Using the S3 service from AWS can be really convenient, but there are also some drawbacks to it: it implies a recurrent cost, and you have to put your data on a datacenter (the cloud). In some cases it would be very convenient to use an S3 storage, but the cost and/or the fact of storing the data off-premises is a problem. For example, in a data science team working with relatively small datasets (some 100th of a GB), using Spark is very convenient and easy to set up. One of the sources of data supported by Spark is S3. In this case, it can be easier and less costly to host a small Spark cluster in premises, as well as an S3 database, rather than using cloud services.
Since S3 is the de facto standard, there are ...
