The current behaviour is to read an entire object into memory on retrieval, which has scaling problems when dealing with high load and/or large object sizes.
Streaming content from a backend would make this kind of use more feasible, but would potentially limit features like verifying data integrity on fetch.
Imported from Launchpad using lp2gh.
The current behaviour is to read an entire object into memory on retrieval, which has scaling problems when dealing with high load and/or large object sizes.
Streaming content from a backend would make this kind of use more feasible, but would potentially limit features like verifying data integrity on fetch.
Imported from Launchpad using lp2gh.