A Company Reduced Their Data Warehouse Cloud Costs by $240,000/Year. Here’s How.
A large healthcare company conducted a data warehouse usage assessment with Teleran prior to migrating their enterprise data warehouse to the cloud. By analyzing how their data warehouse was used by the business, they were able to create an optimized migration plan that reduced migration cost and complexity . They were also able to reduce wasteful usage in the cloud, These were their results:
- Lowered cloud consumption costs by $240,000 per year
- Optimizations improved data warehouse user productivity
- Reduced budgeted migration costs by $40,000
- Reduced migration time by 2 months, delivering faster time-to-value
A large US-based healthcare distribution and supply company, worked with Teleran to migrate their on-premises enterprise data warehouse to the cloud. Over the last several years the company has been reducing costs and streamlining business processes by eliminating data centers and moving applications to the cloud.
We were called by one of their largest divisions that operated a large data warehouse accessed by 1000s of users across many departments. They needed help gaining a deeper understanding of how their current data warehouse was used and what would be the challenges with migrating this critical business asset to the cloud.
Three Key Steps to Ensuring Cloud Data Warehouse Cost-Efficiency and Business Value
After meeting with the data warehouse management team, we identified the three key steps that needed to be taken to ensure a successful data warehouse migration.
- Accurate Cost Projections – Our customer needed to accurately project what it would cost them to operate their data warehouse in the cloud. While cloud pricing might seem straightforward, sizing a specific workload to create an accurate cost projection is challenging. The division needed a cost they could build into their proposal to senior management.
- Optimizations To Lower Ongoing Cloud Costs and Increase Business User Productivity – Our customer needed to “cloud-optimize” their data warehouse and analytics usage to reduce wasteful resource consumption and ensure a cost-optimized cloud data warehouse. Many on premises data warehouses are run on well provisioned servers where they have been able to tolerate or ignore inefficient, poorly formed queries and processes. With the cloud’s usage-based pricing, sloppy queries can drive up costs considerably.
- An Optimized Migration Plan to Minimize Costs, Risks and Level-of-Effort – Our customer needed to develop an optimized migration plan that identified and then minimized the risks, level of effort, cost, and time to migrate their data warehouse to the cloud. Most data warehouses contain sprawling sets of data used by many applications and users. Understanding what’s used and what’s not, what scripts and queries need, and how to subset workloads to prioritize and sequence migrating manageable portions of the data warehouse are just some of the issues to consider when building an effective migration plan.
Data Warehouse Usage Analysis for Cloud Migrations
We deployed Teleran’s automated Cloud Data Warehouse Assessment solution to analyze how their data warehouse was being used. It automates the capture and analysis of data warehouse workloads quickly and unobtrusively. As a SaaS solution, it is fast and simple to implement. It deploys an unobtrusive on premises agent to continuously capture comprehensive data warehouse usage metrics without any overhead on the database. These “fact-based” metrics deliver the foundation to accurately project cloud costs, develop effective and efficient migration plan and identifying the optimizations that minimize ongoing cloud cost and improve data warehouse user productivity.
Generating Accurate Cloud Resource Consumption and Pricing Metrics
The key metrics for cloud data warehouse pricing included concurrency/peak usage, dormant data, and data and compute resource usage by users and applications. Concurrency/peak usage analysis identified high watermark usage to project the computing resources and cost required to meet peak demands. It also identified little or no use periods when compute resources could be shut down to reduce costs. Dormant data analysis identified data that was not used and did not have to be migrated to the cloud, resulting in lower storage costs. This analysis used a combination of Teleran usage capture and database system logs to gain an accurate picture of what data was currently being used / not being used. The analysis also uncovered duplicate data across the data warehouse.
The user/application analysis categorized users and the applications by resource consumption and data used. This information enabled grouping of users and applications by resource consumption, enabling each group to be assigned only the compute resources they needed. The following illustrates data from one of the analyses:
Optimizations to Improve Ongoing Cloud Cost-Efficiency and Business Value
Teleran analysts also identified optimizations that, when addressed, eliminate inefficient user behavior and queries. Such optimizations enable a reduction in cloud-based resources used and improve the business value of the data warehouse. We identified 1000s of “wildcard” or “SELECT*” queries with no constraints that returned huge result sets unnecessarily. Eliminating these queries and requiring users to constrain their queries to only the data they really required can save a material amount of compute resources and improve user productivity.
Also discovered were high-cost queries that returned no data and many queries that resulted in database errors that had not been addressed. These further wasted compute resources and reduced data warehouse quality and analytical application accuracy. In addition, many users leveraged unauthorized inefficient tools to access the data that resulted in wasted resources that would prove costly in the cloud.
Accurate Cloud Pricing
Armed with the detailed usage “facts” and recommended optimizations, the team applied the knowledge gained to cloud pricing calculators enabling more accurate cost projections. Accurate costing enabled the team to establish a clear cloud migration ROI in their recommendation to management to move ahead with the project.
Optimized Migration Roadmap – Minimizing Risks, Costs and Time
The team was also able to identify data dependencies across applications and easily determine what data subsets and applications should be migrated first based on user and application business priority. They also were able to analyze the data extraction, transformation and loading (ETL) cycle and identify all source systems establishing the content and schedule for copying data into the cloud.
Customer Results: Faster Time-to-Value, Improved Business Value, Cost-Optimized
With Teleran’s Migration Assessment the team was able to achieve a faster time-to-value, reducing the overall migration time by two months versus their initial estimate. This translated into saving four man-months, decreasing budgeted migration costs by $40,000. The optimizations identified in the assessment enabled the team to significantly improve data warehouse cost efficiency, lowering ongoing cloud consumption costs by over $20,000 per month or $240,000 per year versus initial cost estimates. These optimizations improved performance and, importantly, increased business user productivity in in the cloud.
Chris is VP Marketing and a co-founder of Teleran. He has over 30 years’ experience in helping companies manage, leverage and protect their business-critical information.