Dit bericht verscheen eerder bij FOSSlife
A model for data management builds on the various sources ranging from traditional databases to business applications and their data storage systems to analytic applications. All of these sources often generate information themselves, which, in turn, can serve as a source for other usage scenarios.
The first integrating layer is metadata management and the resulting data catalogs, which provide an overview of what data (and of what quality) can be found where. However, state-of-the-art tools also provide more detailed information on the data lineage (e.g., the origin of the data) and enable evaluation and collaboration for the information.
A level above is data integration and quality. Data integration products such as extract, transform, load (ETL) and extract, load, transform (ELT) support the integration of information from different sources and the implementation of data formats.
Data quality tools check the quality of information to supplement or correct the data as needed. External sources such as address databases are often accessed for this purpose. Master data management (MDM) builds on this foundation and delivers function- and industry-specific applications for handling information such as product data.
Another level above are the analytical applications and functions for data usage (e.g., for serving up user-specific content in digital services or for decision support).
Specific Data Protection
The functions of data governance act as an overall theme across the layers of a data fabric. One central topic in this context is data security, which in recent years has developed beyond individual technical solutions for security of classic, relational databases.
Database security continues to be an important subarea in this context, protecting databases against breaches of integrity, confidentiality of information, and availability. Security primarily involves functions for the information itself stored and processed on database systems, as well as the underlying server and network infrastructure and access to the information.
However, as infrastructures and technologies for processing and storing data change — especially given cloud-native tools and the resulting hybrid infrastructures of modern and legacy approaches — the requirements change. The core functions of modern data security products include the following functional areas:
- Vulnerability assessment: identifying potential points of attack, configuration errors, and other dangers.
- Data discovery and classification: knowledge of the data and classification of the data in terms of sensitivity (e.g., personal information); tools ideally build on existing infrastructures for metadata management and data catalogs.
- Data protection: encryption, tagging, and other technologies for both storage and data transfer.
- Monitoring and analysis: continuous monitoring of access to and the use of data, and analysis to detect and respond to anomalies, including integration with security information and event management (SIEM) tools.
- Threat prevention: guarding against targeted attacks such as SQL injection.
- Access management: targeted protection of privileged user accounts and dynamic, policy-driven access control; often handled by specialized applications.
- Audit and compliance reporting: automatically generated and ad hoc reports and dashboards for an overview of the current security status.
These tools are fundamental building blocks of a modern, secure data fabric and must be designed to support complex hybrid environments and multicloud platforms.
As mentioned earlier, data governance is more than just data protection and is best defined as an umbrella term covering various functions of protecting and controlling data and data usage. Two other important functional areas in addition to data security are:
- Privacy management for handling information that falls under the scope of the GDPR. Like the entire topic of data management, it is no longer just about structured data, but also about unstructured data that needs to be analyzed, managed, and protected.
- Data governance and risk is a sub-area that focuses on concrete metrics and control functions that can be used to monitor and improve compliance with defined rules for handling data. On the one hand, this sphere includes regulations in the area of data protection such as the GDPR; on the other hand, it encompasses other requirements for handling sensitive information, as well as internal rules for handling and protecting particularly critical and valuable data. Such tools typically dovetail with IT governance, risk, and compliance (GRC) products to deliver data into a higher level of risk management.
A good strategy for data governance is always built on integration of the specific functions with data security, as well as the underlying metadata management and data catalogs.
Flexibility and Coordination
Perhaps the biggest challenge in implementing a data fabric — and in sub-areas such as metadata management or data governance — is that people work with data everywhere in organizations. The multiple areas of use and numerous stakeholders make it difficult to avoid a proliferation of initiatives and technical approaches.
However, precisely here, well-thought-out and comprehensive approaches can help, including an architecture with associated operating models (e.g., target operating models, TOMs) for the data fabric and service-oriented approaches for providing the technical implementation. Experience shows that very few divisions in the corporate environment actually want to implement their own tools if they can turn to a functionally useful service with a suitable operating model. In other words, with a correctly implemented data fabric as an internal service, a corporation has a good chance of containing a significant amount of undesirable growth.
Communication is important so that the different divisions in the company know which services are available. Because the users and areas involved come from both IT and business, advisory support is required on top of technical services. In many cases, today’s tools in the wider data management environment support functions for collaboration; assessing data usability; and collaboration among users, data stewards, and administrative and technical users.
The importance of data security means corporations need to move away from isolated, incomplete tools that are expensive, often fail to cover important security and data governance requirements, or do not do a complete job of providing coverage. An additional consideration is that continually introducing new local solutions simply takes too much time. However, solutions must also serve the quite different requirements of stakeholders from business, IT security, data protection, and other areas.
To deal with data efficiently and effectively, corporations need a strategy in which a holistic view of the required elements (e.g., a data fabric) plays a central role. Individual approaches in the area of analysis are not enough. IT managers need to implement both the foundations with metadata management and data catalogs and the interdisciplinary functions of data governance and data security correctly to be able to work optimally with data and generate the desired added value.
This article originally appeared in ADMIN magazine and is reprinted here with permission.
Dit bericht verscheen eerder bij FOSSlife