ArcelorMittal Gent gains control over process workflow with IBM Tivoli Workload Scheduler

Published on 23-Dec-2010

Validated on 09 Dec 2013

"With the process visibility provided by Tivoli Workload Scheduler, we developed a much better understanding of processes and workflows. With this knowledge, we began to think about what could be changed to optimize processes and operations. The Tivoli Workload Scheduler interface made it easy to change processes and workflows, and the resulting optimization improved customer services and job satisfaction for the staff." - Ann Muys, SAP Systems Specialist, ArcelorMittal Gent

Customer:
ArcelorMittal Gent

Industry:
Industrial Products

Deployment country:
Belgium

Solution:
BA - Business Intelligence, Enabling Business Flexibility, Enterprise Resource Planning, Information Infrastructure, Information Integration, Infrastructure Simplification, Optimizing IT, Optimizing IT

IBM Business Partner:
SAP

Overview

This paper describes how ArcelorMittal Gent is driving its manufacturing and business systems more effectively by introducing IBM Tivoli Workload Scheduler to control very large numbers of highly variable processing tasks.

Business need:
Customer objectives included: Intercept, optimize and manage submitted workload containing essential commercial data from the operations units. Keep financial data up-to-date for executives. Automate workflow control over both SAP and third-party applications. Enable an overview of processes across multiple source and destination systems. Cut the costs of managing a highly variable processing workload. Meet agreed service levels for delivery of business information. Increase business efficiency and achieve operational targets.

Solution:
Deployed IBM Tivoli Workload Scheduler 8.5.1, IBM Tivoli Workload Scheduler for Applications 8.5, four IBM Power 570 servers each with 16 POWER5 processors, four IBM Power 550 servers each with eight POWER6 processors and IBM System Storage DS8100.

Benefits:
Benefits include: Essential reporting and online dashboards for monitoring and control can be achieved by scheduling workload effectively. Business service levels can be achieved by exploiting existing capacity to the maximum with Tivoli Workload Scheduler, avoiding unnecessary investment in hardware capacity. Variant handling allows dynamic modification of jobs and priorities. Flexible priority classes for scheduling and coordination. Standardizing to Tivoli Workload Scheduler provides a shared interface that is easier to use and more productive to operate.

Case Study

About this paper

This paper describes how ArcelorMittal Gent is driving its manufacturing and business systems more effectively by introducing IBM Tivoli Workload Scheduler to control very large numbers of highly variable processing tasks.

Customer Objectives

  • Intercept, optimize and manage submitted workload containing essential commercial data from operations units
  • Ensure business-critical reports due at 8.00am are delivered on time every day
  • Keep financial data up-to-date for executives
  • Automate workflow control over both SAP and third-party applications
  • Enable an overview of processes across multiple source and destination systems
  • Cut the costs of managing a highly variable processing workload
  • Meet agreed service levels for delivery of business information
  • Increase business efficiency and achieve operational targets.

IBM Solution
  • IBM Tivoli Workload Scheduler 8.5.1
  • IBM Tivoli Workload Scheduler for Applications 8.5
  • Four IBM Power 570 servers each with 16 POWER5 processors
  • Four IBM Power 550 servers each with eight POWER6 processors
  • IBM System Storage DS8100.

Customer Benefits
  • Essential reporting and online dashboards for monitoring and control can be achieved by scheduling workload effectively
  • Business service levels can be achieved by exploiting existing capacity to the maximum with Tivoli Workload Scheduler, avoiding unnecessary investment in hardware capacity
  • Job interception allows tasks to be prioritized according to business rules
  • Job throttling helps to improve the SAP scheduling efficiency and reduce the batch window for SAP jobs to a minimum.
  • Variant handling allows dynamic modification of jobs and priorities when required
  • Automation and scheduling of event creation to start and stop jobs
  • Flexible and detailed priority classes for scheduling and coordination
  • Ability to connect with multiple applications on different platforms
  • Standardizing to Tivoli Workload Scheduler reduces the number of products required to connect SAP and other applications, and provides a shared interface that is easier to use and more productive to operate
  • Reviewing and optimizing business workflows and processes during the migration to Tivoli Workload Scheduler itself has a beneficial effect on the design of the underlying business processes
  • Tivoli Workflow Scheduler makes it easier to optimize code and workflow processes for new business executive requests, by identifying dependencies and business priorities as part of the wider enterprise workload.
  • Detailed feedback on performance allows developers to optimize existing workload using current processor, memory and bandwidth resources, rather than investing in additional hardware
  • Integration with IBM Tivoli Workload Manager offers the prospect of managing workload and data storage from a shared interface across the enterprise.

Background, starting point and objectives

About ArcelorMittal Gent
ArcelorMittal is one of the world’s leading steel companies, with operations in more than 60 countries, serving the automotive, construction, household appliances and packaging markets. In 2009, ArcelorMittal had revenues of $65.1 billion and crude steel production of 73.2 million tonnes, representing approximately 8 percent of world steel output.

ArcelorMittal Gent in Belgium specializes in developing innovative steel products, production processes, finishes and coatings, using sophisticated mathematical models combined with advanced production facilities. The software systems play a crucial role in controlling the product quality and in the efficient operation of the production process, which includes the logistics and production scheduling.

Some 15 percent of ArcelorMittal Gent’s research and development investment is directed towards environmentally-friendly measures. For example, each cubic meter of water is reused 25 times to help minimize environmental impact.

Business challenges and project objectives
ArcelorMittal Gent runs a somewhat federated business model. The individual plants are responsible for production planning, quality and operations, while sales orders, finances and administration are carried out centrally.

Details of sales orders, purchases and the related operational data are transferred from local systems to the central SAP applications as a series of files for processing. The local systems are a mix of SAP, custom, legacy and third-party solutions. On some days, more than 2,000 files might arrive, containing valuable sales, financial, quality and similar data, destined to be analyzed and imported into the core SAP systems.

To take invoice generation as an example, using the data from the production plants, the central SAP applications produce daily invoice runs. If the file processing is not completed in time, then sales will be missed from an invoice run, which in turn could delay payment. Because most customers have specific payment terms based on calendar months, a delay of a couple of days in producing invoices could have significant knock-on effects in terms of payment date and cashflow.

The commercial data in the files is extracted overnight from the local systems, consolidated centrally, and provided to managers and executives as standard reports by 8.00am the next day. The reports contain historic as well as current data, which helps management make informed decisions on likely demand and compare performance against forecast.

As the files arrived, they were added to the appropriate queue for processing on an ad-hoc basis. There was no way to predict the arrival times of the files, and no way to balance the demands of importing the new data against the requirements to continue with day-to-day processing needs and provide acceptable system performance.

To minimize the impact on production systems, the ArcelorMittal Gent team analyzed the total workload, and intervened to halt or enable processing queues. Typically it would make sense to complete smaller and faster jobs first, depending on the urgency and importance of the task and whether it affected the critical management reports. The systems administration work was gradually rising, and occasionally manual errors were introduced. Without an automated solution, valuable time was being lost, and business users were not being well served.

ArcelorMittal Gent had deployed a mix of monitoring and scheduling tools to automate as much of the work as possible, but with multiple interfaces and different connectors to the underlying applications, this approach soon became part of the same problem. Changing process workflows was slow and difficult, and there was no single view of end-to-end processes spanning several systems. IT staff productivity was impacted as they tried to discover and rectify root causes of issues – for example where data had not been delivered to a dependent system – and executives were frustrated at the sometimes slow and irregular delivery of reports.

Without centralized control of the workflow, the IT team found themselves at the mercy of the users. If a user sent a stream of very large files for processing, this new workload could significantly impact, and in some cases even halt, production tasks, as the initial task might spawn multiple sub-tasks that could not then be interrupted.

The ArcelorMittal Gent team wanted to:

  • Intercept, optimize and manage submitted workload
  • Automate workflow control over both SAP and third-party applications
  • Enable an overview of processes across multiple source and destination systems
  • Cut the costs of managing a highly variable processing workload
  • Meet agreed service levels for delivery of business information
  • Increase business efficiency and achieve operational targets.

The SAP applications at ArcelorMittal Gent are running on IBM Power 570 servers with IBM POWER5 processors and on IBM Power 550 servers with IBM POWER6 processors, with IBM System Storage DS8100 for data storage. The Gent location runs some 80 SAP instances, supporting up to 4,000 users for ArcelorMittal in Western Europe.

Initial IT environment
In many cases, independent manufacturing systems must complete specific tasks before the next step in the process becomes possible, and information generated from each area must be forwarded reliably and rapidly from application to application. During production, the systems create valuable data, for example on sales order progress, inventory and raw materials requirements. ArcelorMittal Gent uses SAP ERP applications and SAP NetWeaver Business Warehouse to capture and analyze this information, and combine it with financial, logistics and other operational data, to provide managers with a complete picture of business performance.

The existing environment relied on the scheduling capabilities of the SAP software, extensive scripting, calendar-driven programs and basic application scheduling. This mixture of in-house, third-party and SAP event-driven process control was difficult to manage, and required constant attention from the IT team.

The IBM solution


ArcelorMittal Gent prepared a detailed RFP document, providing examples of the processing workload in terms of numbers of files, arrival times and the target schedules. The RFP included analysis of the existing systems and the dependencies of the SAP and third-party applications.

Because a sales order for a specific product might be fulfilled from a manufacturing plant many thousands of miles away from the customer, all the ArcelorMittal Gent divisions exchange data to a greater or lesser extent, and all the operations report their financial and administrative data to the headquarters location. Although the RFP applied to ArcelorMittal in Gent in Belgium and to Dunkirk in France, these two were chosen as Proof of Concept sites with a view to extending a successful solution globally.

The objective was to provide a global view of all job schedules and status, and identify dependencies of the SAP applications awaiting data from external applications, and the dependencies of the third-party software awaiting data from SAP applications.

ArcelorMittal Gent selected IBM Tivoli Workload Scheduler to act as the single point of control for all its workload scheduling, spanning SAP and external applications and across the IBM Power Systems platform, mainframe and distributed systems.

IBM created a pilot solution that would demonstrate the capabilities offered by IBM Tivoli Workload Scheduler, covering approximately 10 percent of the jobs submitted. On this basis, the ArcelorMittal Gent team took the opportunity to plan the larger implementation to best advantage. In particular, the ArcelorMittal Gent looked at specific workflow and scheduling issues and mapped them to the Tivoli Workload Scheduler solution and discovered how the challenges could be resolved.

ArcelorMittal Gent selected four IBM Power 570 servers each with 16 POWER5 processors, and four IBM Power 550 servers each with eight POWER6 processors. The Power servers are both divided into logical partitions (LPARs), sized to provide the most efficient combination of processor, memory and bandwidth to support the applications in each LPAR. The system shares the total resources across the LPARs, assigning greater capacity during a particular LPAR’s peak workload and then reassigning that capacity elsewhere as demand shifts. This allows ArcelorMittal Gent to maximize the workload throughput with the most cost-efficient system design.

Technical Solution in detail


ArcelorMittal Gent is using multiple Tivoli Workload Scheduler SAP scheduling functionalities to monitor and manage SAP jobs, with Tivoli Workload Scheduler 8.5.1 and Tivoli Workload Scheduler for Applications 8.5.

The key deciding functionalities in favor of IBM Tivoli Workload Scheduler were:

  • Job interception, which allows tasks to be prioritized according to business rules
  • Job throttling, which helps to improve the SAP scheduling efficiency and reduce the batch window for SAP jobs to a minimum
  • Variant handling allows dynamic modification of jobs and priorities when required
  • Automation and scheduling of event creation to start and stop jobs
  • Flexible and detailed priority classes for scheduling and coordination
  • Scheduled reporting and online dashboards for monitoring and control
  • Ability to connect with multiple applications on different platforms.


Job Interception
For ArcelorMittal Gent, the job interception capabilities of Tivoli Workload Scheduler have produced the greatest transformative effects. All tasks arrive at Tivoli Workload Scheduler, which intercepts potentially disruptive batch processes. Each job is assigned a priority according to pre-set policies based on the nature of the job, which ensures that workload across the entire system is processed according to business rules rather than submitted time.

Job Interception is a feature of the SAP BC-XBP 2.0 interface. It enables Tivoli Workload Scheduler to have a very sophisticated control over the jobs launched by SAP users from the SAP graphical interface.

The job interception mechanism becomes active when the SAP job scheduler is about to start an SAP job, and examines the start conditions. It checks the job parameters (job name, creator, client) against the entries in the SAP table TBCICPT1, and when the job parameters match the criteria, the SAP job is set back to the scheduled status and is marked with a special flag, denoting that the job has been intercepted.

If Tivoli Workload Scheduler has been set up to handle job interception, it periodically runs its own job to retrieve a list of intercepted jobs and reschedules them. This job can be referred to as the interception collector job.

To enable the Job Interception feature, run ABAP report INITXBP2. This report shows you the current status of the job interception and parent-child features, and allows you to toggle the status of both features.

Because intercepted jobs remain in the scheduled status until they are relaunched, you need to define a Tivoli Workload Scheduler job that uses the SAP interception collector task to collect and restart them.

Job Throttling
The job throttling feature will help ArcelorMittal Gent to improve the efficiency of scheduling on SAP systems and reduce the batch window for SAP jobs to a minimum. Using advanced XBP 2.0 and 3.0 functions, such as the job interception and parent-child, the job throttler function ensures that the SAP system is not overloaded and the number of released jobs does not exceed the total number of SAP background work processes in the system.

Job Throttler is a Java tool that uses JCo (Java SAP Connector), enquiring intercepted jobs and releasing them when the background work processes that they need on the SAP server or SAP server group are available. The queue of intercepted jobs is sorted by scheduling time and priority. When the SAP parent-child feature is enabled, child jobs inherit their progenitor’s priority so that new urgent jobs are run before other planned jobs.

The most interesting use of this feature is that the Job Throttler can be configured to have the intercepted job inherit the priority class from its progenitor (the top-level job in the hierarchy), if the progenitor class is higher than the intercepted job class. This setting automatically enables the parent-child feature on the SAP system.

SAP Job Variants
Tivoli Workload Scheduler allows the team to handle job variants on-the-fly, enabling changed business priorities to be reflected immediately. For example, the names of inbound files sent by other users do not necessarily follow the agreed conventions, and the existing priorities or rules for that job may not meet current needs.

Tivoli Workload Scheduler enables the team to change filenames and over-rule priorities and variants for a single or group of jobs, without having to rewrite each individual job and ensuring that the correct conditions are passed to the SAP applications.

The priority classes are based on operational roles of the submitter, workload type and forecast total run time, offering a very great deal of fine-grained control. The policies take account of the business priorities, and Tivoli Workload Scheduler enables intelligent scheduling of workload to ensure that user requests are met as promptly as possible.

For example, a low-priority job that is known to run quickly can be scheduled before a high-priority job that will take several hours to complete, where there is sufficient time to ensure that both jobs will be finished within agreed service levels. The Tivoli Workload Scheduler Interface allows ArcelorMittal to create/modify/delete standard SAP Jobs. This includes the functionality to manage the variant definition of SAP jobs. For each job variant details are available.

Dynamically Created Jobs
ArcelorMittal Gent will be looking at the use of dynamically created jobs. At submission time, Tivoli Workload Scheduler can create “on-the-fly” a new background job on SAP and submit it. The dynamically created background job is used for that submission only. The Tivoli Workload Scheduler job definition contains all the information needed to create an SAP background job (such as the steps, print parameters). The information is stored in the Tivoli Workload Scheduler database, and at the same time it is possible overwrite a template SAP job from the SAP database with new values from Tivoli Workload Scheduler.

Event Handling
To monitor system and job conditions, Tivoli Workload Scheduler delivers scheduled reports and online dashboards for IT team and business unit use. The reports and dashboards include the ability to drill down to the underlying data, which enables teams to identify the processes and dependencies of each job, should an issue arise.

Job Dependencies are prerequisites that must be satisfied before a job or job stream can start. Internetwork dependencies are dependencies checked by the extended agent workstation to which they belong. In response to an internetwork dependency, the SAP extended agent checks for the occurrence of the SAP background event specified in the dependency. As soon as the SAP event is raised, the SAP extended agent commits the event and instructs Tivoli Workload Scheduler to resolve the corresponding internetwork dependency.

Only the SAP background events stored in the SAP event history table are considered by Tivoli Workload Scheduler to check for internetwork dependencies resolution. As a prerequisite, the SAP administrator must create the appropriate event history profiles and criteria on the target SAP system.

Tivoli Workload Scheduler event handling allows single point of monitoring and control for batch workload and batch environments.

  • Monitor workload jobs events (e.g. job start/end) and alerts (i.e. job in error, late, long)
  • Monitor events related to scheduling SAP application and infrastructure health
  • Monitoring of SAP business critical jobs, critical path, critical predecessors, risk level.

Tivoli Workload Scheduler provides automation and scheduling for both SAP applications and external work requests. Tivoli Workload Scheduler comes with pre-built connectors for an extensive range of applications, removing the need to write custom software each time the ArcelorMittal Gent team wishes to integrate a new business acquisition. This significantly reduces the in-house IT developer workload, and provides a common, central interface for all workload scheduling tasks.

Reporting and analysis
Tivoli Workload Scheduler includes reporting and analysis tools capable of monitoring the entire job workload. As part of production processes, reports are generated which show summary or detail information about the previous or next production day. These reports can also be generated ad-hoc.

For ArcelorMittal Gent, these reports show job statistics and production summaries for operational and business functions to allow the team to compare past performance and adjust future workload.

Central planning and calendar
Another important Tivoli Workload Scheduler option used by ArcelorMittal Gent is the central planning and calendar functionality.

A calendar is a list of scheduling dates. Each calendar can be assigned to multiple job streams. Assigning a calendar to a job stream causes that job stream to run on the dates specified in the calendar. A calendar can be used as an inclusive or as an exclusive run cycle, which helps ArcelorMittal Gent to align the workload to local calendar needs and dependencies.

More Tivoli– Tivoli Storage Manager and Tivoli System Automation
Alongside the Tivoli Workload Scheduling solution, ArcelorMittal Gent uses Tivoli Storage Manager and Tivoli System Automation software to help manage its wider application environment.

ArcelorMittal Gent uses IBM Tivoli Storage Manager, with its wide range of storage management capabilities, to address business continuity, by helping to shorten backup and recovery times and helping to maximize application availability with advanced data recovery management technologies.

Tivoli Storage Manager helps to control storage management costs with ease-of-use features, and to increase visibility into the data protection environment by providing advanced features for operational monitoring and historical reporting.

To improve the availability of the SAP database servers, ArcelorMittal Gent uses Tivoli Storage Manager for Enterprise Resource Planning (ERP), which reduces the SAP database administration workload through automated data protection designed for SAP environments.

Tivoli Storage Manager for ERP is an SAP Certified solution, supporting large-volume data backups and recovery, data cloning and disaster recovery of multiple SAP database servers.

ArcelorMittal Gent uses Tivoli System Automation, which is mainly employed to initiate, execute and coordinate the automated starting, stopping, restarting and failing-over of individual SAP application components or entire composite application landscapes.

Taking advantage of its advanced, policy-based automation capabilities, ArcelorMittal Gent uses Tivoli System Automation to ease the operational management of its complex infrastructure. Using the Tivoli solution, the system administrator can more effectively manage planned outages and protect against operational errors. This reduces costs by automating manual, labor-intensive and error-prone operational tasks.

Though its use of Tivoli System Automation, ArcelorMittal Gent has achieved:

  • easier management of its complex IT infrastructure
  • reduced skill requirements
  • flexible, advanced automation functionality.


Project achievements

Performance improvements
During the Proof of Concept stage, mapping business processes to Tivoli Workload Scheduler highlighted many areas where dependencies had not been clearly defined, and priorities had not been set. Migrating the workload to Tivoli Workload Scheduler allowed the team to build a complete picture of the work processes, managed from a single place, which in turn suggested ways to improve the underlying business processes.

For example, when business executives asked for new processes or reports, using Tivoli Workload Scheduler made it easier for the developers to identify dependencies and schedule the work correctly. This ensures that jobs do not start and subsequently wait for input, and run – as far as possible – correctly, first time. This reduces the total number of jobs running at a particular moment, and reduces the drain on system resources. In addition, with Tivoli Workload Scheduler, developers are now able to examine the wait time and run time for each job, and compare historic performance data.
Where a job waits longer or runs more slowly than previously, the developer uses Tivoli Workload Scheduler to examine root causes and review the job coding, dependencies and priority.

Without this detailed feedback, ArcelorMittal Gent had been forced into addressing workload issues – such as delayed reporting or poor performance – by investing in processor or memory capacity. Tivoli Workload Scheduler allows the team to identify the exact workload challenge and consider re-scheduling work or optimizing processes before investing in hardware.

Benefits and next steps
The ArcelorMittal Gent implementation has demonstrated how the entire ArcelorMittal Group could benefit from integrating and optimizing workload across both SAP and other applications. The current Tivoli Workload Scheduler deployment will be extended to encompass all of the SAP tasks and other applications, providing enterprise-wide control while enhancing the effectiveness of the SAP landscape.

For ArcelorMittal Gent, a worldwide enterprise that has grown in part through acquisition and merger, Tivoli Workload Scheduler offers the opportunity to share a common interface in all the data centers. The lessons learned from ArcelorMittal Gent are:

  • Standardizing to Tivoli Workload Scheduler reduces the number of products required to connect SAP and other applications, and provides a shared interface that is easier to use and more productive to operate
  • Reviewing and optimizing business workflows and processes during the migration to Tivoli Workload Scheduler itself has a beneficial effect on the design of the underlying business processes
  • Tivoli Workflow Scheduler makes it easier to optimize code and workflow processes for new business executive requests, by identifying dependencies and business priorities as part of the wider enterprise workload
  • Detailed feedback on performance allows developers to optimize existing workload using current processor, memory and bandwidth resources, rather than investing in additional hardware
  • Integration with IBM Tivoli Workload Manager offers the prospect of managing workload and data storage from a shared interface across the enterprise.

Products and services used

IBM products and services that were used in this case study.

Hardware:
Power 550 Express, Power 570, Power Systems, Storage: DS8100

Software:
Tivoli Storage Manager, Tivoli Workload Scheduler, Tivoli Workload Scheduler for Applications, Tivoli Storage Productivity Center for Replication Basic Edition for System Z

Service:
IBM-SAP Alliance

Legal Information

© Copyright IBM Corp. 2010 All Rights Reserved. IBM Deutschland GmbH D-70548 Stuttgart ibm.com Produced in Germany IBM, the IBM logo, ibm.com, i5/OS, DB2, Domino, FlashCopy, Lotus, Notes, POWER, POWER4, POWER5, POWER6, System i, System x, and Tivoli are trademarks of International Business Machines Corporation in the United States, other countries, or both. If these and other IBM trademarked terms are marked on their first occurrence in this information with a trademark symbol (® or ™), these symbols indicate U.S. registered or common law trademarks owned by IBM at the time this information was published. Such trademarks may also be registered or common law trademarks in other countries. A current list of other IBM trademarks is available on the Web at: http://www.ibm.com/legal/copytrade.shtml UNIX is a registered trademark of The Open Group in the United States and other countries. Linux is a trademark of Linus Torvalds in the United States, other countries, or both. Microsoft, Windows, Windows NT, and the Windows logo are trademarks of Microsoft Corporation in the United States, other countries, or both. Other company, product or service names may be trademarks, or service marks of others. This brochure illustrates how IBM customers may be using IBM and/or IBM Business Partner technologies/services. Many factors have contributed to the results and benefits described. IBM does not guarantee comparable results. All information contained herein was provided by the featured customer/s and/or IBM Business Partner/s. IBM does not attest to its accuracy. All customer examples cited represent how some customers have used IBM products and the results they may have achieved. Actual environmental costs and performance characteristics will vary depending on individual customer configurations and conditions. This publication is for general guidance only. Photographs may show design models.