|
@ -87,13 +87,13 @@ an upper bound, the *maximum time* assigned to a proposal. |
|
|
There exists a system parameter `MSGDELAY` for end-to-end delays of proposal messages, |
|
|
There exists a system parameter `MSGDELAY` for end-to-end delays of proposal messages, |
|
|
such for any two correct processes `p` and `q`: |
|
|
such for any two correct processes `p` and `q`: |
|
|
|
|
|
|
|
|
- If `p` sends a message `m` carrying a proposal at real time `t`, then if `q` |
|
|
|
|
|
receives `m`, it does that at time `t'` such that `t <= t' <= t' + MSGDELAY`. |
|
|
|
|
|
|
|
|
- If `p` sends a proposal message `m` at real time `t` and `q` receives `m` at |
|
|
|
|
|
real time `t'`, then `t <= t' <= t' + MSGDELAY`. |
|
|
|
|
|
|
|
|
While we don't want to impose particular restrictions regarding the format of `m`, |
|
|
|
|
|
we need to assume that their *size is upper bounded*. |
|
|
|
|
|
In practice, using messages with a fixed-size to carry proposals allows |
|
|
|
|
|
for a more accurate estimation of `MSGDELAY`, and therefore is advised. |
|
|
|
|
|
|
|
|
Notice that, as a system parameter, `MSGDELAY` should be observed for any |
|
|
|
|
|
proposal message broadcast by correct processes: it is a *worst-case* parameter. |
|
|
|
|
|
As message delays depends on the message size, the above requirement implicty |
|
|
|
|
|
indicates that the size of proposal messages is either fixed or upper bounded. |
|
|
|
|
|
|
|
|
## Problem Statement |
|
|
## Problem Statement |
|
|
|
|
|
|
|
|