From mboxrd@z Thu Jan  1 00:00:00 1970
From: Dmitry Yusupov <dmitry_yus@yahoo.com>
Subject: Re: [Iscsitarget-devel] Re: [ANNOUNCE] iSCSI enterprise target
	software
Date: Tue, 01 Mar 2005 21:21:51 -0800
Message-ID: <1109740911.3958.4.camel@mylaptop>
References: <20050301161951K.fujita.tomonori@lab.ntt.co.jp>
	 <1109666441.6293.62.camel@laptopd505.fenrus.org>
	 <20050301183549J.fujita.tomonori@lab.ntt.co.jp>
	 <20050302140413R.fujita.tomonori@lab.ntt.co.jp>
Mime-Version: 1.0
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Received: from smtp106.mail.sc5.yahoo.com ([66.163.169.226]:63325 "HELO
	smtp106.mail.sc5.yahoo.com") by vger.kernel.org with SMTP
	id S262179AbVCBFWE (ORCPT <rfc822;linux-scsi@vger.kernel.org>);
	Wed, 2 Mar 2005 00:22:04 -0500
In-Reply-To: <20050302140413R.fujita.tomonori@lab.ntt.co.jp>
Sender: linux-scsi-owner@vger.kernel.org
List-Id: linux-scsi@vger.kernel.org
To: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
Cc: arjan@infradead.org, linux-scsi@vger.kernel.org, iscsitarget-devel@lists.sourceforge.net

On Wed, 2005-03-02 at 14:04 +0900, FUJITA Tomonori wrote:
> From: FUJITA Tomonori <fujita.tomonori@lab.ntt.co.jp>
> Subject: [Iscsitarget-devel] Re: [ANNOUNCE] iSCSI enterprise target software
> Date: Tue, 01 Mar 2005 18:35:49 +0900
> 
> > The last reason is that user-space cost like memory copy. With 1Gbs
> > Ethernet, is is not critical. However, with 10G, it is critical, I
> > expect. I've been setting up 10G experimental infrastructure to
> > evaluate iSCSI performance.
> 
> If we try to build high-performance iSCSI target software on 10Gbs
> Ethernet, I think that we need to implement it in the kernel
> space, although this topic is still in the research stage.
> 
> IICR, Intel provides open-source iSCSI target software, which uses
> mmap() in the user space, as you suggested. However, they chose a
> different approach to implement iSCSI target software for 10Gbs. They
> modified the Linux kernel TCP stack to fully integrate iSCSI
> functionality with it. It is still not clear how to get the best
> performance out of 10GBs Ethernet, however, it seems that using just
> the socket interface is not good enough. So I think that we need to
> control all the system-resources to do it.
> 
> We've not optimized our code yet. The current code simply uses the
> socket interface, however, moving some functionality to the TCP stack
> can improve the performance, although I don't plan to modify theu TCP
> stack. A possible approach to do it is exploiting sk_data_ready().

www.open-iscsi.org project uses this technique on receiving path and
shows very nice performance numbers for Read operations comparable with
HW accelerators.

To add such a functionality to an existing project will probably require
total re-design.

Regards,
Dima