Fabric computing hones predictive analysis for commodity steel maker

Rolled steel isn't a typical bastion of advanced IT, but one company is getting the most out of its ERP systems through a fabric computing platform.

Advanced IT systems probably don't come to mind when you consider a competitive advantage in rolled steel pipes, but JMC Steel uses fabric computing architecture to their benefit.

JMC Steel Group, which operates in 13 U.S. steel plants with headquarters in Chicago, replaced a server farm with a fabric computing architecture from Unisys to improve customer and industry tracking.

"There aren't many steel companies that have utilized IT like we have," said William Kempert, director of infrastructure at JMC Steel Group.

JMC Steel runs two independent SAP systems. One manages ordering and production planning and the other schedules for rolling steel that goes to the commodity market and made-to-order customers.

The finance department adopted SAP's profitability analysis suite, CO-PA to bring intelligence to their production and sales timing. Kempert's IT team looked into SAP HANA in-memory computing to execute CO-PA, but found the price point a barrier and investigated the Unisys fabric computing platform.

"HANA's price point can be a question of overall investment or licensing costs -- if you don't already use HANA, it can be a big step to go through vendor evaluations, purchase and implement and train on it," said Carsten Hilker, solution manager at SAP.

The in-memory database infrastructure of HANA means that data analysts don't have to load or write-back data, freeing up time for detailed, actionable analysis.

JMC Steel beta tested a two-node Unisys Forward fabric computing platform with Microsoft Windows 2008 Enterprise Edition to run its mission-critical SAP applications. JMC Steel reported 40% better performance for enterprise resource planning (ERP) than its current servers and far less variability in performance than on a virtualized server environment.

The Forward product uses secure partitioning on Intel Xeon processors to isolate resources for each application workload, with two partitions in each node. The business application exists in one partition and a CPU and I/O burner run in the other, with InfiniBand connections between them. The nodes co-exist on one rack.

CO-PA lets JMC Steel analyze business data by customer, product or other metrics, slicing the data by unit sales, pricing and more. The predictive analysis pays off in steel sales, without the bottlenecks JMC Steel experienced running CO-PA on servers.

CO-PA has many dimensions and HANA compiles everything into one big, deep table directly on the memory chip, rather than in disparate small tables, Hilker said.

For companies that adopt CO-PA without IT budgets to invest in a platform like Forward or HANA, Hilker notes a cloud offering of HANA for companies to run their own data at an hourly rate, without the expense of in-house administrators.

The cloud offering usually fits into the discretionary spending of a finance department, rather than IT, he said.

"The biggest thing is better on-time performance to our customers, especially in the commodity steel market," Kempert said. "Customers need it now, so you have to have enough steel ready to meet their demand today, not three weeks from now."

With advanced planning and optimization (APO) on the ERP platform, JMC Steel also predicts pricing trends to sell its product at the best price. Steel prices vary quite a bit compared to more stable industries, according to Kempert.

IT's goal is to give JMC Steel "push-button information," whether through ERP on the fabric computing platform, for call center support by integrating VoIP, or by collecting and analyzing factory floor data to improve maintenance.

Predicting and preventing downtime on the rolling machines and other equipment is on their roadmap, Kempert said. It starts with a baseline of downtime and maintenance from the factories' Wonderware manufacturing execution system, then integrating MES data with SAP to improve order completion.

In the future, JMC Steel's IT systems for APO will go beyond cost and order planning to help get the best margins on products.

"[Steel] plants aren't clean and sexy," Kempert said, but JMC Steel is applying advanced intelligence to how they work. Companies succeed or fail in the steel industry by how they manage inventory, he added.

This Content Component encountered an error

Pro+

Features

Enjoy the benefits of Pro+ membership, learn more and join.

0 comments

Oldest 

Forgot Password?

No problem! Submit your e-mail address below. We'll send you an email containing your password.

Your password has been sent to:

-ADS BY GOOGLE

SearchWindowsServer

SearchEnterpriseLinux

SearchServerVirtualization

SearchCloudComputing

Close