Formal and Empirical Study of Metadata-Based Profiling for Resource Management in the Computing Continuum

TL;DR

This work presents PolarisProfiler, a metadata-driven approach to workload profiling for resource management across the Computing Continuum. By combining static a priori workload metadata with historical execution traces, it builds profile groups via an unsupervised Profile Generator and assigns new workloads to these groups with a metadata-based Profile Classifier, all under a dynamic Feedback Loop that adapts over time. The framework introduces ACQUIRES, a composite quality metric, and demonstrates robust performance on Alibaba PAI and Google cluster data, achieving accurate real-time estimates for most workloads and enabling informed, efficient orchestration. The study further extends the analysis to Google traces and provides a public implementation, highlighting the method's generality, scalability, and potential to reduce overprovisioning while meeting SLOs in diverse, heterogeneous environments.

Abstract

We present and formalize a general approach for profiling workload by leveraging only a priori available static metadata to supply appropriate resource needs. Understanding the requirements and characteristics of a workload's runtime is essential. Profiles are essential for the platform (or infrastructure) provider because they want to ensure that Service Level Agreements and their objectives (SLOs) are fulfilled and, at the same time, avoid allocating too many resources to the workload. When the infrastructure to manage is the computing continuum (i.e., from IoT to Edge to Cloud nodes), there is a big problem of placement and tradeoff or distribution and performance. Still, existing techniques either rely on static predictions or runtime profiling, which are proven to deliver poor performance in runtime environments or require laborious mechanisms to produce fast and reliable evaluations. We want to propose a new approach for it. Our profile combines the information from past execution traces with the related workload metadata, equipping an infrastructure orchestrator with a fast and precise association of newly submitted workloads. We differentiate from previous works because we extract the profile group metadata saliency from the groups generated by grouping similar runtime behavior. We first formalize its functioning and its main components. Subsequently, we implement and empirically analyze our proposed technique on two public data sources: Alibaba cloud machine learning workloads and Google cluster data. Despite relying on partially anonymized or obscured information, the approach provides accurate estimates of workload runtime behavior in real-time.

Figures (16)

• Figure 1: Overview of the PolarisProfiler's model.
• Figure 2: Visual representation of the PolarisProfiler components, actors and their interactions in the model lifecycle.
• Figure 3: Flowchart diagram of the definition of the Profile Generator.
• Figure 4: Flowchart diagram of the definition of the Profile Classifier.
• Figure 5: Flowchart diagram of the process of feedback loop for the PolarisProfiler.