Table 1 Global configurations for the experiments

num of rack = 12; server per rack = 8; simulation unit time = 5 (min)

service generation speed = 300; service lifetime = [300, 300]

service cpu, memory, and network bandwidth runtime usage distribution: beta(2, 4)

memory dirtying rate: beta(2, 4); server models = [1 - 8]

server cpu, memoryt, and network bandwidth under-/over-utilization threshold = {20, 90}

consolidation interval = 12; heterogeneity: {data center, rack}