Which is the exact idea of this enhancement request? GLIBC already has support for interposable malloc, so the usage of an external interposable malloc is straightforward.
If the idea is *replace* current malloc implementation I think a better strategy would bring this up to libc-alpha instead of opening a bug report. This is a huge task, that would require to take into consideration multiple factors, and it will require much discussion.
Also, the malloc implementation has been improved both in security, (for instance 5b06f538c5, c0e82f11735, ebe544bf6e8e), usability and performance (the tache support for instance). So I would prefer to work by identifying issues with the current implementation and work towards fixing them while taking into consideration not only security but performance metrics as well (as we have done in the past).
The proposed allocator is 64-bit-only, so not suitable for glibc anyway.
It would be good if such proposals could at least identify deficiencies in current implementations, instead of generic request to replace a battle-proven implementation that is currently developed.