"As previously explained, however, in-order updates (i.e., better crash behavior) are not practical in multitasking environments with multiple applications."
They say this is the obstacle to the best approach using the same interface. I say look at our best results in DB, concurrency, and distributed computing research to see if there's a solution that might work given a certain number of cores or applications. There's also the possibility of processor enhancements that put this sort of thing in the I/O architecture.
Comments
"As previously explained, however, in-order updates (i.e., better crash behavior) are not practical in multitasking environments with multiple applications."
They say this is the obstacle to the best approach using the same interface. I say look at our best results in DB, concurrency, and distributed computing research to see if there's a solution that might work given a certain number of cores or applications. There's also the possibility of processor enhancements that put this sort of thing in the I/O architecture.