SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


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

    RE: TCP limitations (was Re: ISCSI: Urgent Flag requirementviolates TCP.)



    By the way, you do want to avoid IP layer fragmentation if
    at all possible (fortunately pretty much everyone does
    MTU path discovery now - UDP still causes fragmentation).
    It really adds a lot more processing burden and complexity.
    
    Somesh
    
    > -----Original Message-----
    > From: Randall R. Stewart [mailto:randall@stewart.chicago.il.us]
    > Sent: Tuesday, November 21, 2000 12:25 PM
    > To: ronald lee
    > Cc: ips@ece.cmu.edu
    > Subject: Re: TCP limitations (was Re: ISCSI: Urgent Flag
    > requirementviolates TCP.)
    > 
    > 
    > Ronald:
    > 
    > Comments below...
    > 
    > 
    > ronald lee wrote:
    > > 
    > > I was wondering, is it possible to use TCP option field to
    > > do the framing.  This will require a change to TCP protocol
    > > to handle this option,
    > > but for implementations that want to get the higher performance
    > > they can make this change, else stick with what is there.
    > > Just a suggestion.
    > > 
    > 
    > I believe this is clearly OUTSIDE the scope of the WG. We CANNOT
    > change TCP.. period... if you desire to do this, you must take
    > this to the end2end group (as Scott suggested)...
    > 
    > 
    > > Also, is defragging IP packets any issue with iSCSI hardware
    > > implementations ?  We at MaxStrat did a hardware to accelerate
    > > TCP/IP over HIPPI and I had to deal with fragmented IP packets which
    > > caused
    > > a change in the hardware architecture.  Buffer management is a
    > > big issue when trying to accelerate TCP/IP with hardware.  What
    > > happens if a fragment is missing and lets say many IP packets
    > > have a missing fragment or fragments that arrives late.  We might
    > > get our buffers filled with fragments.
    > 
    > Hmm, seems to me if you implement P-MTU discovery this problem would
    > go away, but I do agree that if there is no P-MTU discovery you may
    > have a problem with fragments....
    > 
    > 
    > R
    > 
    > > 
    > > -- ron
    > 
    > -- 
    > 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 01:06:20 2001
6315 messages in chronological order