Dismiss
InnovationQ will be updated on Sunday, Oct. 22, from 10am ET - noon. You may experience brief service interruptions during that time.
Browse Prior Art Database

Method to instrument event trends using 2D sensors

IP.com Disclosure Number: IPCOM000012448D
Publication Date: 2003-May-07
Document File: 4 page(s) / 38K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a method to instrument event trends using two-dimensional (2D) sensors. Benefits include improved functionality and improved reliability.

This text was extracted from a Microsoft Word document.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 54% of the total text.

Method to instrument event trends using 2D sensors

Disclosed is a method to instrument event trends using two-dimensional (2D) sensors. Benefits include improved functionality and improved reliability.

Background

� � � � � The term sensor refers to a sensing device that is implemented using an embedded controller.

        � � � � � For reliability, servers implement threshold sensors to monitor temperature, voltage and other environmental elements. Fatal and non-fatal errors are sensed and recorded, using discrete sensors, such as single-bit and double-bit memory errors. These threshold and discrete sensors record and report the current state. In addition, they may take some action, for example send a simple network management protocol (SNMP) trap when a double-bit error occurs or a temperature exceeds a maximum threshold.

        � � � � � These sensors are unable to detect subtle but complex problems, such as excessive single-bit errors (SBEs) or rapidly rising temperatures.

        � � � � � Conventional implementations detect and report SBEs. However, the excessive occurrences of SBEs can degrade system performance and predict the likelihood of a fatal double-bit error. Ideally, no limit exists for the number of SBE that a system can encounter. If a system is up and running for 6 months without being reset, 100 SBEs are acceptable. However, 100 SBEs in a 5-minute period is unacceptable.

General description

        � � � � � The disclosed method provides a standard mechanism and an algorithm to detect and alert failure trends. This functionality is native to the hardware (HW), enabling it to function when the operating system (OS) is inactive. It can be easily incorporated into the existing intelligent platform management interface (IPMI) sensor architecture.

Advantages

        � � � � � Some implementations of the disclosed structure and method provide one or more of the following advantages:

•        � � � � Improved functionality due to detecting errors that could not be previously detected

•        � � � � Improved functionality due to providing a HW capability not be previously available

•        � � � � Improved reliability due to improved accuracy in the detection of real system problems

•        � � � � Improved cost effectiveness due to eliminating the requirement for an OS software application

Detailed description

        � � � � � � The disclosed method is a 2D sensor. It detects, reports, and takes action on complex system problems, improving system reliability. A 2D sensor enables trend identification for system changes or events. One of the two dimensions is always time. The other dimension depends on the application but typically it is the number of failures or delta of a sensor measurement.

        � � � � � A 2D sensor can be used to detect an unacceptable failure occurrence rate. For example, one SBE per time unit on a system is acceptable. The sensor value is decreased by one each time unit. Over 16 time periods, a number of SBEs are detected (see Figure 1). The sensor with a threshold of 10 is...