SORT BY:

LIST ORDER
THREAD
AUTHOR
SUBJECT


SEARCH

IPS HOME


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

    Re: iSCSI read/write cost difference




    Did the input data aggregate the final status with data or send it separately.
    If the response was sent separately that may explain why short writes perform better as three interrupts are involved in a short read as opposed to 2 interrupts for writes (I assume you use immediate data).

    Julo


    mingz <mingz@ele.uri.edu>
    Sent by: owner-ips@ece.cmu.edu

    17/04/03 20:03

    Please respond to
    mingz@ele.uri.edu

    To
    "Anshul Chadda" <anshul.chadda@trebia.com>
    cc
    ips@ece.cmu.edu
    Subject
    Re: iSCSI read/write cost difference





    InitialR2T=No , ImmediateData=Yes , and i am not sure about the answer for
    last question.

    under is all the values.
    -----------------------------------------
    ../common/iscsi_manage init restore host=$1
    ../common/iscsi_manage init force v=1800 host=$1
    ../common/iscsi_manage init set
    TargetName=iqn.2002-10.edu.unh.iol.iscsi.draft18-target_$2 host=$1
    ../common/iscsi_manage init set
    InitiatorName=iqn.1999-11.edu.unh.iol.iscsi.draft18-initiator host=$1
    ../common/iscsi_manage init set InitiatorAlias=UNH-Draft18-Initiator host=$1
    ../common/iscsi_manage init set MaxConnections=4 host=$1
    ../common/iscsi_manage init set HeaderDigest=CRC32C,None host=$1
    ../common/iscsi_manage init set DataDigest=CRC32C,None host=$1
    ../common/iscsi_manage init set InitialR2T=No host=$1
    ../common/iscsi_manage init set ImmediateData=Yes host=$1
    ../common/iscsi_manage init set MaxRecvDataSegmentLength=12288 host=$1
    ../common/iscsi_manage init set FirstBurstLength=16384 host=$1
    ../common/iscsi_manage init set MaxBurstLength=32768 host=$1
    ../common/iscsi_manage init set OFMarker=No host=$1
    ../common/iscsi_manage init set IFMarker=Yes host=$1
    ../common/iscsi_manage init set IFMarkInt=4096~32768 host=$1
    ../common/iscsi_manage init set OFMarkInt=1024~8192 host=$1
    ../common/iscsi_manage init set DefaultTime2Wait=10 host=$1
    ../common/iscsi_manage init set DefaultTime2Retain=6 host=$1
    ../common/iscsi_manage init set MaxOutstandingR2T=4 host=$1
    ../common/iscsi_manage init set DataPDUInOrder=No host=$1
    ../common/iscsi_manage init set DataSequenceInOrder=No host=$1
    ../common/iscsi_manage init set ErrorRecoveryLevel=0 host=$1
    ../common/iscsi_manage init snack d=y host=$1
    ../common/iscsi_manage init snack s=y host=$1
    ../common/iscsi_manage init force i=2 host=$1
    ../common/iscsi_manage init force p=2 host=$1


    On Thursday 17 April 2003 12:39, you wrote:
    > What are the negotiated values of InitialR2T and ImmediateData in your
    > setup? Also, does the target piggyback SCSI Response in the Final Datain
    > pdu?
    >
    >
    > -----Original Message-----
    > From: owner-ips@ece.cmu.edu [mailto:owner-ips@ece.cmu.edu]On Behalf Of
    > mingz
    > Sent: Thursday, April 17, 2003 12:10 PM
    > To: ips@ece.cmu.edu
    > Subject: iSCSI read/write cost difference
    >
    >
    > I recently did some experiments on iSCSI benchmarking. I used the UNH iSCSi
    > ref18_03 implementation and use a ram based scsi device for iscsi target
    > use.
    > so there is no any disk overhead.
    >
    > I used a linux kernel module to send fixed size read/write requests to
    > iscsi initiator, which eventually will be filled by iscsi target. i use
    > interl pro1000 gigabit nic and intel 470 gigabit switch. and both initiator
    > and target use same type piii 866 pc with 1g ram.
    >
    > now the IOPS result shows that
    >
    >                                   2k                                  4k                                  8k                                  16k                                  32k
    > read                                  1975                 1781                 1557                 1081                 767
    > write                  2877                 2225                 1588                 812                                  533
    >
    > for small reuqests, read is slower than write, while for large requests,
    > read
    > is faster than write. i redo the experiments on another enviroment, also
    > get similar results.
    >
    > can anybody explain why this happens?
    >
    > thanks a lot.
    >
    >
    > ming

    --
    --------------------------------------------------
    | Ming Zhang, PhD. Student                                    
    | Dept. of Electrical & Computer Engineering      
    | College of Engineering                          
    | University of Rhode Island                      
    | Kingston RI. 02881                                  
    | e-mail: mingz@ele.uri.edu                        
    | Tel. (401) 874-2293 Fax  (401) 782-6422          
    | http://www.ele.uri.edu/~mingz                    
    --------------------------------------------------



Home

Last updated: Fri Apr 18 14:19:29 2003
12534 messages in chronological order