Industrial Data Lake

This provides a solution for storing and exploiting large volumes of information of all types, processes or businesses. Our system brings together information from multiple systems (SCADAs, Historians, MES, ERP, logistics, commercial…) in large data lakes that allow information to be extracted by crossing multiple heterogeneous sources.

• Our Data Engineers create the primary data classification taxonomy.
• Frequently multiple databases are used (streamlined for time series in sensor scenarios, relational for process and non-relational for other types of data).
• These storage environments are deployed on a Kubernetes engine such as OpenShift, designing the optimal storage architecture in OnPremise, Cloud or Hybrid services according to customer requirements.
• On this deployment, redundancy and backup systems are available, alongside other systems for accessing and running Data Intelligence-type information.
• The data accessing process is thus standardised through API systems or data buses.
• An environment is created that allows for these to be run.

Often the end customer requires that all information related to the manufacture of the product be kept for a lengthy period of time (sometimes up to 10 years). This information must also be accessible quickly in the event of a claim. In highly regulated or standardised sectors such as the automobile, food or pharmacy, when faced with a customer complaint regarding a unit part, it is necessary to extract all the traceability information, including the process variables that record its manufacturing parameters. A properly organised Data Lake allows for information to be extracted swiftly with detailed reports that mix data, files and even videos of their manufacture.

Secondly, companies are increasingly aware that applying analytics on those volumes of data can extract highly valuable and non-obvious information. Those insights that are hidden among the data require tools that allow technicians who can manage the process to work with large amounts of data in a simplified way.