Major Broadcaster moves to Azure

Hearst Television, a major US based broadcasting company, benefits from elevated access to key analytics.
Download as PDF

Hearst’s on premise Analytics solution was comprised of data extracted from their operational traffic system to the Decentrix BIAnalytixTM Media Analytics solution which was hosted on a virtual machine. Data was processed into an Analysis Services OLAP cube hosted on physical hardware and subsequently synchronized to a SharePoint application server hosting corporate reports. As configured, the infrastructure was not suitable for delivering data to business users by 6AM daily, with costs impacting Hearst’s operational budget.

Instead of investing more money in expensive on premise hardware solutions, and infrastructure to manage it, Hearst wanted to explore the viability of a cloud implementation of their existing Analytics solution in Microsoft Azure.

Solution:

Decentrix was commissioned to migrate the existing Analytics solution into the most cost effective model in the Azure cloud. Decentrix took a hybrid approach to shift the heavy workloads into Azure without disrupting existing business processes relying on the on premise SharePoint solution.

Decentrix stood up three (3) servers, a domain controller, a data warehouse server, and a cube processing server. The domain controller was provisioned on the smallest VM available to act as the master control machine to coordinate booting and shut downs of other servers in the solution.

During the deployment phase, the data warehouse and the cube processing machine were provisioned on D14 hardware to provide maximum processing power at minimal cost. All virtual machines ran Windows Server 2012 and SQL Server 2014 Enterprise.

"We needed to take urgent action to improve our overall analytic environment for speed and costs. Although I was skeptical, the results achieved by Decentrix through Azure have exceeded all my expectations. The reduced processing times have been stellar, our costs are reduced and the stage is set for future projects in the Cloud. I can’t say enough good things about the team at Decentrix that made this all happen for us in no time at all.”
AL Lustgarten - VP, Information Technology & Administration, Hearst TV

Now in production, each night raw data is extracted from the operational traffic system into BLOB storage on the Azure data warehouse machine. ETL processes consume and transform the data into the desired structure for Hearst’s BIAnalytixTM data warehouse, and the cube processes the aggregated data upon completion. Once cube processing is complete, a multi-threaded synchronization to Hearst’s on premise SharePoint machine enables interactive analysis.

Benefits:

With BIAnalytixTM in the Azure cloud, Hearst is getting data insights faster and at a lower cost than they were when using an on premise architecture. Furthermore, Hearst now has the architectural flexibility to scale operations as data sources and volumes continue to grow.

Performance - By refactoring the ETL process to align better with the Azure architecture, the end to end processing time for nightly work was reduced by 40%. This means more flexibility in the processing schedule for business users. Also, resources in Azure were utilized to their maximum capability whenever possible because the entire Analytics solution was on dedicated equipment.

Cost - By implementing a rolling shutdown and start up routine, Decentrix was able to minimize the uptime of the servers in Azure, reducing overall costs. Machines were brought online at the beginning of the processing schedule and as processes and maintenance routines finished on each machine, the servers were gracefully shut down. By using this carefully managed operational technique, it’s estimated that the overall cost of hardware, software, and infrastructure support was reduced by 35%.

Scalability - Another key benefit gained by migrating to BIAnalytixTM on Azure is the native scalability available. Hearst can now pursue integrating other data sources because of both horizontal and vertical scalabilities afforded by Azure and the BIAnalytixTM architecture. BIAnalytixTM takes advantage of the inherent parallel capability of Azure to ensure that heavy workloads are distributed and processed efficiently. This reduces performance bottlenecks and improves throughput.

By moving Hearst’s current data warehouse into a Azure cloud environment, tightly coupled with Decentrix provided 24/7 Azure operational support, Hearst will now have the peace of mind that the system is being fully supported, is using latest technology enhancements, performing according to Hearst’s current and future business requirements, and helps Hearst more effectively budget for future support costs.

Copyright ©2001 - Decentrix Inc. All rights reserved.