Techniques for monitoring a shared hardware resource

• US 10,489,201 B2
• Filed: 06/09/2015
• Issued: 11/26/2019
• Est. Priority Date: 01/10/2014
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Buendgen, Reinhard, Mayer, Ulrich, Werner, Klaus, Huber, Matthias
• Primary Examiner(s): Lee, Paul D
• Application Number:
  US14/734,554
• Publication Number:
  US 20150268999A1
• Time in Patent Office: 1,631 Days
• Field of Search: 702186
• CPC Class Codes: G01R 29/26 Measuring noise figure; Measuring signal-to-noise ratio
• CPC Class Codes: G06F 11/3003 Monitoring arrangements specially adapted to the computing system or computing system component being monitored
• CPC Class Codes: G06F 11/3409 for performance assessment
• CPC Class Codes: G06F 2009/45591 Monitoring or debugging support
• CPC Class Codes: G06F 21/00 Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
• CPC Class Codes: G06F 21/6245 Protecting personal data, e.g. for financial or medical purposes
• CPC Class Codes: G06F 2201/815 Virtual
• CPC Class Codes: G06F 2221/2135 Metering
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects
• CPC Class Codes: G06F 9/4881 Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
• CPC Class Codes: G06F 9/5011 the resources being hardware resources other than CPUs, Servers and Terminals

Abstract

A technique for monitoring a shared hardware resource in a computer system running at least two partitions includes collecting measurement data from the hardware resource. An aggregation of the measurement data over a certain time interval is computed. A noise level value adapted to a workload is calculated using a workload dependent heuristic. Noise data, based on the calculated noise level value, is generated. The computed aggregation of measurement data is mixed with the generated noise data. The mixed aggregation of measurement data is output to an external interface for monitoring.

Claims

1. A method for monitoring a shared hardware resource, comprising:

collecting, by a data processing system, measurement data from a shared hardware resource that is shared by multiple partitions, wherein the shared hardware resource includes a cryptographic adapter;

computing, by the data processing system, an aggregation of the measurement data over a certain time interval;

using a workload dependent heuristic to calculate, by the data processing system, a noise level value adapted to a workload;

generating, by the data processing system, noise data based on the calculated noise level value;

mixing, by the data processing system, the computed aggregation of the measurement data with the generated noise data to generate monitor data in which details of cryptographic operations performed by a guest using the cryptographic adapter cannot be learned by a hypervisor environment administrator or another guest from the generated monitor data thereby reducing a risk of violating privacy of the partitions that share the hardware resource; and

outputting, by the data processing system, the generated monitor data to an external interface for monitoring, wherein the heuristic generates at least one usage profile for the shared hardware resource comprising a list of measurements measured over a first time interval, and wherein each of the measurements comprises a vector of values and is associated with a specific partition and the vector of values is taken from a simulation or a real computation of the workload over a period of time, and where the heuristic computes for each usage profile;

a first sub-usage profile from the corresponding usage profile by removing measurement data associated with a first partition from the usage profile;

an absolute value of a difference of the aggregation of the measurements in the corresponding usage profile and an aggregation of measurements in the first sub-usage profile; and

a sum of absolute values of vector components representing the difference.

2. The method of claim 1, wherein the measurement data is restricted to data associated with a subset of at least two of the partitions run by the data processing system.

3. The method of claim 1, wherein the heuristic calculates a maximum of approximated noise level values used as noise level value to generate the noise data.

4. The method of claim 1, wherein the heuristic computes a new noise level value while providing measurement data mixed with noise data generated based on an existing noise level value and then replaces the existing noise level value by the new noise level value.

5. The method of claim 4, wherein a first noise level value and a first time interval are replaced by a second noise level value and a second time interval, and wherein the first noise level value is calculated using the first time interval and the second noise level value is calculated using the second time interval which is greater than the first time interval, if the first noise level value is higher than a defined threshold value.