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

RE: [gram-dev] BES



Dear All:

I'd like to see a resource estimate for BES support, along with the related supporting work, so that we can think strategically about the tradeoffs. I do think that several communities are going to be asking for BES support.

Ian.

At 01:07 PM 9/17/2006 -0400, Jarek Gawor wrote:
So is there a decision about the BES interface demo for SC? Because now with this WS-Addressing issue the demo will require two people working on it....

IMHO, if BES will not be in 4.2 in production quality then we should not worry about this demo and instead of focus on the performance issues and on other things that we promised for 4.2.

Jarek

-----Original Message-----
From: owner-gram-dev@xxxxxxxxxx [mailto:owner-gram-dev@xxxxxxxxxx] On Behalf Of Ian Foster
Sent: Sunday, September 17, 2006 12:44 PM
To: Peter G. Lane; Peter G. Lane
Cc: Steve Tuecke; Ravi Madduri; gram-dev@xxxxxxxxxx
Subject: Re: [gram-dev] BES


Right, we need to have our solution for support of multiple interfaces in place.


At 10:40 AM 9/17/2006 -0600, Peter G. Lane wrote:

I just realized that the BES spec causes some of the same problems with GT as does adoption of OGSA WSRF Basic Profile. Specifically, BES uses a different version of WS-Addressing than the GT uses. BES uses "http://www.w3.org/2005/08/addressing" while the GT uses "http://schemas.xmlsoap.org/ws/2004/03/addressing". Unless we upgrade our WS-A implementation, we won't be able to comply with the BES spec.

Peter

Peter G. Lane wrote:

Steve Tuecke wrote:

So you are arguing that we do BES+JSDL as a prototype, and not include either in 4.2?  I would agree with that.  I think adding JSDL to the current GRAM protocol just adds another support and compatibility issue.  Let's stick with the current GRAM interface and job language until BES and JSDL firm up before adopting them as a supported feature of Globus.
This is what we originally agreed upon, but as I understood it there were apparently people pushing for GRAM to have JSDL regardless (it at least made converting to different job descriptions a little easier). Also, part of this, again IIUC, was a fear that people would look elsewhere if we didn't show something with standards. I never really knew who this was, so somebody else will have to elaborate. I think if we participate in this interop demo, that should take care of that. We also have the problem that we've been telling people that JSDL support will be in 4.2. The demo might also make up for this (especially if people can download it and play with it), but I'm not as politically savvy as others here to be certain of this.
Peter


-Steve

On Sep 16, 2006, at 10:00 AM, Peter G. Lane wrote:


Right, one of the reasons we decided against implementing BES for 4.2 was the lack of resource to do it given our other priorities. Even if we do it just as an SC demo, we're still taking valuable developer time away from other things. If we're going to do this, I agree with Ravi that we should not tack on JSDL support to the 4.0 interface. JSDL support was a compromise between getting JSDL support out there soon and implementing all of BES. I still believe just JSDL support is a minor step towards interoperability, so telling people to download the demo code if they absolutely must have JSDL support seems reasonable to me.

Peter

Ravi Madduri wrote:

I agree to this approach. I think we should think hard before we try to implement a changing (?) spec and also on the hook to provide production quality GRAM. We put in considerable amount of effort to make GT4 gram implementation solid and changing some (or most of) it for 4.2 by reimplementing it using BES ( and using JSDL underneath instead of JDD ) should be well thought out. I know we can probably use the same code underneath the hood and just change the service interface but still i think we should just implement a different service as a prototype that supports JSDL/BES and have the current implementation be the production implementation. In the next big release, we can deprecate the current impl and make the JSDL one production quality
My 2c
Steve Tuecke wrote:

We might want to decouple a decision about doing a BES implementation for the SC interop test, from the decision on whether to implement BES in 4.2.  I would argue doing a prototype implementation for the SC interop test make sense.  But I would also argue that we should not rush to put that into 4.2 -- let's keep it in a branch until BES is finished and there are clear indications of adoption, then figure out when to finish it, test it, move it into trunk, and release it in a stable GT.
-Steve

On Sep 15, 2006, at 9:36 AM, Peter Lane wrote:


Yeah, I talked to Chris from Platform about this. I think it comes down to what our priorities. We already decided not to implement BES for 4.2, so doing so requires rethinking what we want to do for 4.2. We should discuss this at our F2F in two weeks or have a phone call after with have a concrete 4.2. feature list ready.

Peter

On Sep 14, 2006, at 10:39 PM, Ian Foster wrote:


Dear All:

I am told that people plan a BES "interop test" at SC. (Actually, I think it may be rather a HPC Profile interop test, i.e., a subset of BES.) The spec won't be finished by then but should be close. It could be politically a good thing for us to participate. We should look at what would be involved.

Ian.




Content-Type: application/x-pkcs7-signature; name="smime.p7s"
Content-Disposition: attachment; filename="smime.p7s"
Content-Description: S/MIME Cryptographic Signature

0?      *?H?÷
 ?0?10  +0? ¬10  +0â?¬     *?H?÷
 ?P0?¦0?? Ã?0
        *?H?÷    *â? Hâ? ÷
0i10    â??&?â??ò,dorg10    â??&?â??ò,dDOEGrids1 s1 0UCertificate Authorities10U
DOEGrids CA 10
051111152456Z
061111152456Z0^10       â??&?â??ò,dorg10    â??&â??°â??ò,ddoegrids10
UPeople10UPeter G Lane 3642430?"0
        *?H?÷
?0??»Ã?â???äÃ??Ã?<¹íM â??Ã?<¹íM â?¹Ã?3¶é=?vU5âHÃ?ñz®w7¼Ah?Ã?Ã?Cñ;?Uâ??d¥°%óÂ?l?lñ?Ã?=­hý??Ã?mE[w"é!Myâ??"c<?2n??[dÿÃ?¢fo®Â¢fo®°5¶an?WAoÂ?Ã?Ã?aé:YÂ??ñbx=Ã?Ã?4Ã?B©?"L5â??@`?Ã?Ã?Ã?Ã?©?Nk2Ã?û?ü°ÿ·Ã??ø·?=D` Ã?¦9æÃ?³ù?oÂ?ÂË?oÂ?¸ðÃ?±Ã?ä1éÃ?¡Ã?Â??F?®r8"[])}QJn>Ã?uï¯j^¸[Ã?ÂÃ?½1¸?Ã?ç5Â?èfãIÃ?oÃ?Ã??~â??Ã?â?¹|CUäÃ?Ã?º?;ù?ùâ?¬Dy£c0a0      `?H?øBà 0Uÿð0U#0?Ã??n¤8]BÃ?1Ã?Ã??1Ã?Ã?Â?
]0U0Â?lane@xxxxxxxxxxxx
        *?H?÷
?|Ã?r
|Ã?r
?¹vçÃ?W29z¾MbÃ?í³þÃ?ãsaTÃ?â??rÃ?x¶ ±<ð<]@Ã?ggzñö#¬ûúÂ?Ã?Ã?â???pxÃ?Ã?¯f^b"Ã?â??cíÂ?cX?>¡Ã?D"D"kPíÂ?òëSÃ?Â??ú¾=æ#Q¼Ã?â?ºÃ¦Ã³hâ?º?é[?½¯´áá0{¡i
õ?âÃ?Ã?ó¨¾?Ã?<â    Â?<ü??W0Æ?Pdæ-±æ-±Ã?-Ã??â
â??ZÂ?}íÃ??*%D6¸=iSpJ|µ0rÃ?i'LKÃ?uOÃ?/l Ã?ccíÂ?Â?Æ?FÃ?*í«Ã?Ã?`c
äÃ?(¯¯â?`â??G®{©æÂ?Ã?>Ã??¾_Ã?÷4¯Ã?Ã?1÷ø ¦¢=ÿ9Ã?Jæð鸾ç0?¦0?? ÃŽ Ã?0
        *?H?÷
0i10    â??&?â??ò,dorg10    â?   â??&?â??ò,dDOEGrids1 0UCertificate Authorities10U
DOEGrrids CA 10
051111152456Z
061111152456Z0^10       â??&?°â??ò,dorg10    â??&?â??ò,ddoegrids10
UPeople10UPeterr G Lane 3642430?"0
        *?H?÷
?0??»Ã?ââ??â??»Ã?â???äÃ??Ã?<¹íM â?¹Ã?3¶é=?vU5âHÃ?ñz®w7¼Ah?Ã?Ã?Ã?Ã?Cñ;?Uâ??d¥°%óÂ?lñ?Ã?=­hý??Ã?mE[w"é!Myâ??"c<?"c<?2n??[dÿÃ?¢fo®°5¶an?WAoÂ?Ã?Ã?aé:YÂ??ñbx=Ã?Ãbx=Ã?Ã?4Ã?B©?"L5â??@`?Ã?Ã?©?Nk2Ã?û?ü°ÿ·Ã??ø·?=ø·Ÿ=D` Ã?¦9æÃ?³ù?oÂ?¸ðÃ?±Ã?ä1éÃ?¡Ã?Â??F?®r8"[r8"[])}QJn>Ã?uï¯j^¸[Ã?½1¸?Ã?ç5Â?èfãIÃ?oÃ?Ã??~â??Ã?ââ?¹|CUäÃ?Ã?º?;ù?Dy£c0a0      `?H?øBà 0Uÿð0UUÿð0U#0?Ã??n¤8]BÃ?1Ã?Ã?Â?
]0U0Â?lane@xxxxxxxxxxxx
          *?H?÷
?|Ã?r
?¹vçÃ?W29z¾MbÃ?í³þÃ?ãsaTÃ?ãsaTÃ?â??rÃ?x¶ ±<ð<]@Ã?gzñö#¬ûúÂ?Ã?Ã?â???pxÃ?Ã?¯f^b"Ã?â??cíÂ?cX?>¡Ã?D"kPíÂ?òëSÃ?Â??ú¾=æ#Q¼Ã?â?ºÃ¦Ãóhâ?º?é[?½¯´á0{¡i
õ?âÃ?Ã?ó¨¾?Ã?<â       Â?<ü??W0Æ?Pdæ-±Ã?-Ã??â
â??ZÂ?}íÃ??*%D6¸=iSpJ|iSpJ|µ0rÃ?i'LKÃ?uOÃ?/l Ã?cíÂ?Â?Æ?FÃ?*í«Ã?Ã?`c
äÃ?(¯¯â?`â??G®{©æÂ?Ã?>Ã?¾_Ã?÷4¯Ã?Ã?1÷ø ¦¢=ÿ9Ã?Jæð°Ã©Â¸Â¾Ã§0?ø0?à 0
        *?H?÷
0u10    â??&â  â??&â?°â??ò,dnet10    â??&?â??ò,dES10UESnet1 0UCertificatee Authorities10UESnet Root CA 10
021205080000Z
080110080000Z0i10       â??&?â??ò,dorg10    â??&?â??ò,dDOEGrEGrids1 0UCertificate Authorities10U
DOEGrids CA 10?"0
         *?H?÷
?0??´õí#a£�>fµ¸�o²?�-�o²š�-jmr¢?�­q÷¼+.��câ,-ÿY"�d
n1Ã?Ã?.kÃ?Ã??Ã??Ã?â??íºdÃ?´ë`Ã?â??? IÃ?wÃ??áQâ??HWKTo@lí²yþúqÂ¥BÃ?Ã?~Â?!t.­7â??«?8ôÃ?5ö^Ã?6ZÃ?í\ö·<­5y9%?þgì²KÅKÅ?Ã?!k´Ã??Ã?Ã??_ú}!Ã?£Ã?ç|%­â??wÃ?íÆ?øÂ?cC¸vâ?vâ???Q[7?Ã?rÃ?oÃ?³µ?QýlW7¾?Ã?Ã??lpk?]-Æ?ðíâ?-Æ?ðíâ??ïcNÃ?ªæÃ?-tö«Ã?°Ã?Z®Æ?Ã?o«<¦¤@ºJ{Mø]Ã?{?êªROÃ?Ã??QZ)ùcw¯?K£Â??0Â?â?º0     `?H?øB?0UÿÃâ?¡0UÿÃ?0UÃ??n¤8]BÃ?1Ã?Ã?Â?
]0U#0?¼]MH/ø5â??Y«\?K>Ã?²Ã?²:ê0Uÿ0ÿ0%U0Â?DOEGrids-CA-1@xxxxxxxxxxxxx
        *? H?÷
?*;b?>Â?3p{)5V#Æ?+±¶(=vÃ?â??53?¡¿(B7â xB7â xQ­{Ã?¤´`-#â??gÃ?Køj1q.ãë?8Ã?Ã?¶Ã?ðRa5Ã???j½?j½BF?¹Ã?Â?ãÃ?]=Ã?£3NQëji_g?kÃ?E´´ªxýÃ?^Ã?¸¿´þ¾BkêÃ?Ã?¹ñî¥@Ã?wè¨(v¶c¶â??JÃ?HÃ?dj?Â?Ã?1rI¥ÿ¿£°ôÃ?¥½Ã?Ã? u<iû²@x1îÃ?Ã?¼¯?;DÃ?â??ê ©Zµkk;­>Ã?z[½TÃ?ï<xG?¸-*Ã?#Ã?)eSÃ?½º ·±?ûþêÃ?(Ã?(ôüÂ?YyÂ?§@éÃ?Gf¢åÿ.#A»1?N0?J0o0i10 â??&?â??ðâ??ò,dorg10    â??&?â??ò,dDOEGrids1 0UCertificate Authoriities10U
DOEGrids CA 1Ã?0 + ?´0   *?H?÷
             1       *?H?÷
0       *?H?÷
        1
060917 060917164010Z0# *?H?÷
        1UÂ?Ã?0YÃ?«f8K?â?ºÂ®Â?üÃ?üÃ?ºâ??0R   *?H?÷
        1E0C0 *?H?÷
0*?Hâ? 0*â? Hâ? ÷
?0
*?H?÷
@0+0
*?H?÷
(0~     +?71q0   +â??71q0o0i10     â??&?â??ò,dorg10    â??&?â??ò,dDOEGrids1 s1 0UCertificate Authorities10U
DOEGrids CA 1Ã?0Â??*?Hâ Hâ? ÷
        1q o0i10        â??&?â??ò,dorg10    â??&&?â??ò,dDOEGrids1 0UCertificate Authorities10U
DOEGridds CA 1Ã?0
        *?H?÷
?w¼?u==8âð=÷·ñ5]·Â·Ã±5]Ã??à ?]8­á¢c$²ªâ????Ã?zO?gÃ??cj·µÂ?&!áÃ?*ÂáÃ?*·Mf¢õ?Ã?¦Â?Y­Ã??  l,ø=¨ã?B
±Ã?®û6Ã?Ã?ø?Ã?øÂ?~Ã?´?í¿hÃ?¢ ?â??çÃ?> &ö<ð+¹PXëÃ?Ã?ºâ¡õKââ??.?6Ã?©ê?³.d?A*O8p?Ã?Ã?¼Â?ÿöÃ?´Ã?Ã?S½Â?ÂS½Â?¡­Z«0)X@¥ç¹Ã?sÃ?¢hð )ôtñ»Ã?£>zùÃ?Z2©eJÃ?½??T=:¡('?é«»Ã?Â?ôÿ{;Ã?ósR<Ã?½¥~Ã?cÆ?VÃ?·94Ã?4Ã?'??½ü´Â?©?4
________________________________________________________________
Ian Foster -- Weblog: http://ianfoster.typepad.com
Computation Institute: www.ci.uchicago.edu & www.ci.anl.gov
Argonne: MCS/221, 9700 S. Cass Ave, Argonne, IL 60439
Chicago: Rm 405, 5640 S. Ellis Ave, Chicago, IL 60637
Tel: +1 630 252 4619 --- Globus Alliance: www.globus.org
     
       

_______________________________________________________________
Ian Foster -- Weblog: http://ianfoster.typepad.com
Computation Institute: www.ci.uchicago.edu & www.ci.anl.gov
Argonne: MCS/221, 9700 S. Cass Ave, Argonne, IL 60439
Chicago: Rm 405, 5640 S. Ellis Ave, Chicago, IL 60637
Tel: +1 630 252 4619 --- Globus Alliance: www.globus.org