Seagate puts Big Data in action – a case study


Image by:

The following guest blog is by Mike Crump, VP, manufacturing quality and supplier quality engineering/Asia, and Harrie Netel, director, quality data analytics.

Seagate creates and stores staggering amounts of data.

Stored on 650 Seagate enterprise-class hard drives and 42 SSDs in Seagate’s Enterprise Data Warehouse (EDW) is everything from the history of more than 2 billion Seagate drives to manufacturing, customer and supplier information.

The amount of data stored at EDW in Oklahoma City defines “Big Data” — data so large and complex it’s difficult to process using traditional processing applications. Big Data is particularly valuable because more information can be obtained from analyzing a single set of large data than from multiple sets of data with the same information.

Storing data is one thing — and we add 400GB of data to EDW every day  — but taking advantage of it is another. After all, it doesn’t matter how much data you have if you can’t do anything with it.

Last year, the Operations Quality Analytics team took on the challenge of turning Big Data into information that can be acted on to uncover threats, make informed decisions, and drive sound business outcomes and factory quality improvements.

First, the team tackled automated ODT (Outgoing DPPM Test) reporting. Collaborating with IT, the team developed a solution using eCube, a powerful Seagate application that slices and dices enormous amounts of data and completes analyses in hours instead of days or weeks.

The solution — implemented six months ago in Suzhou, Wuxi and Korat — freed up 11 engineers, who had been manually compiling and distributing reports, to spend a lot more time searching for hidden patterns and trends in the data that can indicate potential quality issues.

Besides making it easier to pinpoint problems that might not surface until they are in the field, the automated data analysis process provides higher quality information to factory process engineering teams. It also enables quicker decisions and actions closer to the factory floor.

Automating the ODT process already is helping to improve quality. Collaboration between the Quality Analytics and Operations teams has resulted in issues being uncovered at factory workstations and with components coming from a voice coil motor supplier. “Maverick lots,” or abnormal groups of drives, also have been identified.

The improvements are reflected in Seagate’s overall quality performance. Customers also recognize how we’re using analytics. HP and Dell say we are far ahead of our competition in using Big Data to improve business and quality outcomes.

With the automated ODT reporting and analysis process up and running, the team is focused on linking process and performance information from suppliers with data from field quality systems and factories, including Ongoing Reliability Test (ORT) results.

Linking data “end-to-end” (from supplier and factory to end-user performance) will create a true Big Data system that will help identify key data including supplier parameters most important to product performance. That knowledge will remove a lot of risk from our decision making, push decisions to a lower level and make it easier to adjust quality systems and metrics at our suppliers to ensure we receive parts that meet our expectations.

By monitoring and catching issues “upstream” at our suppliers, rather than “testing for quality at the end of the factory line,” we’ll go a long way toward keeping problems out of the factories and sustaining our quality improvements.

Developing end-to-end analytics also will help lay the groundwork for our vision of defect-free Seagate factories.

Related Posts:

Big data, cloud, and the one terabyte round-trip

There’s Data in them there hills!

The costs of dirty data in the US…more than you think.


About the Author:

One Comment

  1. […] Operations per Second. If you look at how Rackspace and Amazon market cloud compute, or how quickly Big Data can be analyzed, it’s all about […]

Leave A Comment