Virtual Data Access Layer
Why it matters?
These days, however, corporations are having to continuously add more types of data (ranging from structured to unstructured) and more locations of data (internal and external to the enterprise) to the set of data sources that matter to their business. This is, in reality, a big challenge for data integration suites that are traditionally limited in the types of data they integrate. As the complexity of data increases, the key challenge for a data integration tool is to offer to the consuming business applications a simple unified abstracted data interface while hiding the diversity and dispersion of the distributed data repositories.
That's exactly what a Virtual Database and data federation does: it provides a unified single repository of federated data from a set of distributed heterogeneous data sources offering the full power of the relational algebra (conveniently extended to deal also with non-relational data sources) to create, combine, transform and federate data views and deliver the required SOA virtualized data services layers. What databases did to improve the interaction of business applications with file systems, Virtual Databases do to databases and other heterogeneous data sources, but taking into account a higher degree of dispersion of data and a wider variety of data sources than was common in the centralized traditional data center.
With a higher degree of data dispersion and a wider diversity of data sources, ETL (Extract, Transform, Load) and the renewed versions of the same old data integration tools, mostly renamed as data integration suites, simply don't work. Often, corporations get ready to pay huge amounts for professional services to build custom-made solutions with these data integration suites. Unfortunately, these suites were designed to operate inside the walls of a centralized data center and do not have the capability to work with all the data sources that matter today.
In the era of the Web data, the Cloud data and SaaS applications, the traditional approach will simply not work. With the Internet globally interconnecting applications and repositories, with the Internet becoming a growing repository of valuable structured and unstructured web data for any business and a unique communication and computing platform ready for enterprise use, it becomes clear that data integration is not anymore just about moving data to new consolidated data stores.
Finally, conventional integration technologies are expensive, as you may have already experienced in your past projects. Not only are the solutions inflexible, every new integration project demands more servers, licenses and staff to deploy, operate and maintain the new systems driving up the total cost of ownership of the integration tool. On the other hand, Virtual Databases and data federation provide flexible solutions, easy change management and lower total cost of ownership. With Data Virtualization and Data Federation solutions from Denodo, you'll have the option to physically store your data views or even write back to the data sources if necessary.