ECE PhD Thesis Defense: Fatih Acun

ECE PhD Thesis Defense: Fatih Acun

Title: From Load to Grid Asset: Coordinated Methods for Multi-Participant Data Center Demand Response

Presenter: Fatih Acun

Advisor: Professor Ayșe Coskun

Chair: Professor Selim Ünlü

Committee: Ayșe Coskun, Professor Ioannis Paschalidis, Professor Emiliano Dall’Anese, Dr. Carole-Jean Wu

Google Scholar Link: https://scholar.google.com/citations?user=iwSBZeYAAAAJ&hl=en&oi=ao

Abstract: The rapid growth of data centers has raised substantial concerns due to their high energy demands, particularly with the surge in artificial intelligence workloads. As large-scale power consumers, data centers put significant stress on power grids, driving up the demand for expanded generation and transmission capacity, as well as increasing electricity costs. To address grid stability and balance power supply and demand, Independent System Operators (ISOs) implement demand response (DR) programs to utilize the flexibility of demand-side consumers and enable sustainable growth of power grids. Although data centers are well-positioned to participate in DR through power management and workload scheduling techniques, real-world adoption remains limited, partially due to concerns over workload performance degradation and compatibility with DR program requirements.

This thesis argues that coordinated DR participation among data centers (or data center tenants) can meet quality-of-service (QoS) requirements while utilizing flexibility across participants to increase reserve capacity, improve power-target tracking, and reduce electricity costs compared with independent participation. Towards this claim, the thesis makes three contributions at the intersection of data center power management and power grids: (1) design of coordinated frameworks for multi-participant data center DR, using the aggregate flexibility to mitigate the risks of violating QoS requirements and improve DR power tracking capabilities, (2) design and open-source implementation of a data center DR simulator to enable testing of power management policies at scale for various DR programs, (3) power–performance analysis of HPC workloads on large-scale production GPU systems, alongside predictive frameworks for GPU power consumption.

The overarching goal of this thesis is to develop methods that enhance the performance of data centers concerning QoS constraints and DR requirements, while enabling their operation as flexible power consumers to relieve strain on power grids. To this end, this thesis approaches the problem from both theoretical and practical perspectives by employing simulation and joint optimization techniques, as well as empirical studies of power management and prediction on production systems. Our results show that, using coordinated methods, data centers can mitigate violations of QoS and power-tracking constraints while providing 22% greater flexibility to power grids than individual DR participation.