Ca, Inc.

Computer Associates Think, Inc.

Ramesh Mani

Brian Zuzga

Ling Thio

Martin Tali

Mark Jacob Addleman

Marco Gagliardi

Chitresh Deshpande

Karmjit Singh

In various embodiments, a method of automatically identifying back-end components of a processing system may be performed. The method may include logging socket communications as a plurality of stack traces. The method may further include filtering the plurality of stack traces to identify a transaction from a particular thread, where the transaction includes a write stack trace and a read stack trace. The method may further include identifying, based on an examination of common stack elements between the write stack trace and the read stack trace, a candidate software component as a back-end component that is communicating with an external software component.

Identifying back-end components based on stack traces

Methods for automatically identifying and instrumenting application classes and methods for a particular application are described. In some embodiments, application code (e.g., bytecode or source code) associated with the particular application may be parsed to identify classes and methods within the application code and to identify terminal components (e.g., methods or function calls) and non-terminal components (e.g., control flow statements). Once the terminal components and non-terminal components have been identified, a complexity model and a corresponding score for each of the classes and methods within the application code may be determined. The complexity model may be used to estimate the number of computations that may be required if a particular class or method is used by the particular application. Application classes and methods corresponding with a score that is greater than a threshold may be instrumented by inserting probes into the identified classes and methods.

Automated instrumentation of applications

A benchmark response time is determined for a browser application request sent to a network server over a network. The response time is determined by performance monitoring code that is loaded into and monitors the browser application from the client. The performance monitoring code automatically sends a request to a network server; the request is not sent in response to user input. The network server receives the request, generates a response and provides the response to the browser application. The response includes a fixed amount of randomly generated data. The browser application receives and processes the response, but does not display the bytes or change the content displayed in the browser application as a result of the response. The browser application sends the times at which the browser application sends the request and the browser application completes processing the response data to the network server for further processing.

Response time benchmarking

A technique for monitoring software which efficiently communicates transaction trace data, including static and dynamic data, from an agent to a manager is disclosed. A tree data structure describes transactions as a sequence of transaction elements. Each branch may correspond to one type of transaction. The nodes along a branch may be associated with components of monitored software that are invoked to perform the transaction. A provisional branch may be constructed when an agent is unable to initially identify the transaction. Transaction data may be stored in association with the nodes in the provisional branch during the monitoring. After information needed to identify the transaction becomes available to the agent, the provisional branch is updated. The transaction data is reported in response to determining the branch in the tree data structure that matches the provisional branch.

Data structure for efficiently identifying transactions

Multiple transaction components that comprise a transaction are correlated using a GUID generated at a browser application. The transaction components may occur asynchronously or synchronously between a network browser and one or more applications. An identifier is generated for a set of network browser requests corresponding to a single user input or some other event and the identifier is included in each request. Server traffic and the servers processing the request are monitored and data which includes the identifier is generated. Data for the transaction with multiple transaction components is reported using the GUID associated with the transaction.

Growing community of inventors

Fremont, CA, United States of America

Ramesh Mani