Adaptive dynamic buffering system for power management in server clusters

Abstract

A system, method and computer program product for managing power consumption in a system comprising at least two computing devices. The at least two computing devices are organized as part of active computing device group or a ready computing device group, the ready computing devices configured into various modes of readiness including a standby mode, a hibernating mode, and a power-off mode. A control device receives obtained sensed parameters indicative of the processing load at the group of active computing devices and determines a “Readiness” index value representing a state of system readiness which is a function of a transition time parameter representing a time for moving one or more computing devices from the second group of ready computing devices to the first group of active devices in response to a service level violation. To achieve maximum power saving and yet meet various performance targets, the ready computing devices are dynamically allocated from amongst three states: standby, hibernating or power-off, based on the current readiness index value.

Description

RELATED APPLICATIONS[0001]This application is a continuation of U.S. Ser. No. 12 / 013,952, filed Jan. 14, 2008, the entire contents of which are incorporated herein by reference.BACKGROUND[0002]1. Field of the Invention[0003]The invention relates generally to power management techniques for computing systems, generally, and more particularly, the invention relates to effective and efficient and, a dynamic server buffering system for optimizing power consumption in servers while adhering to specified Service Level Agreements (SLAs).[0004]2. Description of the Related Art[0005]The cost of data center server power consumption and the cooling systems to dissipate the generated heat are major expenses in modern data centers where thousands of servers are densely packed in relatively small racks. To maintain effective operation and sustain profitability, it becomes necessary to have power management systems to optimize the power usage with respect to customer requirements. In other words, ...

AI Technical Summary

Benefits of technology

The present invention provides a dynamic buffering system that includes a set of servers as a workload buffer in various stages of readiness to absorb sudden increased of traffic without frequent power on / off of servers. The system includes a control device that determines a system transition time for moving computing devices from one group to another, based on the state of processing operations, and dynamically allocates the computing devices in the second group of ready computing devices in various stages of stand-by, hibernating, and power-off readiness, to minimize the time it takes to bring a computing device from the second group to an active readiness state in the first group while maximizing energy savings in the system. The technical effects of the invention include improved system readiness, reduced power consumption, and improved energy efficiency.

Problems solved by technology

The cost of data center server power consumption and the cooling systems to dissipate the generated heat are major expenses in modern data centers where thousands of servers are densely packed in relatively small racks.

It is well established that a typical server consumes a relatively high amount of power even it is idle, due to chip leakage current and other supporting components such as disk drive and network gears.

The amount of effort for the system administrators, and the priority of data center to satisfy customer requirement over the cost of power consumption negate the effective use and benefit of these power management systems.

It is obvious that there is a finite cost of shutting down a system and turning it back on.

The amount of time to power down / up and restore active applications means lost time for real work and usually maximum power is used to power on / off a server, not to mention the frequent repeated power on / off cycle increases the wear-and-tear of these servers, resulting in higher failure rate.

Due to the time lapsed in restoring a server from shut-down state to fully available for use, usually from 1 to 5 minutes, depending on system applications and configurations, power on a shut-down server may not react fast enough to the sudden increased traffic demand, resulting in frequent violation of customer SLAs.

All these potential problems prevent this power management strategies from being widely used.

Images