From owner-stp@oss.sgi.com Fri Jul 7 14:06:52 2000 Received: by oss.sgi.com id ; Fri, 7 Jul 2000 14:06:42 -0700 Received: from pneumatic-tube.sgi.com ([204.94.214.22]:35413 "EHLO pneumatic-tube.sgi.com") by oss.sgi.com with ESMTP id ; Fri, 7 Jul 2000 14:06:35 -0700 Received: from lhotse.engr.sgi.com (lhotse.engr.sgi.com [163.154.35.41]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id OAA05081 for ; Fri, 7 Jul 2000 14:12:11 -0700 (PDT) mail_from (aman@engr.sgi.com) Received: from engr.sgi.com (localhost [127.0.0.1]) by lhotse.engr.sgi.com (SGI-8.9.3/8.9.3) with ESMTP id OAA01517; Fri, 7 Jul 2000 14:06:14 -0700 (PDT) Message-ID: <39664646.AB6B952@engr.sgi.com> Date: Fri, 07 Jul 2000 14:06:14 -0700 From: Aman Singla Organization: SGI X-Mailer: Mozilla 4.74b2C-SGI [en] (X11; I; IRIX 6.5-ALPHA-1286569120 IP32) X-Accept-Language: en MIME-Version: 1.0 To: bill.main@bigfoot.com CC: Ben Segal , Pekka Pietikainen , Jean-Michel Pittet , stp@oss.sgi.com Subject: Re: looking seriously at STP References: <3965EC74.5D03373B@gis.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-stp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;stp-outgoing > Folks- > We are looking seriously at STP as a jumping off point to the ipSCSI > endeavor. Toward that end we would like to work with you on the > Linux/STP project. It will take a a little time to set up our test bed > and get the same kernel up that you are using. I gather also that you > have twiked a NIC driver, so I think we should use that net card as well > (which one is it?). Though the Linux/STP code will work on most ethernet and/or IP devices without any modification of the driver/NIC (by simulating most STP reqd. functionality in s/w), we've been principally interested in enhancing GbE for STP. Toward this end, we've modified the GbE driver and the firmware to a limited extent. The NIC we're using for our purpose is the Tigon-2 based alteon GbE NIC. A group at CERN is also very interested in running SCSI over STP, and I'm CCing them on this email to introduce your intentions to each other. In the future, it might be useful to have such discussion on the stp mailing list (stp@oss.sgi.com) to have everyone participate in the discussion and be aware of work happening in this dimension. thanks, :a > > After that stage, I will be happy to share the work load with you on > pushing it forward. > > Our near term goal is to get a iSCSI like system up on IP or TCP/IP > to study perfomance issues with general traffic on the same net. > > Regards, > Bill Main > Cambex Corporation From owner-stp@oss.sgi.com Mon Jul 10 05:42:08 2000 Received: by oss.sgi.com id ; Mon, 10 Jul 2000 05:41:49 -0700 Received: from smtp1.cern.ch ([137.138.128.38]:54541 "EHLO smtp1.cern.ch") by oss.sgi.com with ESMTP id ; Mon, 10 Jul 2000 05:41:26 -0700 Received: from cern.ch (IDENT:ben@lxplus003.cern.ch [137.138.161.124]) by smtp1.cern.ch (8.9.3/8.9.3) with ESMTP id OAA24576; Mon, 10 Jul 2000 14:41:13 +0200 (MET DST) X-Authentication-Warning: smtp1.cern.ch: Host IDENT:ben@lxplus003.cern.ch [137.138.161.124] claimed to be cern.ch Message-ID: <3969C469.D69818B0@cern.ch> Date: Mon, 10 Jul 2000 14:41:13 +0200 From: Ben Segal Organization: CERN X-Sender: "Ben Segal" <@smtp.cern.ch> (Unverified) X-Mailer: Mozilla 4.7C-CERN UNIX lxplus003 45 [en] (X11; I; Linux 2.2.12-20smp i686) X-Accept-Language: fr-FR, en MIME-Version: 1.0 To: bill.main@bigfoot.com CC: Aman Singla , Pekka Pietikainen , Jean-Michel Pittet , stp@oss.sgi.com, Arie Van Praag , Les Robertson Subject: Re: looking seriously at STP References: <3965EC74.5D03373B@gis.net> <39664646.AB6B952@engr.sgi.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-stp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;stp-outgoing Hallo Bill, Just to support Aman's answer: we at CERN (myself, Pekka and Arie) are very happy that the ipSCSI and SST ("stSCSI") efforts collaborate as much as possible. With Cambex's IBM connections (via your HBA's) and ipSCSI contacts, you must know the Satran group from IBM Haifa. They visited us here at CERN recently and we spoke then of our hope that ipSCSI and SST got closer together. Ben ================================================================================== Aman Singla wrote: > > > Folks- > > We are looking seriously at STP as a jumping off point to the ipSCSI > > endeavor. Toward that end we would like to work with you on the > > Linux/STP project. It will take a a little time to set up our test bed > > and get the same kernel up that you are using. I gather also that you > > have twiked a NIC driver, so I think we should use that net card as well > > (which one is it?). > > Though the Linux/STP code will work on most ethernet and/or IP devices > without any modification of the driver/NIC (by simulating most STP > reqd. functionality in s/w), we've been principally interested in > enhancing GbE for STP. Toward this end, we've modified the > GbE driver and the firmware to a limited extent. The NIC we're using > for our purpose is the Tigon-2 based alteon GbE NIC. > > A group at CERN is also very interested in running SCSI over STP, > and I'm CCing them on this email to introduce your intentions to > each other. > > In the future, it might be useful to have such discussion on the > stp mailing list (stp@oss.sgi.com) to have everyone participate in > the discussion and be aware of work happening in this dimension. > > thanks, > > :a > > > > > After that stage, I will be happy to share the work load with you on > > pushing it forward. > > > > Our near term goal is to get a iSCSI like system up on IP or TCP/IP > > to study perfomance issues with general traffic on the same net. > > > > Regards, > > Bill Main > > Cambex Corporation -- Ben M. Segal / CERN - IT Division, 1211 Geneva 23, Switzerland. Tel: +41-22 767 4941 / Mobile: +41-79 201 0618 Fax: +41-22 767 7155 WWW: http://home.cern.ch/ben/ **** See you at INET'2000, Yokohama, 18-21 July 2000 **** From owner-stp@oss.sgi.com Tue Jul 11 15:37:34 2000 Received: by oss.sgi.com id ; Tue, 11 Jul 2000 15:37:15 -0700 Received: from smtp1.cern.ch ([137.138.128.38]:57099 "EHLO smtp1.cern.ch") by oss.sgi.com with ESMTP id ; Tue, 11 Jul 2000 15:37:00 -0700 Received: from cern.ch (IDENT:ben@lxplus003.cern.ch [137.138.161.124]) by smtp1.cern.ch (8.9.3/8.9.3) with ESMTP id RAA10059; Tue, 11 Jul 2000 17:39:30 +0200 (MET DST) X-Authentication-Warning: smtp1.cern.ch: Host IDENT:ben@lxplus003.cern.ch [137.138.161.124] claimed to be cern.ch Message-ID: <396B3FB1.2465DBD2@cern.ch> Date: Tue, 11 Jul 2000 17:39:29 +0200 From: Ben Segal Organization: CERN X-Sender: "Ben Segal" <@smtp.cern.ch> (Unverified) X-Mailer: Mozilla 4.7C-CERN UNIX lxplus003 45 [en] (X11; I; Linux 2.2.12-20smp i686) X-Accept-Language: fr-FR, en MIME-Version: 1.0 To: bill.main@bigfoot.com CC: Jean-Michel Pittet , don@genroco.com, Aman Singla , Arie Van Praag , stp@oss.sgi.com, mike@genroco.com Subject: Re: looking seriously at STP References: <3965EC74.5D03373B@gis.net> <39664646.AB6B952@engr.sgi.com> <3969C469.D69818B0@cern.ch> <3969CB64.E7A5C5DE@gis.net> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-stp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;stp-outgoing Bill Main wrote (Mon, 10 Jul 2000 09:11:00 -0400): > > Ben- > Good to hear from you again. I believe we communicated while I was on the Digital > Hippi program some years back. I have just joined the project at Cambex and we are > just getting started. It is my opinion that both STP and ipSCSI must exist because > there are different perfomance problems at different distances to be solved. > Hi Bill, yes we did interact during the Dec Hippi installation phase here. I agree that if SCSI must run over WAN's then ST alone won't be adequate as a carrier and IP needs looking at. But I believe stSCSI (or "SST" as SCSI-over-ST is normally written) is more advanced in the standards and implementation process than ipSCSI. > Given we are in ramp up stage, any help getting caught up to everybody else would > be helpful. In particular a list of equipment you are using on the Linux/STP so that > we can all stay in the same envirnoment. > We and SGI are using Intel boxes (single or multiple CPU) with Tigon 2 GigE adapters (i.e. Alteon, 3COM, Bay Networks cards). > How far along is the project at this point? > ST runs over GigE under Linux, the main implementors being Aman Singla and the SGI team under Jean-Michel Pittet in Mountain View (see the URL http://oss.sgi.com/projects/stp for details). At CERN, I have a Master's student (Pekka Pietikainen) beginning to implement OS Bypass within this Linux ST/GigE tree. So far, under Linux, file systems haven't had their SCSI layer linked up to SST; however, this has been done for file systems on True64, Solaris and Irix platforms, for both GSN and GigE I believe. The leader in this latter work is Genroco - you can contact Don Woelz or Mike Barratt about this. SGI and Compaq have also worked on it and I believe IBM may be active too. Genroco may get into Linux SST some time this year (ask Don), but SGI isn't doing it so far and at CERN we haven't yet committed to it, as Pekka must do the OS Bypass first. > Has anyone worked up a spec for stSCSI yet? > Yes indeed! It's an ANSI standard, developed by the T.10 committee. For all their latest docs on ST, see the site (including a simple ST FAQ) at: http://www.hippi.org/cNEW.html and for SST see: http://www.hippi.org/dSST-r03.pdf > Is there a mailing list for stSCSI? > Well, for Linux ST it's at: http://oss.sgi.com/projects/stp/mail.html > Regards, > Bill Main > ================================================================================== > Ben Segal wrote (Mon, 10 Jul 2000 14:41:13 +0200): > > > Hallo Bill, > > > > Just to support Aman's answer: we at CERN (myself, Pekka and Arie) are > > very happy that the ipSCSI and SST ("stSCSI") efforts collaborate as > > much as possible. > > > > With Cambex's IBM connections (via your HBA's) and ipSCSI contacts, you > > must know the Satran group from IBM Haifa. They visited us here at CERN > > recently and we spoke then of our hope that ipSCSI and SST got closer > > together. > > > > Ben > > > > ================================================================================== > > > Bill Main wrote: > > > > > > > Folks- > > > > We are looking seriously at STP as a jumping off point to the ipSCSI > > > > endeavor. Toward that end we would like to work with you on the > > > > Linux/STP project. It will take a a little time to set up our test bed > > > > and get the same kernel up that you are using. I gather also that you > > > > have twiked a NIC driver, so I think we should use that net card as well > > > > (which one is it?). > > > > > Aman Singla wrote (Fri, 07 Jul 2000 14:06:14 -0700): > > > > > Though the Linux/STP code will work on most ethernet and/or IP devices > > > without any modification of the driver/NIC (by simulating most STP > > > reqd. functionality in s/w), we've been principally interested in > > > enhancing GbE for STP. Toward this end, we've modified the > > > GbE driver and the firmware to a limited extent. The NIC we're using > > > for our purpose is the Tigon-2 based alteon GbE NIC. > > > > > > A group at CERN is also very interested in running SCSI over STP, > > > and I'm CCing them on this email to introduce your intentions to > > > each other. > > > > > > In the future, it might be useful to have such discussion on the > > > stp mailing list (stp@oss.sgi.com) to have everyone participate in > > > the discussion and be aware of work happening in this dimension. > > > > > > thanks, > > > > > > :a > > > > > > > > > > > After that stage, I will be happy to share the work load with you on > > > > pushing it forward. > > > > > > > > Our near term goal is to get a iSCSI like system up on IP or TCP/IP > > > > to study perfomance issues with general traffic on the same net. > > > > > > > > Regards, > > > > Bill Main > > > > Cambex Corporation > > Ben -- Ben M. Segal / CERN - IT Division, 1211 Geneva 23, Switzerland. Tel: +41-22 767 4941 / Mobile: +41-79 201 0618 Fax: +41-22 767 7155 WWW: http://home.cern.ch/ben/ **** See you at INET'2000, Yokohama, 18-21 July 2000 **** From owner-stp@oss.sgi.com Tue Jul 11 16:08:53 2000 Received: by oss.sgi.com id ; Tue, 11 Jul 2000 16:08:43 -0700 Received: from genroco.com ([205.254.195.202]:1542 "EHLO gifw.genroco.com") by oss.sgi.com with ESMTP id ; Tue, 11 Jul 2000 16:08:41 -0700 Received: from gi2.genroco.com (gi2.genroco.com [192.133.120.3]) by gifw.genroco.com (8.9.3/8.9.3) with ESMTP id LAA11663; Tue, 11 Jul 2000 11:01:21 -0500 Received: from don (don.genroco.com [192.133.120.101]) by gi2.genroco.com (8.9.3/8.9.3) with SMTP id LAA17180; Tue, 11 Jul 2000 11:01:16 -0500 From: "Don Woelz" To: , Cc: , , "Arie Van Praag" , "Aman Singla" , "Jean-Michel Pittet" Subject: RE: looking seriously at STP Date: Tue, 11 Jul 2000 11:01:15 -0500 Message-ID: <001201bfeb51$3e22a740$657885c0@don.genroco.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook 8.5, Build 4.71.2173.0 X-MimeOLE: Produced By Microsoft MimeOLE V4.72.3155.0 In-Reply-To: <396B3FB1.2465DBD2@cern.ch> Importance: Normal Sender: owner-stp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;stp-outgoing Hello Bill, Good to hear from you again. I see you are no longer with the CSS group that used to be in Merrimack. > Hi Bill, yes we did interact during the Dec Hippi installation phase > here. I agree that if SCSI must run over WAN's then ST alone won't be > adequate as a carrier and IP needs looking at. But I believe stSCSI (or > "SST" as SCSI-over-ST is normally written) is more advanced in the > standards and implementation process than ipSCSI. Yes. I am the editor of the SST standard. Most of the work on the standard at this time is editorial, not technical. We expect to be able to forward the standard for public review this fall. > So far, under Linux, file systems haven't had their SCSI layer linked up > to SST; however, this has been done for file systems on True64, Solaris > and Irix platforms, for both GSN and GigE I believe. The leader in this > latter work is Genroco - you can contact Don Woelz or > Mike Barratt about this. SGI and Compaq have also > worked on it and I believe IBM may be active too. Genroco may get into > Linux SST some time this year (ask Don), but SGI isn't doing it so far > and at CERN we haven't yet committed to it, as Pekka must do the OS > Bypass first. We have indeed targeted Linux (and NT) for SST. Your time estimate is as good as I can do...later this year. > > Is there a mailing list for stSCSI? We use the ANSI T11.1 HIPPI committee's mailing list. To subscribe to the e-mail reflector send a message: Send to: majordomo@network.com Message body: subscribe hippi The subject line is ignored. Avoid adding a signature line(s) if possible. To send to everyone on the reflector, send e-mail to: hippi@network.com. Best regards, Don Donald D. Woelz Tel: 262-644-2505 V. P. of Marketing Mobile: 414-732-7705 GENROCO, Inc. Toll Free: 800-243-6762 255 Info Hwy Fax: 262-644-6667 Slinger, WI 53086 From owner-stp@oss.sgi.com Wed Jul 19 15:30:37 2000 Received: by oss.sgi.com id ; Wed, 19 Jul 2000 15:30:27 -0700 Received: from deliverator.sgi.com ([204.94.214.10]:3932 "EHLO deliverator.sgi.com") by oss.sgi.com with ESMTP id ; Wed, 19 Jul 2000 15:30:05 -0700 Received: from nodin.corp.sgi.com (nodin.corp.sgi.com [192.26.51.193]) by deliverator.sgi.com (980309.SGI.8.8.8-aspam-6.2/980310.SGI-aspam) via ESMTP id PAA04937 for ; Wed, 19 Jul 2000 15:22:10 -0700 (PDT) mail_from (aman@engr.sgi.com) Received: from lhotse.engr.sgi.com (lhotse.engr.sgi.com [163.154.35.41]) by nodin.corp.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via ESMTP id PAA43319 for ; Wed, 19 Jul 2000 15:29:10 -0700 (PDT) Received: from engr.sgi.com (localhost [127.0.0.1]) by lhotse.engr.sgi.com (SGI-8.9.3/8.9.3) with ESMTP id PAA29562; Wed, 19 Jul 2000 15:27:27 -0700 (PDT) Message-ID: <39762B4F.8A6CAE7B@engr.sgi.com> Date: Wed, 19 Jul 2000 15:27:27 -0700 From: Aman Singla Organization: SGI X-Mailer: Mozilla 4.74b2C-SGI [en] (X11; I; IRIX 6.5-ALPHA-1286569120 IP32) X-Accept-Language: en MIME-Version: 1.0 To: Ian Pratt CC: stp@oss.sgi.com, Naveen RatnaSabapathy , Jean-Michel Pittet , stp@engr.sgi.com Subject: Re: zero copy and early demultiplex on Alteon gigabit ethernet ca rds References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-stp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;stp-outgoing Hi Ian- I'm very impressed and interested in the work that you guys have done. I've been working with a protocol called Scheduled Transfers (ST) and doing some work similar to what you guys have done using the alteon GbE card on linux platforms. You can look at http://oss.sgi.com/projects/stp to get more information regarding ST. I'm particularly impressed by the UDP payload throughput of 956Mb/s for 1480 bytes. In my experience I found the alteon firmware loop times to be closer to 12-16us for 1500byte frames. Did you guys do the firmware from scratch? The second question I had was: how easy/difficult is to run a different protocol (stack) using the modified firmware by you guys. thanks, :a Ian Pratt wrote: > > > I believe that Ian Pratt did something along these lines for a project at > > the University of Cambridge. He may have some specific insights with this. > > I am not sure whether he did this work in the context of Linux, but I do > > know that he did this for another OS (Nemisis) that was developed at the U. > > of Cambridge. > > Sorry to take sooo long to reply to this. > > We developed some new firmware for the Alteon NIC last year, and > wrote appropriate Linux drivers and user-space protocol stacks. > > Our firmware has the following features: > 32+ `virtual interfaces' > each virtual interface has its own TX/RX descriptor rings > applications register pages with OS for use for DMA > early packet demultiplex using packet filters uploaded by OS > RX packet filter ensures correct header/payload split into user buffers > transmit packet filters for verifying application's packets before TX > TX traffic shaping with token buckets and round-robin MUX'ing > > We then integrated the `user-accessible' network access into > Linux, enabling applications to use protocol stacks implemented > in shared-libraries (we ported Linux's TCP/UDP stacks to > user-space to enable a fair performance comparison). The RX/TX > packet filters and TX traffic shaping enable the operating system > to remain in control of what packets an application is allowed to > send and receive. > > Performance is good. We get 974Mb/s TCP payload data throughput > with 8192 octet MSS, and 796Mb/s for 1472. UDP payload data > throughput is 992Mb/s for 8192 MSS, 956Mb/s for > 1480. Non-blocking app-to-app ping-pong latency is 45us. > > The following paper has more details: > http://www.cl.cam.ac.uk/~iap10/gige.ps > > Hope this is useful, > Ian From owner-stp@oss.sgi.com Thu Jul 20 12:04:15 2000 Received: by oss.sgi.com id ; Thu, 20 Jul 2000 12:04:05 -0700 Received: from pneumatic-tube.sgi.com ([204.94.214.22]:19739 "EHLO pneumatic-tube.sgi.com") by oss.sgi.com with ESMTP id ; Thu, 20 Jul 2000 12:03:43 -0700 Received: from nodin.corp.sgi.com (nodin.corp.sgi.com [192.26.51.193]) by pneumatic-tube.sgi.com (980327.SGI.8.8.8-aspam/980310.SGI-aspam) via ESMTP id MAA00044 for ; Thu, 20 Jul 2000 12:09:02 -0700 (PDT) mail_from (aman@engr.sgi.com) Received: from lhotse.engr.sgi.com (lhotse.engr.sgi.com [163.154.35.41]) by nodin.corp.sgi.com (980427.SGI.8.8.8/980728.SGI.AUTOCF) via ESMTP id MAA48637 for ; Thu, 20 Jul 2000 12:02:50 -0700 (PDT) Received: from engr.sgi.com (localhost [127.0.0.1]) by lhotse.engr.sgi.com (SGI-8.9.3/8.9.3) with ESMTP id MAA23306 for ; Thu, 20 Jul 2000 12:01:10 -0700 (PDT) Message-ID: <39774C75.14090A39@engr.sgi.com> Date: Thu, 20 Jul 2000 12:01:09 -0700 From: Aman Singla Organization: SGI X-Mailer: Mozilla 4.74b2C-SGI [en] (X11; I; IRIX 6.5-ALPHA-1286569120 IP32) X-Accept-Language: en MIME-Version: 1.0 To: stp@oss.sgi.com Subject: Release 0.32 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-stp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;stp-outgoing Hi- I've just made another release of the ST support for linux. This release is primarily for libst / OSBypass functionality. The README file at the release site explains what you CAN (not much!) and CANNOT do with libst for now.. Comments/questions can be addressed to the mailing list. As always: http://oss.sgi.com/projects/stp thanks, :a From owner-stp@oss.sgi.com Thu Jul 27 07:27:59 2000 Received: by oss.sgi.com id ; Thu, 27 Jul 2000 07:27:50 -0700 Received: from smtp1.cern.ch ([137.138.128.38]:8206 "EHLO smtp1.cern.ch") by oss.sgi.com with ESMTP id ; Thu, 27 Jul 2000 07:27:19 -0700 Received: from cern.ch (IDENT:ben@lxplus003.cern.ch [137.138.161.124]) by smtp1.cern.ch (8.9.3/8.9.3) with ESMTP id QAA26691; Thu, 27 Jul 2000 16:26:33 +0200 (MET DST) X-Authentication-Warning: smtp1.cern.ch: Host IDENT:ben@lxplus003.cern.ch [137.138.161.124] claimed to be cern.ch Message-ID: <39804698.FEEB73C3@cern.ch> Date: Thu, 27 Jul 2000 16:26:32 +0200 From: Ben Segal Organization: CERN X-Sender: "Ben Segal" <@smtp.cern.ch> (Unverified) X-Mailer: Mozilla 4.7C-CERN UNIX lxplus003 45 [en] (X11; I; Linux 2.2.12-20smp i686) X-Accept-Language: fr-FR, en MIME-Version: 1.0 To: stp@oss.sgi.com, Don Woelz , Stephen Bailey , Roger Ronald , Wolfgang Hoschek , Pekka Pietikainen , Arie Van Praag , greg@tspan.com Subject: BXXP Protocol - a fit with STP? Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-stp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;stp-outgoing Hi STP enthusiasts, I've just learned about BXXP (the "Blocks eXtensible eXchange Protocol" recently proposed by Marshall Rose). In Marshall's words "BXXP is a generic application protocol framework for connection-oriented, asynchronous request/response interactions. The framework permits multiplexing of independent request/response streams over a single transport connection, supporting both textual and binary messages...". As I read about it, the similarity of many of its concepts with STP were compelling. Currently BXXP runs over TCP but it seems a natural to run it over STP too. I'd like to know if any STP'ers already knew about BXXP and if there has been contact yet with Rose or his followers. The next step would be to ask the BXXP'ers if they know about STP... For an introduction to BXXP, look at: http://mappa.mundi.net/Internet-Drafts/blocks-appldesign.html and follow links therein. Ben -- Ben M. Segal / CERN - IT Division, 1211 Geneva 23, Switzerland. Tel: +41-22 767 4941 / Mobile: +41-79 201 0618 Fax: +41-22 767 7155 WWW: http://home.cern.ch/ben/ **** See you at INET'2000, Yokohama, 18-21 July 2000 **** From owner-stp@oss.sgi.com Fri Jul 28 08:34:40 2000 Received: by oss.sgi.com id ; Fri, 28 Jul 2000 08:34:20 -0700 Received: from mta1.cl.cam.ac.uk ([128.232.0.15]:19984 "EHLO wisbech.cl.cam.ac.uk") by oss.sgi.com with ESMTP id ; Fri, 28 Jul 2000 08:34:09 -0700 Received: from striker.cl.cam.ac.uk ([128.232.8.39] helo=cl.cam.ac.uk ident=iap10) by wisbech.cl.cam.ac.uk with esmtp (Exim 3.092 #1) id 13IC8z-0003oA-00; Fri, 28 Jul 2000 16:33:49 +0100 To: Aman Singla cc: Ian Pratt , stp@oss.sgi.com, Naveen RatnaSabapathy , Jean-Michel Pittet , stp@cthulhu.engr.sgi.com Subject: Re: zero copy and early demultiplex on Alteon gigabit ethernet ca rds In-reply-to: Your message of "Wed, 19 Jul 2000 15:27:27 PDT." <39762B4F.8A6CAE7B@engr.sgi.com> Date: Fri, 28 Jul 2000 16:33:49 +0100 From: Ian Pratt Message-Id: Sender: owner-stp@oss.sgi.com Precedence: bulk Return-Path: X-Orcpt: rfc822;stp-outgoing > You can look at http://oss.sgi.com/projects/stp to get > more information regarding ST. Thanks for the pointer. It's about time we had some new protocols... > I'm particularly impressed by the UDP payload throughput > of 956Mb/s for 1480 bytes. In my experience I found the > alteon firmware loop times to be closer to 12-16us for > 1500byte frames. Did you guys do the firmware from scratch? We completely re-wrote the send and receive path. > The second question I had was: how easy/difficult is to run > a different protocol (stack) using the modified firmware > by you guys. Easy. The firmware is entirely agnostic as to the protocol being used -- all the demuxing and packet validation is done with a generic packet filtering mechanism. Ian