Reduced upgrade costs with Snowflake Evaluation

Challenge

One of the largest multi-channel specialty retailers was encountering issues meeting regular cadence for reports and difficulty adding data or supporting analytics with their on-premise enterprise. Looking to reduce the costs of upgrading the following year, the client wanted to off-load some, or all, of its workload. Further, they sought to compare the cost of ownership over a five to ten year period for three separate solution offerings.

Solution

As part of our solution, Clarity conducted a 12-week Proof of Concept (POC) comparing LLAP on Hortonworks, Impala on Cloudera, and Snowflake. Tests were conducted to evaluate concurrent and constant load, and both data and compute scalability.

By extrapolating sizing for all platforms and accounting for yearly growth, Clarity determined migration costs and benefits for each of the three technologies—ranging from lift and shift to a complete rewrite—as well as development of multiple cost of ownership models that account for migration, support, infrastructure, and resources.

Outcome

From the evaluation of the three solution scenarios, Snowflake was determined to provide the most elasticity, scalability, and cost savings. During the evaluation, Clarity:

• Used approximately 10% of a 50+ TB warehouse
• Applied production data that included more than four fact tables of over a billion rows, with the largest having 45 billion rows
• Produced reports that represented their largest processing window, proving compatibility
• Evaluated total cost of ownership models with capital and expense costs and different years of depreciations

By migrating to Snowflake, the client is expected to reduce cost by 78% with Snowflake and get a full return on investment within 1.5 years of retiring their existing platform.