Significant cost reduction after onboarding to Snowflake16.08.2021
Elektro Ljubljana is the largest electricity distribution company in Slovenia with their HQ in the city of Ljubljana. The company operates the largest distribution network in Slovenia, covering the central and south-eastern parts of Slovenia and ensuring that electricity efficiently reaches more than 338,500 customers. The electricity infrastructure covers 6,166 km2 (30.4% of the country’s area).
Elektro Ljubljana carries out various (i.e., statutory) network activities and provides a wide range of commercial services related to the electricity infrastructure.
Due to the growing need for reporting, the company decided to modernize its existing data warehouse in order to meet the new requirements that have arisen in recent years as a result of more and more data being scattered across various sources or applications.
Elektro Ljubljana recognized In516ht as the best fit for their data management project. During the first phase, they decided to build a data warehouse for the purchasing/procurement application area, pulling data from two separate source systems MS Dynamics AX and IBM Maximo platform for electricity billing.
The customer’s initial idea was to build the data warehouse on their existing on-premise infrastructure using MS SQL and SSIS for ETL procedures and to create the reports and dashboards using Cognos Analytics BI tool and MS PowerBI.
Shortly after the project started, we noticed performance issues on the existing infrastructure because of under-provisioned hardware and compute resources. Due to the insufficient resources Elektro Ljubljana would have to invest into new infrastructure (HW processing and storage and SW licenses).
In516ht did the TCO calculation for the investment to current infrastructure and proposed modern architecture with Snowflake Data Cloud running on Microsoft Azure Public Cloud as an alternative. The TCO calculation clearly showed that investing in cloud resources would cost 50% less than investing in the new on-premise infrastructure. In addition, Snowflake brings unlimited scalability, workload isolation, a free testing and development environment, the ability to securely share data for regulatory reporting, and most importantly, the ability to introduce new use cases without any compromise. During the first phase, In516ht performed a lift’n’shift migration from MS SQL to the Snowflake Data Cloud and continued to develop new use cases on the new platform.
In addition to the new modern cloud services, In516ht introduced a modern ELT approach with the In516ht asset DataMerlin for data warehouse automation. This is how a new ELT data integration approach was introduced to take advantage of cloud resources. Data Merlin accelerated ELT development, introduced the DQ (data quality) module, data lineage and an “out of the box” automatic documentation.
Elektro Ljubljana data warehouse is running on Snowflake’s Enterprise Edition to take advantage of the multi-cluster architecture and time travel functionality. The data warehouse is based on the Kimball dimensional model. In516ht migrated existing business areas and rebuilt the ETL using DataMerlin.
Based on In516ht data warehouse methodology, we split the development into several phases starting with procurement/purchase to minimize the risk and get results faster. The client continues to expand the data warehouse with new use cases: General Ledger, Balance Sheet, Financial Dimensions, Work Orders, Employee Productivity, HR Analytics, Account Receivable/Payable and Fixed Assets Analysis. The overall goal is to expand reporting capabilities across the entire enterprise.
Thanks to In516ht assets, which were developed with the aim of supporting data migrations, the entire migration process was completed in a very short period of time, in a single month, and development continued on the Snowflake Data Cloud.
It should be noted that we have reduced costs by 1000% in terms of infrastructure and that there are no licenses with Snowflake. We have managed to combine the major ERP systems Microsoft AX and IBM Maximo, alongside other dedicated applications. For users, we have prepared analytical reports with deep mining capabilities, as well as advanced control panels where users can create their own reports.Iztok Bizjak, Head of Business Intelligence (BI) at Elektro Ljubljana d.d.
Outcomes of the Project & Success Metrics
- Snowflake was faster in all respects when compared to the existing MS SQL Enterprise edition. The TCO is calculated significantly lower at 50%.
- Users can now access reports without resource contention.
- Customer has a proper development and testing environment at his disposal.
- Customer is able spin up additional resources when needed without upfront investment (month and quarter end reporting).
- Daily batch job is finished 80% faster than on the old database.
- Significantly lower DBA activities. Regular DBA is not needed any more, resulting in resource savings.
The data warehouse user is usually not interested in technology that allows him to see the results of the analysis. Therefore, technology decisions are the responsibility of the ICT department. As the CIO at Elektro Ljubljana, I am not the only one making the decision, as the analysts and technologists are the ones actually working with this technology. With this in mind, I decided to listen to my colleagues who are more knowledgeable in this particular field. They assessed (and later proved) that using the Snowflake Data Cloud would cause less hassle in terms of scalability, security, and future development.Alenka Kolar, Executive director ICT at Elektro Ljubljana