Print
PDFTable of Contents
Data is the lifeblood of any organization. With it, companies can understand current issues, analyze trends, and determine actionable results to provide superior products to their customers. Enterprise resource planning (ERP) and customer relationship management (CRM) systems have done this for the business intelligence world for many years. In the test and production arena, R&D departments are realizing that proper test data management can be a competitive differentiator in getting high-quality products to market faster. Three trends are driving this increased need for a data management strategy, and if handled incorrectly, they can make test data a big problem.
Trend 1 – Decreasing Cost of Memory Storage
Fifty years ago, IBM invented the first hard drive, which weighed more than 1,000 lbs, cost $250,000 USD per year to lease, and provided 5 MB of storage capacity. Today, the price of memory is exponentially cheaper. You can find a 500 GB hard drive for less than $100 USD, and consumer electronic devices such as Apple iPods, portable media players, and mobile phones have made it even less expensive to store large amounts of data.
This same memory storage technology has moved into labs and production lines, giving you access to large amounts of inexpensive disk space. Therefore, when making design decisions, you now can choose to store as much data as possible. The problem is that few companies have well-defined processes or strategies for recording and managing large amounts of data.
Figure 1. Consumer electronic devices have made it even easier and less expensive to store large amounts of data.
Trend 2 – Increasing Need to Quickly Analyze and Share Data
Performing automated analysis on your data is crucial to efficiently finding anomalies, trends, and results as well as sharing this information with others. The technical knowledge required for this type of analysis is growing while skilled engineering labor is declining, and compensation rates continue to rise for qualified candidates. More than ever before, you cannot afford to waste time and talent copying and pasting data into Microsoft Excel, manually performing analysis, and building reports with no basis on which to start. Software should be finding trends in the massive amount of data collected and compiling the results into reports automatically. You should be spending your time analyzing these reports and making decisions about design and production based on them versus spending hours just getting the data organized and formatted.
Trend 3 – Growing Lack of IT Expertise on Handling Test Data
For many companies, large system management software or tools are controlled by information technology (IT) departments. However, while the IT department succeeds at ERP systems, CRM systems, and operating system controls, it has little understanding of the data being collected in R&D characterization or production functional testers. Engineers and scientists work with complex waveforms, often consisting of thousands or even millions of data points. Typically, the IT department works with charts and graphs that have tens of data points. The same database used by the sales department is not scalable to work for the R&D department. And many times, R&D does not have the time and/or expertise to create its own data management system.
Hybrid and Off-the-Shelf Solutions for Managing Test Data
As the data problem has emerged, so have the technology and solutions available to help solve the problem. Some solutions take a hybrid approach for storing and managing measurement data. With these hybrid solutions, you can store your data in familiar file formats and organize and manage the data with behind-the-scenes software. For example, NI DIAdem software stores only the meta information from flat files in a behind-the-scenes index, which you can then query for specific property values. You can search on these property values and return the files that match the search results. You can then further analyze and report on the returned values. This type of hybrid approach using both flat files and a data index is ideal when there are large amounts of existing data in files and you want to continue storing the bulk of your data in flat files instead of a database.
Figure 2. DIAdem takes a hybrid approach to data management by storing only meta information in an index for searching.
Another approach is to use off-the-shelf solutions that automatically store all of your test data directly into a database system, so you can directly link to other common postprocessing tools. An example of this is IntraStage, a middleware product that records data from test rack applications – written in NI LabVIEW, NI TestStand, Visual Basic, or C – into a Microsoft SQL server database and provides links to Web-based reports; Excel; Minitab software; DIAdem; The MathWorks, Inc. MATLAB® software; and ERP systems for postprocessing. This type of system abstracts data recording so you can focus on more pressing issues such as the data acquisition and postprocessing components.
Figure 3. IntraStage can help solve your data management problem by automatically collecting, reporting, and analyzing test data.
As the amount of data engineers and scientists can record grows, so do the potential issues that can arise from being overloaded with data. However, vendors are introducing new technologies and products into the marketplace to address these issues. At the same time, the overall goal of these tools is not only to keep you from drowning under all your data, but also to help you gain more information from your data using trending, reporting, and automated analysis. As the three trends outlined above continue to gain momentum, the need for technical data management solutions will continue to grow.
Patrick Kelly is the vice president of strategy at Cal-Bay Systems and has published articles and held seminars for 12 years on trends in the test and measurement industry. He began his career at National Instruments after receiving his bachelor’s degree in artificial intelligence and MBA in finance from Vanderbilt University.
Caroline Bright is the NI DIAdem and data management product manager and has been educating engineers and managers on the benefits of proper data management during her career at NI. She holds a bachelor’s degree in computer engineering from Vanderbilt University.
Watch an overview video on IntraStage.
Learn more about DIAdem by viewing short videos and reading more about data management.
This article first appeared in the October 7, 2008, issue of NI News.
MATLAB® is a registered trademark of The MathWorks, Inc.
Reader Comments | Submit a comment »
Thank you!
Your article on ...Test Data a Big Problem
describes perfectly some of my frustrations
in trying to get my upper management to
understand our R&D computing needs. I will
use some of your points in upcoming
discussions. Thanks!
- Eric Wilson, NAWCWD China lake. anon2091391 - Oct 7, 2008
Legal
This material is protected under the copyright laws of the U.S. and other countries and any uses not in conformity with the copyright laws are prohibited, including but not limited to reproduction, DOWNLOADING, duplication, adaptation and transmission or broadcast by any media, devices or processes.