- Nov 12, 2021
-
-
Michael Eischer authored
-
Michael Eischer authored
-
- Jul 08, 2020
-
-
Michael Eischer authored
There's no need for the current replica to further process the request.
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
If a reply message completes a certificate while in the same moment the lock acquire inside invoke times out then the following retry instruction can reset the certificate even though the lock was already released.
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
This ensures that the output of multiple threads won't be mixed up.
-
Michael Eischer authored
As both stdout and stderr point to the same file this can cause garbling of log outputs between threads.
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
This simplifies handling of only part-time active clients
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
The first sequence number on the replicas is zero
-
Michael Eischer authored
Conflicting read request replies are not the leaders fault. So just resend the request normally the first time.
-
Michael Eischer authored
A client could also reset it's certificate and retry the read request for a newer sequence number.
-
- Jul 03, 2020
-
-
Michael Eischer authored
-
Michael Eischer authored
This should lead to less mangled stack traces
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
-
Michael Eischer authored
The minimum is necessary for generating the keys not when loading them
-
Michael Eischer authored
A replica directly replies with its current state to a weak read requests. Therefore the client has to send the request itself to all replicas.
-
Michael Eischer authored
This fixes hashed replies and weak read requests. In BFT mode all replica should return a reply to the client. The only question is whether a hashed reply is enough or not.
-
Michael Eischer authored
As the benchmark policy and the interval statistics start at nearly exactly the same time this could cause the interval statistics to report one interval too much or not. This change slightly delays to policy start and suppresses printing of the last incomplete interval. That way the first interval captures slightly too little data (but that doesn't matter as it's normally cut off anyways) and the last incomplete interval only captures data for the fraction of a second.
-
Michael Eischer authored
This is useful for experiments where not all clients are active right from the start
-
Michael Eischer authored
-
Michael Eischer authored
-