Apache/Arrow with Mainframes

[Bdartlon]

Hello All,

I am new to this community and was actually directed here by a colleague recently. My primary role at work is helping make mainframe data reliably and securely available off host via REST APIs. As mainframes are inherently designed to store and process large amounts of data, my question is, would there be any need/benefit to adding some mainframe support to this project?

I am uncertain what this might look like, but I am hopeful that if it has any merit, we can explore the options here.
If this is the wrong format/location for this query or if it has no value, please let me know!

[lidavidm]

We did test on S390x in the past, but I think with some CI migrations that was lost.

What would you envision for "mainframe support"? As long as you have a compliant and reasonably recent C++ compiler, the core libraries here should build (if all you have is a C compiler, something like nanoarrow may work better). As you mention REST APIs, there's also some work on demonstrating what Arrow + HTTP looks like: https://github.com/apache/arrow-experiments/tree/main/http