BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260730T152637Z
LOCATION:
DTSTART;TZID=America/Los_Angeles:20260728T124400
DTEND;TZID=America/Los_Angeles:20260728T124400
UID:dac_DAC 2026_sess315_RESEARCH2970@linklings.com
SUMMARY:GREEN: Towards Scalable Energy-Efficient Workload Scheduling and P
 lacement in the Cloud
DESCRIPTION:Jinghua Wang (University of Illinois Urbana-Champaign); Asser 
 Tantawi, Olivier Tardieu, Alaa Youssef, Tamar Eilam, Chen Wang, and Pradip
  Bose (IBM Research); JIAXIN WAN and Klara Nahrstedt (University of Illino
 is Urbana-Champaign); and Deming Chen (UIUC)\n\nThe workload scheduling an
 d placement problem has been the core of resource management in various di
 stributed systems, including modern cloud computing systems. As modern clo
 ud systems become larger, more heterogeneous, dynamic, complicated, and en
 ergy-hungry, developing a scalable, flexible, adaptive, and effective ener
 gy-efficient resource manager for such a cloud system is becoming very cha
 llenging. Specifically, existing machine learning-based resource managers 
 cannot scale well, and existing heuristic algorithm-based scalable resourc
 e managers do not generate the most energy-efficient solutions. To address
  the limitations in existing methods and solve the problem more effectivel
 y, we propose GREEN, a graph neural network-reinforcement learning (GNN-RL
 ) based cloud resource manager that generates high energy-efficiency workl
 oad scheduling and placement solutions and scales up to large cloud system
 s with hundreds and thousands of servers. GREEN reduces this challenging p
 roblem to a graph optimization problem and then uses our novel RL formulat
 ion and GNN architecture for generating scheduling and placement solutions
  for cloud systems in various scales. Through extensive cloud simulations 
 using COSCO and real-world experiments using CloudLab, we found that GREEN
 's solutions save energy by up to 2.17x than those generated by the best p
 revious state-of-the-art (SOTA) resource managers without compromising Ser
 vice Level Objective (SLO) metrics. Most importantly, GREEN can generate s
 imilarly high quality scheduling and placement solutions on systems with 1
 00 to 1000 servers in COSCO cloud simulations.\n\nTopics: Quantum\n\nTrack
 : DES6. Quantum Computing, Systems\n\nSession Chairs: Hokeun Kim (Arizona 
 State University); BaekGyu Kim (Daegu Gyeongbuk Institute of Science and T
 echnology); Xiang Chen (Peking University); Sangyeob Kim (Yonsei Universit
 y); Robert M. Radway (University of Pennsylvania); Prashanth Mohan (Univer
 sity of Pittsburgh); Xiaoxuan Yang (University of Virginia, Stanford Unive
 rsity); Ramtin Zand (University of South Carolina); Hadi Kamali (Universit
 y of Central Florida); Jiafeng Xie (Villanova University); Prabuddha Chakr
 aborty (University of Maine); Junichiro Kadomoto (University of Tokyo); an
 d Yosuke Ueno (RIKEN)\n\n
END:VEVENT
END:VCALENDAR