SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


    [Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

    RE: iSCSI: Flow Control



    Hi Doug:
    
    See comments below
    
    > -----Original Message-----
    > From: Douglas Otis [mailto:dotis@sanlight.net]
    > Sent: Friday, October 06, 2000 6:06 PM
    > To: Charles Monia; John Hufferd/San Jose/IBM; IPS@ece.cmu.edu
    > Subject: RE: iSCSI: Flow Control
    > 
    > 
    > Charles,
    > 
    > The problem with the iSCSI control schemes is that it does 
    > not relate to the
    > actual FIFO being controlled.  A dynamic credit scheme as I 
    > indicated in the
    > example draft would allow an accurate tracking of buffer 
    > space and allow
    > adjustment to changes in use to be tracked rather quickly.
    
    I stand by my original posting.  Any dynamic proposals need to be backed up
    by convincing real-world data on behavior.  Even the simplest of such
    schemes are prone to exhibit unexpected instabilities.  To me, this is
    something that will take some trial and error to get right.
    
    > The internal
    > FIFO feeding the medium is the critical parameter being 
    > controlled as you
    > would not want one FIFO to become deep.  Otherwise this would 
    > allow frames
    > to become stale.  It would be better to keep these frames at 
    > the client in
    > the case where the FIFO feeding the medium is congested.  For good
    > operation, all these FIFOs must be kept as lean as possible.  
    > That would
    > imply Class 3 control.  This type of control is already in 
    > use, so you might
    > say the iSCSI scheme has not been tested.
    > 
    > The present iSCSI scheme makes no provisions for such 
    > control.  There are
    > not any meaningful constraints in adhering to the limits 
    > imposed by FCP in
    > limiting to 256 commands.
    
    
    What limits?
    
    If you are referring to the command reference number defined in FCP-2, use
    of that feature is optional and has nothing to do with any sort of
    limitation on the number of pending commands. 
    
    Charles
    
    <snip>
    


Home

Last updated: Tue Sep 04 01:06:46 2001
6315 messages in chronological order