Modul

At the end of the lecture the participants shall be able to explain what is specific to database functionality in the cloud, and what the advantages and disadvantages are. They shall have understood how cloud-enabled database technology differs from conventional technology of this kind, but also where the commonalities reside. Participants shall be able to explain the core ideas and approaches that define cloud-enabled database technology and discern them from each other.

We currently witness owners of large data sets, be they big organizations, be they startups, to rent database functionality to a significant extent, rather than providing it themselves. The total costs of ownership just happen to be much lower in many cases. This lecture features database technology that facilitates exactly this. This concerns you if you want to make use of such services at some time in the future, but is also is of interest if you will have to do with 'conventional' database technology.

According to my perspective, the following features of "cloud-enabled" database technology are key, and the lecture will cover them:

• Fully automated tuning of individual database – the option to interact with a database administrator does not exist any more!
• Approximate query results suddenly are attractive. The monetary costs of evaluating a query are commensurate with the necessary effort – on the other side, high fixed costs that typically occur with owner-operated databases do not incur any more.
• Multi-tenancy. I.e., how to ensure tenants that are completely independent from each other to have a DBMS for their applications available, not only without interfering with each other, but also with performance guarantees for each tenant individually?
• Secure storage. Administering data and evaluating queries shall take place in the cloud. At the same time, the cloud provider must not be allowed to see the data. Both objectives in full beauty currently are incompatible – we will discuss possible compromises.

In this setting, conventional, established concepts like distributed transactions and distributed data management and query processing play an important role as well, and the lecture will address them equally.