Saturday, June 16, 2018

Locking the python interpreter

Seems simple if we have a processor function called 'crc code block' that generate an immutable code of the executable memory, just before re-entry (of the microprocessor code into user space).

In other words, a generic verification of executable space, done with an internal, or kernel protected code. Then the python users have an asynchronous, and rapid check on the python interpreter, but that can be proven apriori using language grammar theory.

The interpreter can, in turn, guarantee the python bots on the botchain.  Python bots can be made provable, apriori.    This is all doable, done (Inhaven;t looked), designed, or in the lab.  Likely the botchain start up will handle the entire security chain.

But, note, this is not complete pipeline control except the kernel based crc code check. But these trades run as fast as the python can interpret. If a greater segment of the interpreter can run in cache,  then we have less context switching, the switching between bot trades and notaries checking CRC. 

The mid level security system (not pipeline cache control) works, it causes asynchronous background 'checks', an additional hack opening due to induced context switching.

OK, great. Works well enough, it is insurable. Charge a traders a fee, remit to vendor of trusted interpreter, then sell bot codes by entry fees into he proven closed trading group.  We have stepped up the swap net layer.

No comments: