IBM Cloud Object Storage

IBM Cloud Object Storage [is] used for training and and data storage. It is mainly used by the data science department. It enables users who are not the most familiar with difficult devops procedures to work with big data and manage their data. It addresses problems like data synchronization, user training, and so on.

It hasn't been the most useful to our team, because our stack don't necessary expose apache spark; familiarity with apache spark is sometimes an issue with data scientist who have tensorflow or pytorch backgrounds, as it requires a decent amount of framework specific knowledge. Although SQL queries are used more extensively and helps with on-demand analysis.

Functionalities like data transfer to multiple regions has been very helpful. It allows us to build batch processes whenever and wherever, and not needing to worry much about underlining physical connections. It also enables things like multi-part upload and download with ease. Overall a high-speed transfer increases fault tolerance of applications that's sitting on top.

Yes, it has. Different storage tiers allows programmatic access of changing costs. It helps devops to manage necessary load and fine tune tiers for specific applications. We are able to leverage these pre-built, vetted solutions to build our production environment faster and plugin such storage to solve common customer problems.