Browse Prior Art Database

Streams Timing Characterization Map

IP.com Disclosure Number: IPCOM000243862D
Publication Date: 2015-Oct-22
Document File: 1 page(s) / 20K

Publishing Venue

The IP.com Prior Art Database

Abstract

Disclosed is a timing profile which tracks tuple timings as they move through a graph of operators in a streaming application.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 84% of the total text.

Page 01 of 1

Streams Timing Characterization Map

Given a huge graph of operators in a streaming application, a developer may want to debug a particular error in a deployment of the streams application in an IDE, a test lab, or machine. Or, they may wish to only debug a subset of the streams application.

    The performance of the lab (or subset of a stream app) may be very different than the production environment where the problem occurred. Since many problems are timing dependent, a mechanism is needed to recreate the timing from the production machine where the problem was detected.

    The invention is a timing profile which tracks tuple timings as they move thru the graph. At any time, the profile will keep track of timing of different streams, operators, and/or tuples. Example: Source X producing 1000 tuples per second, and Operator 1 processes 500 tuples per second, etc. or tuple y sitting in buffer B for 5 seconds, etc. Thus, the timing profile will move tuples approximately as they would be in the live environment (example: only producing X tuples per minute from some host, etc.). Additionally, in a more complex embodiment, a rolling buffer keeps the actual tuples for the last time frame (example 30 seconds) so that if an error is encountered, a snapshot of that rolling buffer can be saved. Thus, it can be played back at a later time to debug problems with the actual timing from the production environment.

    The same technology can be used for simulating and/or sizing stream...