Author: adrian.chadd
Date: Sun Jun  7 01:12:52 2009
New Revision: 14087

Modified:
    branches/LUSCA_HEAD/src/cf.data.pre
    branches/LUSCA_HEAD/src/structs.h

Log:
Start fleshing out a configurable knob for setting the download speed from  
upstream servers.
The option will default to "download at the slowest speed" which is the  
current behaviour.

This commit doesn't include the behavioural modification; just the config  
knob.



Modified: branches/LUSCA_HEAD/src/cf.data.pre
==============================================================================
--- branches/LUSCA_HEAD/src/cf.data.pre (original)
+++ branches/LUSCA_HEAD/src/cf.data.pre Sun Jun  7 01:12:52 2009
@@ -5976,6 +5976,22 @@
        This option currently does nothing.
  DOC_END

+NAME: download_fastest_client_speed
+COMMENT: whether to download the object at the rate of the slowest or  
fastest client
+TYPE: onoff
+LOC: Config.onoff.download_fastest_client_speed
+DEFAULT: off
+DOC_START
+       The default behaviour of the read-ahead code is to limit the read speed
+       to that of the slowest client. In many situations this may be desirable
+       but in some situations the opposite may be more desirable. If this
+       is turned on then Lusca will pace the store read at the speed of the
+       fastest disk client. It also has the side-effect of disabling the
+       read-ahead gap logic if all clients are disk clients.
+
+       If a client is a memory client (ie, data is being buffered in memory
+       for them) then the read-ahead gap logic still applies.
+DOC_END


  EOF

Modified: branches/LUSCA_HEAD/src/structs.h
==============================================================================
--- branches/LUSCA_HEAD/src/structs.h   (original)
+++ branches/LUSCA_HEAD/src/structs.h   Sun Jun  7 01:12:52 2009
@@ -683,6 +683,7 @@
        int WIN32_IpAddrChangeMonitor;
        int load_check_stopen;
        int load_check_stcreate;
+       int download_fastest_client_speed;
      } onoff;
      int collapsed_forwarding_timeout;
      acl *aclList;

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"lusca-commit" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/lusca-commit?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to