Using set for CID allocations provides better API.
There's also some slow behavior in releasing of existing CID, which can cause conflicts with random allocation. Therefore the CID allocation here uses sequential approach, jumping over already-reserved ones, so that probability of hitting a recently released CID is as low as possible.
Using set for CID allocations provides better API.
There's also some slow behavior in releasing of existing CID, which can cause conflicts with random allocation. Therefore the CID allocation here uses sequential approach, jumping over already-reserved ones, so that probability of hitting a recently released CID is as low as possible.