Architectural Automation with ActiveBatch
We had the pleasure this week of speaking with a long standing customer for an update about their implementation. Turkiye Finans has been an ActiveBatch user since 2009 and Mucahit Yavuz, IT Operations Manager at Turkiye Finans, has been part of the ActiveBatch project from the beginning.
The conversation underscored a number of points, but most importantly, Turkiye Finans’ implementation of ActiveBatch highlights a strategy we mentioned in our previous post: taking a phased approach where singular processes were identified and automated, then complimentary processes and technologies have been added to the implementation, building an “architectural” IT automation solution.
For Turkiye Finans, it started with a common pain point: a collection of disparate, “point” scheduling tools and scripts, in this case an assortment of PowerShell scripts executed via Windows Task Scheduler and SQL Server Agent, .NET Assemblies and Web Services. This loosely cobbled together collection of tools represented the foundation for the automation of many of the bank’s key business processes, including updating of customer accounts, bill payments and transaction processes.
Defining and executing scripts via Task Scheduler and SQL Server Agent meant scheduling jobs across individual servers. If a script failed to run, finding the machine it was located on was half the battle, and once it was located, security credentials had to be entered each time. Using SQL Server Agent also meant the inability to trigger jobs based on an IT event, such as a file being downloaded or a customer email being received.
The result was an IT operations team that spent more time putting out fires and managing existing processes via handshakes and flag waving then concentrating on automating and optimizing new processes for the benefit of the IT organization and business as a whole.
Since consolidating multiple schedulers, the IT operations team has driven governance and visibility, including receiving an automated alert/incident within System Center Operations Manager when a job fails, as opposed to the business calling the IT Helpdesk when a customer-facing application hasn’t been updated. They’ve also eliminated security loopholes, including removing the need to enter SQL Server security credentials each time by simply associating a specific user with his or her security credentials the first, and therefore, only time.
Less than a year later, the implementation has expanded and now includes the datacenter to manage a rapidly expanding environment of virtual machines, including VMware and Microsoft Hyper-V. The goal was to optimize the execution of the business workflows already executed within ActiveBatch to the provisioning of the underlying resources within the datacenter. Doing so has allowed the IT organization to balance the execution of business workflows versus “the runbook and datacenter house-keeping tasks,” therefore increasing the overall availability of virtual resources. Most importantly, using a single, unified solution has virtually eliminated the failure of a standalone job scheduled via SQL Server Agent or Task Scheduler because the server it resides on was down for maintenance.
There’s much more to this story, including more on Turkiye Finans’ use of System Center Operations Manager and IT organization’s heavy reliance on Web Service and .NET Assemblies…too much to go into length here. To read the complete story, feel free to download the case study.