Abstract: Cloud service providers are distributing data centers globally to reduce operating costs while also improving the quality of service by using intelligent cloud management strategies. The development of time-of-use electricity pricing and renewable energy source models has provided the means to reduce high cloud operating costs through intelligent geographical workload distribution. However, neglecting essential considerations such as data center cooling power, interference effects from workload co-location in servers, net-metering, peak demand pricing of electricity, data transfer costs, and data center queueing delay has led to sub-optimal results in prior work because these factors have a significant impact on cloud operating costs, performance, and carbon emissions. In this dissertation, we propose a suite of cloud management techniques that take a holistic approach to the cloud operating (energy and data transfer) costs and carbon emissions minimization problem for geo-distributed data centers. Our algorithmic techniques perform intelligent workload management across geo-distributed data centers while considering heterogeneity in data center compute capability, cooling power, interference effects from workload co-location in servers, time-of-use electricity pricing, renewable energy, net metering, peak demand pricing distribution, network costs, and carbon emissions. We demonstrate the value of utilizing such information by comparing it against state-of-the-art geo-distributed workload management techniques considering varying amounts of system information. Our experimental analysis indicates that the proposed techniques can more effectively minimize cloud energy expenditures and carbon emissions than existing approaches.

Adviser: Dr. Sudeep Pasricha
Co-Adviser: N/A
Non-ECE Member: Dr. Chuck Anderson
Member 3: Dr. H. J. Siegel
Addional Members: Dr. Anthony Maciejewski

Publications:
Journals
[1] N. Hogade and S. Pasricha, "Game Theoretic DRL for Reducing Carbon Emissions and Operating Costs of AI Workloads in Geo-Distributed Data Centers," [In Preparation - IEEE Transactions].
[2] N. Hogade and S. Pasricha, "A Survey on Machine Learning for Geo-Distributed Cloud Data Center Management," IEEE Transactions on Sustainable Computing, vol. 8, no. 1, pp. 15-31, 1 Jan.-March 2023
[3] N. Hogade, S. Pasricha, and H. J. Siegel, "Energy and Network Aware Workload Management for Geographically Distributed Data Centers," IEEE Transactions on Sustainable Computing, Vol. 7, No. 2, pp. 400-413, June. 2021.
[4] N. Hogade, S. Pasricha, H. Jay Siegel, A. A. Maciejewski, M. A. Oxley, and E. Jonardi, "Minimizing Energy Costs for Geographically Distributed Heterogeneous Data Centers," IEEE Transactions on Sustainable Computing, Vol. 3, No. 4, pp. 318-331, Oct.-Dec. 2018.

Other
[5] G. Rattihalli, N. Hogade, A. Dhakal, E. Frachtenberg, R. Pablo Hong Enriquez, Pedro Bruel, A. Mishra, and D. Milojicic, "Fine-Grained Heterogeneous Execution Framework with Energy Aware Scheduling," to appear in IEEE Cloud, 2023
[6] C. Bash, N. Hogade, D. Milojicic, C. D. Patel, G. Rattihalli, "Sustainability: Fundamentals-based Approach to Paying it Forward," IEEE Computer, Volume: 56, Issue: 1, January 2023.
[7] C. Bash, N. Hogade, D. Milojicic, C. D. Patel, "IT for Sustainable Smart Cities," National Academy of Engineering, Special Issue on "Advances in Smart Cities," 2023.
[8] S. Pasricha, N. Hogade, H. J. Siegel, and A. A. Maciejewski, "Green Computing with Geo-Distributed Heterogeneous Data Centers," International Green and Sustainable Computing Conference (IGSC), 6 pp., Oct. 2019.

Program of Study:
GRAD-510 Fundamentals of High-Performance Computing
ECE-561 Hardware/Software Design of Embedded Systems
ECE-554 Computer Architecture
CS-575 Parallel Processing
CS-545 Machine Learning
CS-420 Introduction to Analysis of Algorithms
ENGR-510 Engineering Optimization: Method/Application
GRAD-511 High-Performance Computing and Visualization