Browse Prior Art Database

Method to Predictive Avoid Application Crash Due to Out of Resource Disclosure Number: IPCOM000234874D
Publication Date: 2014-Feb-12
Document File: 5 page(s) / 57K

Publishing Venue

The Prior Art Database


Some monitor mechanism is introduced and it will predict the out of resource cases like out of memory, or reach max file handlers, then it allows user to configure some corrective action to avoid the application crash. The invention will have monitor components which will monitor kinds of resource usage like file handlers, memory and so on, an analysis component which will analysis the cause for resource occupied like whether a single thread occupied most of resources, whether there is a infinite loop, whether the infinite loop is caused by a bug, and action taken component which will execute some actions configured by end user like running the application stop command to soft stop the application, break out of infinite loop as it is caused by bug and so on. With this invention, application and software robust is improved and can continue work even there is some bugs in the error path. And the mechanism can be used by a single application or used by an operating system.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 55% of the total text.

Page 01 of 5

Method to Predictive Avoid Application Crash Due to Out of Resource

In this invention, there is a monitor component, it will monitor the resources used by each component and here the system file handler is taken as the example.

There is one table for each resource, and it will record the resources used per thread of each process.

Monitor components will get all the available resources and the threshold configured by the end user when it is initialized .

When a thread request some resources, systems will allocation the required resources, then the counter of allocation will be checked for the thread to see whether it's the time to check whether the resources occupied exceed the threshold, and if yes, the monitor component will check whether the thread occupied the too much resources and need to kick off the analysis components.

The allocation counter is introduced to for threshold check is for performance improvement, and if the application or systems always have big resources allocated at one time, the number can be set as a small number like 2. And if the application or systems always allocate the resources in small number, the counter can be set as a bigger number like 20. Control will give back to the thread until the resource allocation check is completed.

And a thread repeated request resource is not a concern here as when a thread is detected that occupied lots of resources , the analysis model is started to analysis the thread and the thread is put in sus...