Browse Prior Art Database

Method and apparatus for efficienttly calculating end-to-end response time across transaction paths that include asynchronous interactions.

IP.com Disclosure Number: IPCOM000248101D
Publication Date: 2016-Oct-26
Document File: 8 page(s) / 102K

Publishing Venue

The IP.com Prior Art Database

Abstract

Transaction tracking technologies focus on tracking composite applications across multiple technologies, protocols, domains (middleware stacks) and operating systems. Tracking is often achieved by instrumenting targeted software with tracking agents which generate tracking events at strategic points in the application flow. Collected tracking events can be analysed to determine application metrics and topology. One of the common requirements for customers using transaction tracking technologies to is the calculation and display of the end-to-end or round-trip time. Customers need to know how long a transaction took in total as this often corresponds to an end user's experience of the system. In transaction paths where interactions between nodes are synchronous the end-to-end time calculation is simple. When the transaction path includes one or more asynchronous interactions the calculation is much more difficult. This invention proposes an efficient method and apparatus for calculating end-to-end response time across transaction paths that include asynchronous interactions, and exposing this response time metric to a user via a graphical user interface for the purpose of transaction response time monitoring.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 37% of the total text.

Page 01 of 8

Method and apparatus for efficienttly calculating end

Method and apparatus for efficienttly calculating end-
--to

to

to-

--end response time across transaction paths that include

end response time across transaction paths that include

end response time across transaction paths that include

asynchronous interactions

asynchronous interactions. .

Transaction tracking technologies focus on tracking composite applications across multiple technologies, protocols, domains (middleware stacks) and operating systems. Tracking is often achieved by instrumenting targeted software with tracking agents which generate tracking events at strategic points in the application flow. Collected tracking events can be analysed to determine application metrics and topology.

One of the challenges for transaction tracking is topology building. Making the necessary associations between requests/responses from an application in one domain with the corresponding requests/responses in the adjacent domain can be difficult. In addition, monitored transactions may traverse multiple domains increasing the complexity of the application topology map.

Typically transaction tracking data from each domain in the transaction path is collected separately and stored/persisted in some kind of cache (document store, relational database or other data store). The data typically contains context (data describing the transaction and data describing the next 'hop' transaction) plus interaction data (data describing the interactions or calls from one domain to the next). Once a complete set of data has been collected for a transaction instance (the complete set of transaction tracking events that describes

a single transaction flow from end to end is named an 'instance') the next step is to correlate the various tracking events that describe the transaction path.

Note: A transaction topology graph is represented as a directed graph with nodes identifying the transaction 'hops' (where the processing took place) and edges showing the links (or interactions) between the nodes. The edges indicate the direction of the interaction (request or response) between the nodes. For example, A ---> B (indicates A calls B), A

B (indicates A calls B and B returns to A).

One of the common requirements for customers using transaction tracking technologies to is the calculation and display of the end -to-end or round-trip time. Customers need to know how long a transaction took in total as this often corresponds to an end user's experience of the system. In transaction paths where interactions between nodes are synchronous the end-to-end time calculation is simple. When the transaction path includes one or more asynchronous interactions the calculation is much more difficult.

This invention proposes an efficient method and apparatus for calculating end-to-end response time across transaction paths that include asynchronous interactions, and exposing this response time metric to a user via a graphica...