Who owns the data?

At least once, and usually several times in every IT or integration project, someone asks the question: "Who's the data owner?" This question reflects older system architectures which considered owning the data as a measure of system importance. In old architectures, systems controlled data and shared it reluctantly, if at all, with other systems.

By Dennis Brandl, BR&L Consulting December 1, 2005

At least once, and usually several times in every IT or integration project, someone asks the question: “Who’s the data owner?” This question reflects older system architectures which considered owning the data as a measure of system importance. In old architectures, systems controlled data and shared it reluctantly, if at all, with other systems.

Today we realize that information sharing is valuable and a desirable system trait. Asking for the data owner ignores the different roles that need to be performed with data. I find that it is better to ask several questions, such as: What system is the data repository? What system is the data custodian? What systems are data producers? What systems are data consumers? What systems are data distributors?

Data repositories have responsibility for ensuring that data are available when needed. To achieve this goal they include capability for storing, archiving, and restoring data, and they provide redundancy and resiliency to meet their service commitments. Data repositories in IT projects are often commercial database systems and document management systems.

Data custodians have responsibility for ensuring the correctness of data. Data custodians execute data integrity rules to ensure that bad or inconsistent data is not stored in the repository. Data integrity rules can be implemented using database integrity constraints, XML schema rules, and specialized code. Data custodians are typically implemented in application programs, such as the back end of an ERP system. However, other systems are also data custodians. Usually an MES system or data historian is the custodian for site production data, and LIMS is the custodian for site laboratory data.

Data producers have responsibility for collecting information from people or other systems. Many times the data custodian is also a data producer, but when information is shared, there may be several data producers. For example, there may be multiple systems that can set the quality status for a material lot. The ERP system may be a producer of the original quality status when the material is received into the company, a LIMS system may update the quality status based on test results, and an MES system may update the quality status based on a production event, such as a spill.

Data consumers use data and also have responsibility for presenting the data in the correct format to people or other systems. Many times the data custodian is also a data producer. For example, the front-end of an ERP system is both a data consumer and producer.

Data distributors have the responsibility for ensuring that consumers and producers have access to the data when needed. Data distributors have local caches for data to ensure data availability when the custodian system is unavailable. Not all systems have data distributors, but when there is a single data custodian, multiple producer, and consumer systems with different uptimes, then data distributors are often used. The need for data distributors is often based on a risk/cost analysis. If the producers and consumers can be down when the custodian is down, then no data distributor is needed. Often there is a data distributor to between ERP, MES, and LIMS systems to handle their different uptime requirements. For example, a data distributor may distribute material lot status information from an ERP to an MES, or distribute test results from a LIMS to an ERP.

Another question often asked is: “What is the system of record?” This question is usually associated with some form of regulatory compliance. The system of record is the system that has responsibility for the quality of data and manages the data validation rules the system of record is the data custodian.

When asked a question about who owns the data, respond by trying to understand all of the roles that systems will have and don’t get involved over a battle of data ownership.

Author Information

Dennis Brandl, dbrandl@brlconsulting.com , is the president of BR&L Consulting, a consulting firm focusing on manufacturing IT solutions, based in Cary, N.C.