts.
BERT/LLAMA3 test completion times
Llama2 Inference against existing infrastructure.

Refer to the test report for more information.