West Corporation (West) processes “billions of minutes” of voice-related transactions every year. Its extensive line of communications products automate phone and SMS notifications for businesses, medical practices, schools, and public safety agencies. West also provides its thousands of customers with the technology to conduct a wide variety of business transactions through phone and SMS systems.
To ensure reliable delivery of the enormous volume of notifications associated with 15 products, West developed more than 4,000 Windows tasks. In addition to scheduled jobs associated with its core products, the company maintains many additional Windows tasks and Cron jobs to support custom add-ons provided to commercial customers.
Enterprise Administrator Michael Rose, and his IT Operations team managed the schedule of jobs through years of growth, but they were well aware of potential weaknesses in the system. “The lack of a failover mechanism with Windows Task Scheduler could prevent the company from meeting an SLA,” he noted. “Windows tasks were frequently unreliable, and we often never knew if one failed.” Nick Vella, who develops and manages many of the custom processes for customers, added that Cron jobs lack the error handling necessary to quickly recover failed or stalled jobs.
West had employed several workarounds to make sure the master schedule proceeded as reliably as possible. One such workaround was the constant execution of a script that logged into an FTP server every five minutes to check for the presence of a required file before running the next job. The company also used a daily, repeated reminder for the team to log in to each server and scan log files for errors. This manual task pulled the company’s IT staff off higher value projects. Troubleshooting large-scale distributed processes posed an additional challenge because the exact server where a job ran was often unknown.
As the company approached 6,000 jobs across multiple servers, the business recognized the need for a centralized solution.
West initially focused on file transfer applications, but expanded its search process to job scheduling and workload automation once they saw the potential value of centralized cross-platform automation. Rose said, “We looked at several products and selected JAMS based on its simplicity. The thick client offered by [a competing product] was overwhelming and would have required extensive training. JAMS was straightforward and required only about 30 minutes of training.” He added, “We have interns who, without extensive technical training, can immediately leverage JAMS.”
Rose and his team called the implementation “easy” and found the native Windows Task Scheduler conversion utility built into JAMS useful in streamlining the company’s migration. They also praised the JAMS Support team for its continued assistance. He recounted, “When upgrading on a Saturday morning, we ran into an error and called support. We got voice mail. We decided to roll back, but in less than 5 minutes, we got a call back. JAMS Support helped us complete the upgrade. They really went above and beyond.”
Since installing JAMS, West has converted 75% of its batch processing into JAMS.
In addition to freeing West’s IT Operations staff from low-level batch processing tasks, JAMS improved the company’s ability to deliver notifications and custom processes (e.g. specialized reports). Vella said that JAMS contributes directly to customer satisfaction. He said, “Our large commercial clients often have special requirements and requests. These single client solutions are not built into our core platform, so those client customizations execute as batch jobs. Now [with JAMS], if one of those custom processes fails, it can easily be re-run.” He also pointed out the benefit of JAMS parameters in setting up custom code for commercial customers.
JAMS is fully integrated with West’s ticketing system. This enables the IT Operations team to stay well ahead of support issues. “When a customer’s custom process isn’t executing as expected, we know before they do!” said Vella.
Rose pointed out that JAMS gives West the advantage of truly distributed workflows, in which jobs can run in parallel and be triggered by dependencies. Whereas the old system may have kicked off jobs based on the estimated time a file is expected to be present, JAMS kicks off jobs based on the actual event of that file arriving and notifies IT if files don’t arrive when they are supposed to.