Skoda Auto Makes Data-Driven Decisions Using an Advanced Analytics Platform
servers use the solution implemented by Adastra
terabytes is the total data storage capacity
months to connect Adoki to the required data sources
In 2018-19, Adastra built an on-premise Data Analytics Platform (DAP) at ŠKODA AUTO. Its purpose? To visualize data and use advanced analytics and artificial intelligence to perform sophisticated tasks with large volumes of data.
A detailed look at the Data Analytics Platform at ŠKODA AUTO
At ŠKODA AUTO, the platform is used for Production and Logistics, Quality Control, Marketing and Technical Development, as well as in the Warranty department, which takes advantage of smart scoring for warranty claims to be invoiced.
The platform also makes it possible, for example, to gauge customer loyalty, predict customer behavior, evaluate needs and tailor product and service offers accordingly.
Jiří Boček, the DAP’s product owner, says its potential lies primarily in combining data from different departments and finding connections between individual indicators.
Interest in using the DAP is growing steadily alongside the need to process data, both at ŠKODA AUTO and within the Volkswagen Group as a whole. Today, it is already a hybrid platform based in the ŠKODA Data Center in Mladá Boleslav and in the cloud.
Voice of the Customer
months took to connect Adoki to the required data sources
has consolidated most other ETL tools and technologies
transfers are checked and recorded in an auditable history
IT specialist manages all data transfers
on-premises and cloud environments are brought together
Adoki provides a complete metadata service for all transfers
It takes system performance into account, and offers ŠKODA AUTO’s data specialists an intuitive front end.
In the initial phase of the project, we synchronized 40 data transfers from 8 of the car manufacturer’s key databases. After 6 months, there were almost twice as many transfers; one year on, the number has increased fivefold and is still going up.
In terms of volume, Adoki consistently transfers hundreds of GB of data and ensures complete data governance.
All the data transfers can be managed by a single IT specialist using Adoki.
We combine on-premise and cloud environments
The benefits of the on-premise component include
- its large capacity (149 TB, which is roughly 300 typical laptop disks),
- fast data processing,
- the wide variety of formats used,
- an extensive set of tools for reporting, analytics and artificial intelligence.
At the same time, users from ŠKODA AUTO appreciate that the Cloud:
- offers unlimited capacity for storing data,
- is easily scalable,
- provides analytical tools that are always up to date,
- ensures GPUs are available for compute-intensive operations, and more.
With the growing interest in uploading and storing data on this platform, ŠKODA AUTO faced the dilemma of how to manage data transfers to and from the company’s various systems and the DAP, as well as between the DAP’s on-premise and cloud components. Before Adoki was deployed, the situation required the involvement of an entire team of specialists and, even then, did not provide a full user service.
For example, setting up transfers could only be done manually; each transfer had to be newly configured and parametrized, and no one received any notifications when a transfer finished. Programming new functionalities would have been possible but costly. This is precisely why ŠKODA AUTO decided to implement Adoki, a central tool that manages and monitors all data transfers. Adoki will also be used for transfers between the on-prem and cloud parts of the DAP.
Data transfers are performed several times a day
Adoki’s first major task was to integrate data from 8 of the car manufacturer’s production databases to the DAP. Today, it integrates data from many other relational systems, including MS-SQL, Oracle, PostgreSQL and SAP HANA, as well as non-relational systems like Jira.
- On average, hundreds of GB of data are replicated every month.
- Each transfer is completed within 3 minutes.
At the beginning, data transfers were carried out 3 times a day, after each production shift ended. Now, they are performed several times a day, ad-hoc, according to business needs and requirements.
- Transfers are batched: smaller tables are uploaded in their entirety; large transactional tables are done incrementally.
- The DAP and Adoki also serve to archive production data.
During the initialization process, when historical data was uploaded to the DAP (from as far back as 2016 for some tables), Adoki had to deal with significantly higher volumes of data. However, individual transfers still completed in under 5 minutes. After transferring the data from the primary systems and checking that the entire process had completely finished, the data specialists could safely delete the data from the source systems to free up capacity.
Implementing Adoki was more cost effective than supporting and developing the previous solution
At ŠKODA AUTO, the tool is managed by a single specialist and carries out data transfers to hundreds of target tables. The cost of implementing Adoki was more financially viable than supporting and further developing the solution that was previously in use.
Adoki has accelerated, standardized and streamlined data transfers from key systems to the DAP.
It is entirely automated and unconstrained – it does not limit either the number or size of tables to be transferred, nor does it force ŠKODA AUTO to transfer only select sources.
At the same time, it also ensures data and data structures remain consistent in both the source and target environments.
Audited and monitored transfers that only require a single specialist
All data transfers are fully auditable and monitored
If an error or interruption occurs during the course of a transfer, the tool immediately notifies the transfer administrator, who can then easily track down the marked transfer and restart it.
All data transfers, of which there are now 250 a day, are managed by a single person. This is possible because, among other things, the tool uses metadata and parametrization.
Adoki transfers data quickly and securely; the transfer process is covered from end to end. The tool also ensures that the data transfers are automatically optimized so that they maximize computational performance and do not limit the source databases’ operation.
The tool is user-friendly and intuitive thanks to its sophisticated graphical interface.
This is one of the reasons why the specialists working with the tool could be trained in one day. After their training, they can configure data transfers entirely on their own, independent of the Adoki administrator. Users can set up simple transfers within an hour, more complex ones within a day.
