DATA WAREHOUSE OPTIMIZATION
Traditionally, data had been stored centrally in enterprise data warehouse where it provided business insights to the organization. In the last few years, tremendous growth has been observed in terms of data. Data has been growing at a 40% compound annual rate and faster in some of the industries. Technologies around data are evolving faster. The cost to store the new type and amount of data motivate the organization to explore the alternative architecture. It is difficult to start a big new data warehouse project even though benefits to create modern data warehouse architecture and solution are clear. We see that many organizations start their modern data warehouse journey with extending and Data Warehouse Optimization. Our analysis shows, that it generally costs between $15,000 and $25,000 per terabyte per year, at list prices, to store and run a good-sized data warehouse. Our recommended approach is based on the following three principles.
- Use Hadoop ecosystem technologies for storing and processing the data. Load only the relevant data in the existing warehouses. This will ensure that you are not changing the existing tools and processes for larger organization.
- Use modern architecture for new data sources. There is possibility to move some components of data warehouse solutions to modern cost effective architecture without affecting a lot of processes. We recommend use columnar data warehouse solutions e.g. Azure data lake, redhshift which provide upto 5X compression on data.The use of cloud based data solutions has advantage of ‘pay as you go’ model and spot pricing.
- Start new initiatives with cloud based data warehouse solutions. Plan for an extended data warehouse solution from your on-premise solution and use that as a candidate case for all the new initiatives.
At Innovyt, we can plan to save 20%-50% of current data warehouse cost and flexible architecture to accommodate future growth of data. We have built some accelerators to plan and architect Data Warehouse Optimization and extension.