To perform channel access read/writing we intend to move from cothread to aioca. Aioca uses the python asyncio standard imports and so will be better supported in the future.
As part of the transition we are also hoping to improve the performance of the virtual accelerator, possibly by optimising the requests queueing system, but this is still unknown.