|
|
|
|
||||||
| comp.protocols.tcp-ip TCP and IP network protocols. |
![]() |
|
|
LinkBack | Outils de la discussion |
|
|
#1 |
|
Messages: n/a
Hébergeur: |
We are running a customized TCP/IP stack on Windows. The problem is how
do deal with the case of hibernation with respect to the stack timers. There are 3 possible ways : 1. To do nothing 2. To jump the clock to the real time after coming back from hibernation. In this case most of the TCP timers are going to time out (assuming the hibernation time is long enough) and behavior is that of the timers suddenly timing out. 3. To fast forward the clock, that is actually run the TCP/IP timers faster, so that all the events that would have actually been trigerred by the timers during hibernation would be compressed in a very small amount of time, possibly resulting in a flurry of retransmit segments or keep-alives. I wonder how windows deals with the situation, and what could be relative pros/cons of these approaches.. Can anybody ? --Deb |
|
|
|
#2 |
|
Messages: n/a
Hébergeur: |
Here is an idea:
Store the time when hibernation occured. Determine the time when it resumed from hibernation. Calculate the interval. Add the interval to all timers, timestamps etc. Bye, Skybuck. "debdut" <debdut@gmail.com> wrote in message news:1143312423.004991.260760@v46g2000cwv.googlegr oups.com... > We are running a customized TCP/IP stack on Windows. The problem is how > do deal with the case of hibernation with respect to the stack timers. > There are 3 possible ways : > > 1. To do nothing > 2. To jump the clock to the real time after coming back from > hibernation. In this case most of > the TCP timers are going to time out (assuming the hibernation time > is long enough) and > behavior is that of the timers suddenly timing out. > 3. To fast forward the clock, that is actually run the TCP/IP timers > faster, so that all the events > that would have actually been trigerred by the timers during > hibernation would be compressed > in a very small amount of time, possibly resulting in a flurry of > retransmit segments or > keep-alives. > > I wonder how windows deals with the situation, and what could be > relative pros/cons of these > approaches.. > > Can anybody ? > > --Deb > |
|
![]() |
| Outils de la discussion | |
|
|