SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


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

    Re: Symmetric vs Asymmetric



    Black_David@emc.com wrote:
    > 
    > > Since I have seen no one else that made arguments for the Symmetric
    > > implementation, you maybe able to call the agreement on the Asymmetric
    > > approach a consensus.
    > 
    > While I appreciate the invitation, I'm going to hold off on doing that.
    > The current state of the discussion doesn't feel like consensus (rough
    > or otherwise) to me yet, and it seems to me that there have been
    > people other than John advocating the Symmetric model.
    > 
    > I do want to kill off a red herring that found its way into the discussion.
    > Squeezing additional bandwidth out of a single network link (HTTP-style)
    > has not been an important motivation for multiple connections.  
    
    I guess I in part introduced this red herring... :0  sorry about that!
    But
    I am very glad to hear it is dead!! It kill one of my big concerns :)
    
    > The aims
    > of multiple connections have been to use parallelism in the network for
    > failover and load balancing.  A few weeks ago, there was rough consensus
    
    You can easily get the failover from SCTP. It will automatically retry
    on
    alternate addresses if they exist... on the first retran.
    
    So SCTP could be used to get:
    
    A) CC
    B) Fail-over from one NIC to another on retransmissions
    C) Limited retransmission (i.e. only retransmit 0 or 1 times, for 
                               those custom configured networks where
                               loss should not happen... doing 1 retran
                               looks good here since it would go on an
                               alternate NIC)
    D) Un-ordered delivery
    C) Multiple conversations seperated in the Streams without
       blocking between the streams.
    
    It can NOT help you with
    
    1) Load balancing
    
    
    R
    
    > that use of multiple connections to squeeze additional bandwidth out of
    > single network would not be a requirement on iSCSI, and I haven't seen
    > anything that leads me to believe that this consensus has changed.
    > 
    > Progress is being made -- carry on ...
    > 
    > --David
    > 
    > ---------------------------------------------------
    > David L. Black, Senior Technologist
    > EMC Corporation, 42 South St., Hopkinton, MA  01748
    > +1 (508) 435-1000 x75140     FAX: +1 (508) 497-8500
    > black_david@emc.com       Mobile: +1 (978) 394-7754
    > ---------------------------------------------------
    
    -- 
    Randall R. Stewart
    randall@stewart.chicago.il.us or rrs@cisco.com
    815-342-5222 (cell) 815-477-2127 (work)
    


Home

Last updated: Tue Sep 04 01:07:30 2001
6315 messages in chronological order