Browse Prior Art Database

Series of Methods for Provisioning of Resources

IP.com Disclosure Number: IPCOM000198996D
Publication Date: 2010-Aug-20
Document File: 3 page(s) / 27K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed are methods to optimize the deployment of workloads within a cloud infrastructure such that the least amount of energy is consumed by the data center, lowering the total operating costs of the data center and increasing profit.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 40% of the total text.

Page 1 of 3

Series of Methods for Provisioning of Resources

There are a lot of managed environments where different customers' systems must be installed and run in a shared data center environment. Such a scenario is quite common in the emerging cloud computing space. In the cloud model, customers do not choose on which equipment their workloads are executed. The owner of this data center environment might want to allocate different workloads/data to different resources to minimize the overall cost of running the data center by reducing power usage.

Data Center Power Use

Within the United States, the total cost of powering data center servers represented about 0.6 percent of total electrical use within the country. Factor in the additional costs of cooling and other usage, and that number jumps to 1.2 percent. The total power demand in 2005 (including associated infrastructure) is equivalent (in capacity terms) to about five, 1000 MW [megawatt] power plants for the U.S. and 14 such plants for the world. As the demand for new technology grew, the number of installed, low-end volume servers-typically systems under $25,000, which also includes blades- increased. This trend seems to have driven the skyrocketing energy consumption of the last five years more than the actual energy usage per server. Almost all of this growth
is attributable to growth in the numbers of servers (particularly volume servers), with only a small percentage associated with increases in the power use per unit. (1)

Presently, methods do not exist to optimize workloads to reduce the total power consumption of a data center. Methods are needed to optimize workloads in relation to power consumption.

The proposed invention discloses multiple methods to enhance workload placement within a data center, or cloud center, in order to decrease the amount of energy expended. The disclosed system provisions resources into a system in a manner that offers maximum overall power efficiency. Characteristics of the workload and provisioning of resources is done based on predictive methods and real time monitoring. Provisioning decisions are also based on in-flight transactions, application characteristics, seasonality, etc.

Points of Novelty:

1. A method of managing a computing environment in a manner that reduces overall energy consumption, comprising:

• Monitoring of data characteristics of one or more computing environment resources;

• Using forecasting methods to predict future requirements of data storage and access for each resource;

• Maintaining a database of available resources, including power consumption characteristics of each;

• Using one or more criteria to determine best resource to provision into

1

Page 2 of 3

which computing environment to save on overall power consumption,

• Provisioning additional computing resources according to optimized plan

2. A method of Point 1 further comprising: de-provisioning of resources to remove them from database cluster when not n...