nvPAX: Constrained Optimization for Dynamic Power Allocation in Hierarchical and Multi-Tenant Systems

TL;DR

nvPAX is a hierarchical, constrained-optimization power allocation policy for datacenters that maximizes utilization and satisfaction ratio through a three-phase hybrid QP/LP approach.

Contribution

It introduces a novel three-phase hybrid QP/LP method for dynamic power allocation respecting hierarchical and tenant constraints.

Findings

nvPAX achieves a 98.92% satisfaction ratio in large-scale GPU simulations.

It runs with a mean of 264.69 ms per allocation interval.

Outperforms static equal-share and greedy proportional allocation methods.

Abstract

Power oversubscription is increasingly central to datacenter operation as power density grows, making it necessary to dynamically allocate limited power budgets across devices based on real-time demand. Existing approaches typically assume flat power domains, whereas in practice power distribution is hierarchical and allocation decisions must additionally respect tenant-level contractual constraints. We present nvPAX, a constrained-optimization policy that computes feasible power allocations at every control step via a three-phase hybrid QP/LP procedure. Phase I allocates power with minimum deviation from each device's power request, while respecting job priorities. Phase II fairly distributes excess power among active devices. Phase III fairly distributes any remaining power to idle devices. The rationale behind the three phases is to allow power oversubscription while maximizing datacenter utilization. On a trace-driven large-scale simulation using GPU power telemetry from a production datacenter, nvPAX runs with a mean wall-clock time of 264.69 ms per allocation interval and achieves a mean satisfaction ratio of 98.92%, outperforming static equal-share allocation and providing robustness beyond greedy proportional allocation in the presence of non-uniform hierarchical bottlenecks.