Compute Credit Management

Overview

The IUDX-Novo platform implements a compute credit management model to ensure fair usage, cost transparency, and controlled access to analytical compute resources.

As described in earlier sections, all platform systems are deployed on Kubernetes, and user workloads are orchestrated using Kubeflow with Dex for identity and access control. This architecture enables fine-grained isolation and accurate accounting of compute usage.

Namespace-Based Isolation

Each user’s analytical activities are isolated using dedicated Kubernetes namespaces.

This design enables:

• Isolation of user workloads and resource usage

• Independent lifecycle management of user environments

• Strong security and fault isolation between users

Resource Provisioning Model

The platform supports flexible and efficient resource provisioning:

• User compute environments can be scaled independently within their isolated namespaces

• All compute resources are provisioned starting from minimal baseline configurations

• Resources scale dynamically based on workload demands and platform policies

This approach ensures optimal utilization while preventing resource overconsumption.

Compute Credit Calculation

Compute credits associated with each user are derived by monitoring Kubernetes resource utilization, including but not limited to:

• CPU usage

• Memory consumption

• Storage usage

• Execution duration of workloads

Metrics and cost attribution frameworks such as OpenCost can be deployed to:

• Collect detailed resource usage metrics

• Map resource consumption to platform costs

• Translate usage into standardized compute credit units based on the underlying machine pool configuration

Credit Storage and Enforcement

The computed compute credit usage is recorded and managed within the Data Exchange platform.

This information is used to:

• Enforce authorization decisions related to compute usage

• Control access to analytical and sandbox resources

• Support payment, quota enforcement, or credit replenishment workflows

Role in the Platform

• Enables fair and transparent usage of shared compute resources

• Supports sustainable platform operation and cost recovery

• Integrates governance, authorization, and billing with analytics execution

---