g302ge commented on issue #13424: URL: https://github.com/apache/arrow/issues/13424#issuecomment-1218345514
> > When I have updated the arrow::plasma project to support some features in our business. > > Can you explain what this means exactly? > > Also, regardless, Plasma is currently unmaintained and will probably be deprecated at some point... Sure, we have design a new data application, whose bottleneck is File I/O (some FXXK history reason). So these day we are trying do some hot patch to reduce the IO using big main memory to cache them. In our design every column in file could be regard as an object in Plasma, but we have to introduce some new feature designed for our situation. For example we use the `SHA1(file_path)` + `FileInfo.mtime()` as version + `Column Index` as an ObjectID, since we expand the origin ObjectID into 32bit. Sounds really tricky, but recent days the only road to reduce the IO is to build a new Database to replace File I/Os . LOL -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
