Browse Prior Art Database

Method Of Providing On-Demand-Computing For Server Blades Disclosure Number: IPCOM000019249D
Original Publication Date: 2003-Sep-08
Included in the Prior Art Database: 2003-Sep-08
Document File: 1 page(s) / 44K

Publishing Venue



The method disclosed here is to enable IBM to provide on-demand computing to customers have installed IBM BladeCenter processing blades and therefore bill customers based on their actual usage of the blades. The on-demand billing method is described below.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 53% of the total text.

Page 1 of 1

Method Of Providing On-Demand-Computing For Server Blades

       Customers install the desired number of BladeCenter chassis(s) that are fully populated with the maximum number of processor blades. Each blade is installed with a customer desired applications. A customers signs up for a specific Quality of Service (QoS) level, for example, number of hits per day of a web server, to start. Other methods for billing besides hits can also be used. A management agent would monitor the system for CPU utilization, Network bandwidth and consumption, memory usage, response time to web requests, etc . The collected statistical data would be available for use in planning for the future computing needs of the customer. Based on the selected QoS number and upon the purchase policies chosen by the customer, the Management Module (MM), which is controlled by a higher level management software such as IBM Director, decides when to turn on/off the individual blades basing on computing demands. The MM also keeps track of actual usage of blades for accounting and billing purposes. This type of monitoring by the MM would also allow for workload balancing among the blades for the different applications/services to provide better utilization of computing resources.

     The MM will compute the number of blades to be turned on in order to provide the selected service. The MM also keeps a record of the computing consumption. If the consumption is going beyond the customer's prediction, the MM will...