Implement epoch friendly binary tree

eBPF object IDs should be stored in order to permit fast enumeration by id.

Currently there is no support for a lexicographical traversal of the object ID table, with a O(N^2) work around.

Begin by finding the eBPF object ID table and its current O(N^2) enumeration workaround in the repository. Read the surrounding ID-management code to determine the required traversal and epoch constraints. Done means IDs are stored in an epoch-friendly binary tree and can be traversed lexicographically without the current workaround.

Written by the indexing model from the issue text.