Self-organizing sensing infrastructure for autonomic management of green datacenters

Abstract

34 0890-8044/11/$25.00 © 2011 IEEE atacenters are a growing component of society's information technology (IT) infrastructure, enabling services related to health, banking, commerce, defense, education, and entertainment. Due to the rise in demand for computing and storage, energy consumption, heat generation, and cooling requirements of datacenters have become critical concerns in terms of both the growing operating costs as well as their environmental and societal impacts. It is predicted that datacenter energy consumption in the United States will reach 100 billion kWh/year by the end of 2011 with a corresponding energy bill of approximately $7.4 billion [1]. Impact on the environment and society may include increase in CO 2 emissions, overload of the electricity supply grid, and rising water usage for cooling leading to water scarcity. The scale and complexity of modern datacenters are growing at an alarming rate due to the rising popularity of the cloud computing paradigm as an effective means to cater to the aforementioned increase in demand [2]. Their management is rapidly exceeding human ability, making autonomic (self-configuration, self-optimization, self-healing, and self-protection) management approaches essential.From our feasibility study and proof-of-concept experiments on our testbed [3] at the National Science Foundation (NSF) Center for Autonomic Computing (CAC) -a multiinstitutional research center that comprises four university sites: the University of Florida, University of Arizona, Rutgers University, and Mississippi State University -we have inferred that one of the fundamental problems in existing datacenters is the local unevenness in heat generation and heat extraction rates. The former can be attributed to non-uniform distribution of workloads among servers and the heterogeneity of computing hardware, while the latter can be attributed to non-ideal air circulation, which depends on the layout of server racks inside the datacenter and on the placement of computer room air conditioning (CRAC) unit fans and air vents. The heat generation and extraction rates may differ, which over time causes what we call heat imbalance. A large negative heat imbalance at a particular region inside a datacenter will result in energy-inefficient overcooling and hence a significant decrease in temperature. Conversely, a large positive heat imbalance in a particular region will lead to a significant temperature rise, which may increase the risk of equipment overheating and hence the chances of server system failures due to operation in the unsafe temperature range [4]. Thus, thermal awareness, which is the knowledge of heat imbalance in different regions inside a datacenter, is essential to maximize energy and cooling efficiency as well as to minimize server system failure rates.Autonomic datacenter management, which includes thermal-and energy-aware resource provisioning, cooling system optimization, and anomaly detection, can help minimize both the impact on the environment and the total cost of ownership (...

VMAP: Proactive thermal-aware virtual machine allocation in HPC cloud datacenters

“…The blade servers at both sites are equipped with a host of internal sensors that provide information about server subsystem operating temperatures and utilization. In addition, the machine room at RU is instrumented with an external heterogeneous sensing infrastructure [5] to capture the complex thermodynamic phenomena of heat generation and extraction at various regions inside the machine room. The sensing infrastructure comprises of scalar temperature and humidity sensors placed at the server inlet (cold aisle) and outlet (hot aisle), airflow meters at the server outlet, and thermal cameras in the hot aisle.…”

“…Information from the environment layer, comprising of an heterogeneous sensing infrastructure (with scalar temperature and humidity sensors, thermal cameras, and airflow meters) is key to characterize the thermal behavior of a datacenter under a given load (information from the application layer) [5]. As mentioned earlier, the estimation of heat imbalance requires estimation of the heat-generation and heat-extraction rates.…”

Thermal Prediction Models for Virtualized Data center Servers by using Thermal-profiles

“…Thermal monitoring systems [15,16] that relied upon thermal cameras could progress efficiently by detecting the hotspots from a distance, without the assistance of thermal sensors. Only three thermal cameras were required to accurately identify and locate the coordinates of the hotspot.…”

“…The characteristics of hotspots were extracted and propagated over network, instead of sending multimedia images. Thus [15,16] presented a network efficient thermal monitoring systems for data centers. However, thermal cameras had a drawback of huge cost of setup.…”

Discrete control for ensuring consistency between multiple autonomic managers

“…In contrast with [27], which addresses the management of datacenters based on thermal awareness with external sensing infrastructure for energy and cooling efficiency, the work, presented in this paper, focuses on coordinating multiple workload-aware managers to ensure an energy efficiency.…”