Containment Domains C++ API
0.1
Containment Domains C++ API v0.1
|
Functions | |
float | cd::CDHandle::GetErrorProbability (SysErrT error_type, uint error_num,) |
Ask the CD framework to estimate error/fault rate. More... | |
float | cd::CDHandle::RequireErrorProbability (SysErrT error_type, uint error_num, float probability, bool fail_over=true) |
Request the CD framework to reach a certain error/failure probability. More... | |
float cd::CDHandle::GetErrorProbability | ( | SysErrT | error_type, |
uint | error_num | ||
) |
Ask the CD framework to estimate error/fault rate.
Each CD will experience a certain rate of failure/error for different failure/error mechanisms. These rates depend on the system, the duration of the CD, its memory footprint, etc. The CD framework can estimate the expected rate of each fault mechanism given its knowledge of the CD and system properties.
[in] | error_type | Type of |
error_num | error/failure queried [in] Number of errors/failures queried. |
float cd::CDHandle::RequireErrorProbability | ( | SysErrT | error_type, |
uint | error_num, | ||
float | probability, | ||
bool | fail_over = true |
||
) |
Request the CD framework to reach a certain error/failure probability.
Each CD will experience a certain rate of failure/error for different failure/error mechanisms. These rates depend on the system, the duration of the CD, its memory footprint, etc. The CD framework may be able to apply automatic resilience techniques, such as replication, to ensure certain errors/failures will be tolerated (or simply) detected.
error_num
errors/failurse of the type queried will occur during the execution of this CD, after the automatic techniques are applied. Should be less than or equal to requested probability
if successful.[in] | error_type | Type of error/failure queried. |
[in] | error_num | Number of errors/failures queried. |
[in] | probability | Requested maximum probability of num_errors errors/failures not being detected or even occurring during CD execution. |
[in] | fail_over | Should redundancy be added just to detect the specified error type (false ) or should enough redundancy be added to tolerate the error (fail-over/forward-error-correction/...) |