The term “data texture” is used through the technology industry, however its definition and implementation can vary. I saw this across the sellers: in the fall of last year, the British Telecom Company (BT) talked about the data texture in an analyst event; Meanwhile, in storage, NetApp has redirect its brand into a smart infrastructure but was previously using the term. Platform Platform Appian has a data tissue product, and the Mongodb database provider also talks about data fabrics and similar ideas.
In essence, the data tissue is a uniform structure that is bridging and integrating the different data sources to create a smooth data layer. The principle is to create a synchronous uniform layer between divergence sources and work burdens that need access to data – your applications, work burdens, increasingly, AI algorithms or learning engines.
There are a lot of reasons for the desire of such an overlap. Data tissue acts as a generalized integration layer, communicating different data sources or adding advanced possibilities to facilitate access to applications, work burdens, and models, such as enabling access to those sources while keeping them simultaneous.
So far, so good. However, the challenge is that we have a gap between the principle of data tissue and its actual implementation. People use the term to represent different things. To return to our four examples:
- BT defines data texture as an overlapping over the network level designed to improve data transfer over long distances.
- The explanation of NetApp (even with the term smart data infrastructure) emphasizes the efficiency of storage and central management.
- Appian places the Fabric Data Form as a tool to unify the data in the application layer, allowing the development and allocation of tools facing the user faster.
- Mongodb (and other structured data service providers) consider the principles of data fabric in the context of data management infrastructure.
How do we cut all this? One answer is accepting that we can handle it from multiple angles. You can talk about the concept data texture – realizing the need to combine data sources – but without exceeding. You do not need a global “Uber-Fabric” that covers everything at all. Instead, focus on the specific data you need to manage.
If we are returning for two decades, we can see similarities with the principles of architectural engineering directed to the service, which are looking to separate the service provision of database systems. At that time, we discussed the difference between services, operations and data. The same applies now: you can order a service or request data as a service, focusing on what is required of your work burden. Create, read, update and delete, remain the most direct data services!
I also remember the assets of the network acceleration, which will use cache to accelerate data transfers by keeping copies of data locally instead of accessing the source over and over again. Akamai has built its works on how to transfer irregular content such as music and movies efficiently and at long distances.
This does not suggest that the data fabrics invent the wheel. We are in a different world (cloud) technically; In addition, it brings new aspects, not the least of which is to manage descriptive data, track lineage, compliance and security features. This is especially important for the burdens of artificial intelligence, as data governance, quality and the source directly affect the performance of the model and the merit with confidence.
If you are considering spreading data tissue, the best starting point is to think about what you want for data. This will not only help you direct you towards the type of data texture that may be the most appropriate, but this approach also helps to avoid the trap of trying to manage all data in the world. Instead, you can give priority to the most valuable sub -group of data and consider the level of data texture that works better for your needs:
- Network level: To merge data through multi -sodium environments and edge environments.
- Infrastructure level: If your data is central with one storage seller, focus on the storage layer to serve the coherent data gatherings.
- The level of application: To assemble the different data collections of specific applications or platforms.
For example, in the case of BT, they found an internal value in using their data tissue to unify the data from multiple sources. This reduces duality and helps to simplify operations, making data management more efficient. It is clear that it is a useful tool for unifying silos and improving the rationalization of the application.
In the end, the data tissue is not a homogeneous solution, suitable for everyone. It is a strategic conceptual layer, supported by products and features, you can apply as it makes sense to add flexibility and improve data delivery. Publishing fabric does not represent a “appointment and forgetting” exercise: it requires a continuous effort to achieve and preserve publishing and publishing – not only the program itself but also forming data sources and integrating them.
Although data tissue can be found in concept in multiple places, it is important not to repeat delivery efforts unnecessarily. Therefore, whether you collect data together via the network, or within the infrastructure, or at the application level, the principles remain the same: use it where the most appropriate for your needs, and enable it to develop with the data you serve.