In an article 'Cloud interoperability: Problems and best practices' of Computerworld, June 1st, 2011, Bill Claybrook states 'As more applications find their way to the cloud, data portability and other issues are coming to the fore'.
Data portability, by nature, is a platform issue, for the technical data format is usually optimized for the software that manage that data. To solve this problem we already have seen many initiatives, like SQL and XML. However, we need a more semantic format that descibes data in a way that is technologically agnostic and rich enough to describe the peculiar differences of the many datamodeling techniques. No abstraction but a formal description. Business intelligence companies have the knowledge to define such a standard, because they have experience with a variety of technologies for their ETL-tools.
And when the standard is in place, why not separate your data from your cloud software provider? This will solve both the portability issues and the vendor lockin.
My prediction is, that before 2020 data is separated from applications and freedom of choice for Software-as-a-Service is no longer limited because of the data portability issue.
No comments:
Post a Comment