Subject: NetSolve revision for C:PE special issue From: Jack Dongarra Date: Sun, 30 Dec 2001 11:56:36 -0500 To: Geoffrey Fox CC: Dorian Arnold , Henri Casanova X-UIDL: ~(j!!9B8"!OnE"!"c="! X-Mozilla-Status: 0001 X-Mozilla-Status2: 00000000 Return-Path: Received: from grids.ucs.indiana.edu (localhost [127.0.0.1]) by grids.ucs.indiana.edu (8.10.2+Sun/8.10.2) with ESMTP id g011Bh619005 for ; Mon, 31 Dec 2001 20:11:43 -0500 (EST) Resent-Message-Id: <200201010111.g011Bh619005@grids.ucs.indiana.edu> Delivery-Date: Sun Dec 30 14:37:02 2001 Received: from snorkel.uits.indiana.edu (snorkel.uits.indiana.edu [129.79.6.186]) by grids.ucs.indiana.edu (8.10.2+Sun/8.10.2) with ESMTP id fBUJb1626624 for ; Sun, 30 Dec 2001 14:37:01 -0500 (EST) Received: from moose.cs.indiana.edu (moose.cs.indiana.edu [129.79.254.191]) by snorkel.uits.indiana.edu (8.10.1/8.10.1/IUPO) with ESMTP id fBUJb0j09589 for ; Sun, 30 Dec 2001 14:37:00 -0500 (EST) Received: from localhost (localhost [127.0.0.1]) by moose.cs.indiana.edu (8.11.6/8.11.6/IUCS_2.43) with ESMTP id fBUJb0i01888 for ; Sun, 30 Dec 2001 14:37:00 -0500 (EST) Prev-Resent-Message-Id: <200112301937.fBUJb0i01888@moose.cs.indiana.edu> Replied: Sun, 30 Dec 2001 14:36:41 -0500 Replied: dongarra@cs.utk.edu Received: from cs.utk.edu (cs.utk.edu [160.36.56.56]) by moose.cs.indiana.edu (8.11.6/8.11.6/IUCS_2.43) with ESMTP id fBUH0Ai28194 for ; Sun, 30 Dec 2001 12:00:10 -0500 (EST) Received: from cs.utk.edu (marvin@localhost) by cs.utk.edu with ESMTP (cf v2.9s-UTK) id LAA11961; Sun, 30 Dec 2001 11:56:46 -0500 (EST) Received: from cs.utk.edu (12.64.66.29 -> slip-12-64-66-29.mis.prserv.net) by cs.utk.edu (smtpshim v1.0); Sun, 30 Dec 2001 11:56:47 -0500 Message-ID: <3C2F4744.88FA52D4@cs.utk.edu> Reply-To: dongarra@cs.utk.edu Organization: ICL @ University of Tennessee X-Mailer: Mozilla 4.78 [en] (Windows NT 5.0; U) X-Accept-Language: en MIME-Version: 1.0 Content-Type: multipart/mixed; boundary="------------3215893B06D2EF476B77C0BF" Prev-Resent-To: gcf@indiana.edu Prev-Resent-Date: Sun, 30 Dec 2001 14:37:00 -0500 Prev-Resent-From: Geoffrey Fox Resent-To: ccpe Resent-Date: Mon, 31 Dec 2001 20:11:43 -0500 Resent-From: Geoffrey Charles Fox X-UIDL: ~(j!!9B8"!OnE"!"c="! Geoffrey, Attached is an updated version of the NetSolve paper for the special issue of Concurrency: P&E on Grid based computing. Below are the details of what was done to address the reviewers' comments. Sorry for the delay in returning this back to you. Best wishes, Jack Referee 1 ******************************************* E: Referee Comments (For Author and Editor) ------------------------------ This is a very nice paper which needs just some fixes in its presentation. Some additional comments I would like to see addressed in the paper are: comment: Page 8: There seems to be a flaw with the Proxy Architecture in Netsolve allowing a denial of service attack due to the lack of port protection with credentials or other methods. resolution: added a citation to a paper giving full details about the security integration of Netsolve. comment: Page 9: comment on exceptions, what is if DSI goes down? resolution: added a citation to a paper giving full details about NetSolve/DSIs comment: Page 14: It is not clear if netsolve server is used per user, per compute center, ...... It would be nice to be more precise with the usage scenario. resolution: elaborated/addressed in section "Administrative Issues" comment: Page 15: Please rework the Performance model using the same latter P for the performance and the given problem does not make me believe in your formula [;-)] Maybe you should use the Fox, et.al. book for a sample on how to chose your subscripts. Also your performance work model introduces the constant workload which is unfortunately a function over time. Thus it should probably be w(t). You need to motivate why that performance model is sufficient for your system. It is definitely necessary to put a reference to a thesis/paper/.... explaining why this model does work. Other performance model could be averaging over time thus this culd be _ resolution: NOT addressed. There is another paper entitled "Netbuild" maybe a summary (1 page) of that work could be included in this paper? F: Presentation Changes comment: The use of "figure" in the text must be capitalized "Figure" do not abbreviate to Fig. (page 18) resolution: FIXED comment: The paragraph on page 6 needs clarification if integration of Condor and Legion has ben done or not. Use of the word can is ambiguous. In case it has been done we need more detailsabout this. resoultion: clarified. comment: Page 8 : Figure ??? resolution: fixed. comment: Page 10: do not start sentences with a reference [15] discusses resolution: fixed. comment: Page 14: The headline is not placed correctly. resolution: fixed comment: Page 18: add references to the nuclear engeneering section. Such as CENTS, MCNP, resolution: not addressed Referee 2 ******************************************* E: Referee Comments (For Author and Editor) comment: The description of the NetSolve has appeared in the literature many times. The current draft does not add any significant new information about the system. Already, there are alternatives to Netsolve. Some performance results must be added and NetSolve be compared to other alternatives. resolution: added an empirical evaluation of the system's overhead. comparison not done -- the target publication is a special issue all about GCEs grid computing environments, so ... F: Presentation Changes comment: Can be significantly improved. Existing publications about Nesolve are superior to this presentation. resolution: ??? Referee 3 ******************************************* E. Referee comments for author and editor: comment: The paper was concise, to the point, and fairly complete. It needs some elaboration on security model/implementation: are sessions or connections encrypted, and how? Is Kerberos the only authentication used, if so, are there plans to integrate GSI (note: the term proxy is used with GSI/PKI authentication, so there is some slight confusion here). resolution: elaborated on the current implementation and cited more detailed reference. -- ********************************************************************* Jack Dongarra Innovative Computing Lab Computer Science Dept University of Tennessee 1122 Volunteer Blvd Knoxville TN, 37996-3450 Phone: 865-974-8295 dongarra@cs.utk.edu http://www.cs.utk.edu/~dongarra/