You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm defining a suite of core tests that can be run instead of a physical sensor node.
This is following on from #705
The advantage of this is it can provide near instant feedback for alpha changes in an ODM2 instance, and allow characterization of the implementation.
This also supports a developer to understand what type of logs are needed as per #718
This is very typical of software best management practices, and attempting to achieve the lowest cost of hours invested in software development and quickest turn around time.
Years of experience with software systems has shown that if software isn't tested, then software bugs dribble out over subsequent years - requiring heroic efforts to find them, and costly budgets to pay for it.
Engineering theory describes the benefit in providing a sharp impulse to a system, to test its response. This is likened to the old process of testing a bell cast in bronze. If it rings with a pleasant characteristic frequency and gracefully decays the system is good. If it cracks, or has a chaotic response - then its failed. In software terms - a failure is usually easy to detect when it is being tested, and much harder to determine root causes when it is running with real world conditions.
I would hope to do this in conjunction with alpha staging of the ODM2 whenever possible.
It could be used to exercise #649
defined in #674
and then in #688
It does require that there be some system objectives for the performance of a specific implementation, and typically this is done as a forward looking prediction of system loading.
The text was updated successfully, but these errors were encountered:
@ptomasula implemented a virtual node for stimulating and characterizing MMW with #729
which is a great short term alpha release check and excellent long term stability characterization.
Not being clear about the full extent of what is implemented with https://www.site24x7.com/, I'm envisioning that this test is for the alpha release stage, to create a pulse of known data, that can be verified as being delivered.
I'm defining a suite of core tests that can be run instead of a physical sensor node.
This is following on from #705
The advantage of this is it can provide near instant feedback for alpha changes in an ODM2 instance, and allow characterization of the implementation.
This also supports a developer to understand what type of logs are needed as per #718
This is very typical of software best management practices, and attempting to achieve the lowest cost of hours invested in software development and quickest turn around time.
Years of experience with software systems has shown that if software isn't tested, then software bugs dribble out over subsequent years - requiring heroic efforts to find them, and costly budgets to pay for it.
Engineering theory describes the benefit in providing a sharp impulse to a system, to test its response. This is likened to the old process of testing a bell cast in bronze. If it rings with a pleasant characteristic frequency and gracefully decays the system is good. If it cracks, or has a chaotic response - then its failed. In software terms - a failure is usually easy to detect when it is being tested, and much harder to determine root causes when it is running with real world conditions.
I would hope to do this in conjunction with alpha staging of the ODM2 whenever possible.
It could be used to exercise #649
defined in #674
and then in
#688
It does require that there be some system objectives for the performance of a specific implementation, and typically this is done as a forward looking prediction of system loading.
The text was updated successfully, but these errors were encountered: