June 2010 - pgi-wg - lists.ogf.org

Re: [Pgi-wg] PGI Telcon 2010-05-20: Requirements List Spreadsheet - OGF29 + PGI F2F in Chicago - Next telcon
by Etienne URBAH 09 Jun '10

09 Jun '10

Morris, Concerning the OGF PGI telephone conference held on 20 Mai 2010 : - Thank you for leading the discussion. We made much progress. - Could you publish in GridForge the updated MS-Excel source file of the Condensed Requirements List ? - OGF29 + PGI F2F in Chicago : Can you confirm that we can book our travel back home for Friday 25 June 2010 ? - For the next PGI telephone conference : - I heard that there will be NONE next week. - On Thursday 03 June 2010, it would be difficult for me to attend after 16h CET, but I will attend until 17h CET if no other schedule is possible. - Could you propose a date and time ? Thank you in advance. Best regards. ----------------------------------------------------- Etienne URBAH LAL, Univ Paris-Sud, IN2P3/CNRS Bat 200 91898 ORSAY France Tel: +33 1 64 46 84 87 Skype: etienne.urbah Mob: +33 6 22 30 53 27 mailto:urbah@lal.in2p3.fr ----------------------------------------------------- On Thu, 20/05/2010 15:10, Morris Riedel wrote: > Dear PGI team, > > just a short reminder for our meeting today at 3:30 pm CET: > > > (1) > TelCon System using HiDef (Attention: New Number! Not the old PGI one) > > Free of charge when calling from Skype: > +9900827045746500 > > International Numbers: > United States +1 (201) 793-9022 > Canada +1 (201) 793-9022 > Austria +43 (0) 82040115470 > Belgium +32 (0) 7 0357134 > France +33 (0) 826109071 > Germany +49 01805009527 > Ireland +353 (0) 818270968 > Italy +39 848390177 > Spain +34 (9) 02885791 > Switzerland +41 (0) 8 48560397 > United Kingdom +44 (0) 8454018081 > > Room number: 5746500 > > No costs from Skype and nothing to install on the computer. > > > (2) > Screen-Sharing (Voice Backup) > > http://webconf.vc.dfn.de/ogf-pgi-wg/ > > No login is required. Just enter as 'guest' with your real world name. > > > > (3) > Proposed agenda: > > (a) > Sessions at OGF in Chicago > > (b) > Requirement List > > (c) > AOB > > > Please feel free to add other items on the agenda as you like. > > > A high participation from the PGI and GIN members would be appreciated. > > Your co-chair, > Morris > > ------------------------------------------------------------ > Morris Riedel > Distributed Systems and Grid Computing Division > > Jülich Supercomputing Centre (JSC) > Info: http://www.fz-juelich.de/jsc/JSCPeople/riedel

1 2

Call tomorrow
by Johannes Watzl 09 Jun '10

09 Jun '10

Dear Morris, all, I can't make the call tomorrow - I will be at the SIENA kick-off meeting. Best, Johannes -- _ _ _ _ _ _ Johannes Watzl |\/| |\ | |\/| Institut für Informatik / Dept. of CS | | | \| | | Ludwig-Maximilians-Universität München ======= TEAM ======= Oettingenstr. 67, 80538 Munich, Germany Room E 005, Phone +49-89-2180-9162 Munich Network Management Team Email: watzl(a)nm.ifi.lmu.de Münchner Netz-Management Team http://www.nm.ifi.lmu.de/~watzl

1 0

OGF - Useful Official and De facto Standards
by Etienne URBAH 08 Jun '10

08 Jun '10

David and Andrew, Concerning Useful Official and De facto Standards : At the 6th e-Infrastructure Concertation Meeting in Lyon on 24 November 2008, David WALLOM performed a presentation entitled 'Standardisation: Recent progress review and best-practices sharing: Middleware Track', where slides 26, 27 and 28 showed relationships between OGF recommendations. These slides permitted me to understand much better the work undertaken at OGF, and I thank again David very much for them. But since then, I have NOT seen that OGF has published any update of these slides. Therefore, I have used the open source ArgoUML tool to create a diagram containing : - updated information about standards and OGF recommendations shown in these slides, - relationships between official standards, de facto standards and OGF recommendations about security (AUTHN and AUTHZ). I have published : - This diagram at http://forge.gridforum.org/sf/go/doc15990?nav=1 - The source UML at http://forge.gridforum.org/sf/go/doc15977?nav=1 I would like that everyone verifies this diagram, in order to correct error, add improvements, and finally permit official publication. In particular : - David, can you verify this diagram (in particular for recommendations about Data and Job) ? - Andrew, can you verify this diagram (in particular for the SAML and WS standards and recommendations about security) ? Thank you in advance for your contribution. Best regards. ----------------------------------------------------- Etienne URBAH LAL, Univ Paris-Sud, IN2P3/CNRS Bat 200 91898 ORSAY France Tel: +33 1 64 46 84 87 Skype: etienne.urbah Mob: +33 6 22 30 53 27 mailto:urbah@lal.in2p3.fr -----------------------------------------------------

1 0

Re: [Pgi-wg] OGF PGI Use Cases and Scpoe
by Etienne URBAH 08 Jun '10

08 Jun '10

Morris and all PGI members, Concerning OGF PGI Uses Cases and Scope : UML diagrams ------------ I have updated inside http://forge.gridforum.org/sf/docman/do/listDocuments/projects.pgi-wg/docma… the source file and the pictures of UML Class Diagrams (designed with the ArgoUML tool) : - showing useful official and de facto standards, - showing INTERFACES of Distributed Data Processing which are necessary or useful to standardize, - permitting to assess the impact of SCOPE and architecture on the list of INTERFACES which are absolutely required to be standardized to permit minimum interoperability. For each interface, I have added the known relevant standard(s) inside square brackets. Arrows in RED depict relationships NOT using interfaces. These relationships hinder interoperability and should be avoided when possible. Terminology / Vocabulary ------------------------ - INSTRUMENT : Anything generating large sets of scientific data (HEP experiment, Medical imaging, Observation instrument, Petaflopic computer, ...) as taken into account by the DORII project - PAYLOAD : Any Grid or Cloud Entity directly useful for a scientist (Data, Activity, Instrument, ...) - SUPPORT : Any other Grid or Cloud Entity (Security, Info, Application, License, VM image, Log, Accounting, ...) Info, Security, Monitoring, Log and Accounting are absolutely needed to operate a Production Grid. Files ----- http://forge.gridforum.org/sf/go/doc15977?nav=1 ArgoUML source file http://forge.gridforum.org/sf/go/doc15990?nav=1 Useful Official and De facto Standards http://forge.gridforum.org/sf/go/doc15979?nav=1 Abstract Class Diagram http://forge.gridforum.org/sf/go/doc15978?nav=1 Detailed Class Diagram with Abstract Dependencies http://forge.gridforum.org/sf/go/doc15980?nav=1 Detailed Class Diagram with Detailed Dependencies http://forge.gridforum.org/sf/go/doc15981?nav=1 Scope with BES + JSDL http://forge.gridforum.org/sf/go/doc15982?nav=1 Interoperability with a Monolithic Execution Service Standardization priorities -------------------------- The diagrams with detailed dependencies show in particular that : - The interfaces for Activity Management are only connected to the Instrument Manager, the Activity Managers and the Computing Resources. So, it is easy to build a gateway to bridge Activity Management between 2 completely different Grid or Cloud infrastructures (The EDGeS 3G bridge is in full operation between gLite- and BOINC-powered infrastructures). Therefore, Activity Management alone does NOT require urgent standardization. - The interfaces for Security, Info, Log and Accounting are directly connected to most functionalities. Therefore, interoperability between Production Grids absolutely require urgent standardization of these interfaces. Criticism, remarks and suggestions are welcome. Best regards. ----------------------------------------------------- Etienne URBAH LAL, Univ Paris-Sud, IN2P3/CNRS Bat 200 91898 ORSAY France Tel: +33 1 64 46 84 87 Skype: etienne.urbah Mob: +33 6 22 30 53 27 mailto:urbah@lal.in2p3.fr ----------------------------------------------------- On Fri, 14/05/2010 18:03, Etienne URBAH wrote: > Balazs, Morris, Luigi, Johannes and all PGI members, > > > Lot of thanks to Oxana, Aleksandr and Andrew for their contribution to > OGF PGI : > > We finally come to what we all should have begun with, that is the USE > CASES and the SCOPE. > > > +-------------+ > | USE CASES | > +-------------+ > I suggest that : > > - We all provide clear descriptions of Use Cases inside > http://forge.gridforum.org/sf/docman/do/listDocuments/projects.pgi-wg/docma… > > In particular, Use Cases for the five LHC categories described by Oxana > are welcome. > Could the already published GROMACS Use Case be included inside one LHC > category ? > > - Each Use Case (in particular the GROMACS Use Case) must clearly > indicate if the required file stagings should be performed : > - automatically by the Execution Service (from locations specified > inside the Job Description), or > - manually by the Submitter of the Activity, which then requires to be > in a 'Hold' state, and its 'session directory' (or alike) published. > > - We describe these Use Cases with a graphical UML tool, such as ArgoUML > (which runs on MS-Windows, MAC OS X and Linux), and publish the > corresponding source files inside GridForge (open collaborative process). > > - Simple example of Use Case : > The Submitter of an Activity MAY be a scientist using a Scientific Grid > Portal permitting to submit predefined Applications. > Therefore, this Activity Submitter MAY have very little knowledge of > Grids, and MAY have very little knowledge of the Application executed as > Payload on the Computing Resource. > > > +---------+ > | SCOPE | > +---------+ > Execution of Activities or Jobs is only a little part of the > 'Distributed Data Processing' functionalities that a Production Grid is > required to provide. > > In particular, we MUST clearly understand the difference of 'Quality of > Service' required by : > - Transient entities (such as Activities or Jobs), which MAY fail at any > time for any reason, > - Persistent entities (such as Grid resource descriptors, Security > descriptors, Data sets, Accounting Records, Log records, ...), which > SHOULD be securely kept. > > I will publish very soon the source file and the pictures of UML > Collaboration Diagrams (designed with the above mentioned ArgoUML tool) > showing : > > - INTERFACES of Distributed Data Processing which are necessary or > useful to standardize > > - The impact of ARCHITECTURE on the list of INTERFACES which absolutely > require to be standardized to permit minimum interoperability. > > > +------------------------+ > | FOUNDATION STANDARDS | > +------------------------+ > In order to ease mutual understanding and general agreement, we need > Foundation Standards as sound basis. > > Requirement NF6 (162) : JSPG (Security Policies) > ------------------------------------------------- > Without agreement on AUTHN and AUTHZ, there is simply NO interoperability. > > Requirement IS1 (1) : GLUE model > --------------------------------- > - At the Amsterdam meeting, someone said about GLUE : 'Information model > does not concretely say anything'. > - I have the totally opposite opinion : I strongly suggest to use the > GLUE model (currently GLUE 2.0) as one of these Foundation Standards, > and to describe as much concepts as possible using GLUE entities. > > > +----------------------------+ > | TERMINOLOGY / VOCABULARY | > +----------------------------+ > For clarity, I suggest following definitions : > > Client > ------ > Holder of credentials belonging to a member of a GLUE UserDomain. > For example, a Client MAY submit an Activity to an Execution Service, or > query the Status of an Activity. > > Activity = Job (fully synonyms) > -------------- > Remote processing which a Client describes in a 'Job Description', which > the Client then submits to an Execution Service. > > Payload > ------- > Anything (Application, Script, Pilot Job, ...) executed by a Computing > Resource on request of the Activity. The payload MAY completely ignore > that is is executed inside a grid Activity. > > Simple Activity > --------------- > Simple Job Description containing only ONE local job executed by only > ONE batch system, WITHOUT 'Hold' states NOR manual staging. > - This is a suggestion of restrictive evolution of requirement JM5 (55). > - An Activity requiring 'Hold' states and/or manual staging is then > called 'Interactive Activity'. > > > Criticism, remarks and suggestions are welcome. > > Best regards. > > ----------------------------------------------------- > Etienne URBAH LAL, Univ Paris-Sud, IN2P3/CNRS > Bat 200 91898 ORSAY France > Tel: +33 1 64 46 84 87 Skype: etienne.urbah > Mob: +33 6 22 30 53 27 mailto:urbah@lal.in2p3.fr > ----------------------------------------------------- > > > On 13/05/2010 21:00, Oxana Smirnova wrote: >> Hi Andrew, all, >> >> allow me to start from very starters, to explain "typical" workload >> Aleksandr referred to. >> >> Both ARC and gLite "grew" from the requirements of the High Energy >> Physics community, more specifically - those of LHC experiments. I'll >> come back later to what it means in practice. >> >> The basic difference between gLite and ARC starting requirements is that >> gLite is designed for resources owned or controlled to a large extent by >> their users, while ARC is designed to support resources shared between >> different users and controlled by fairly independent resource providers. >> >> The immediate difference is policies: while gLite community is largely >> expected to comply with policies devised by the LHC Grid (WLCG) Joint >> Security and Policy Group, ARC community has no unique set of policies. >> ARC sites that contribute to LHC computing in general tend to respect >> the WLCG policies, but not too close, giving priority to local policies >> of resource owners. Needless to say, this introduces extra complexity >> into requirements, and reduces the rate of simple use cases. >> >> Now, LHC experiments are huge communities, between 500 and 3000 members >> in each. All well-educated computer-savvy people who never hesitate >> coming with their own brilliant solutions. Even within one experiment >> divergence is huge, and there are sites that support 4 experiments. >> This adds to the complexity: not only we have diverging or contradictory >> requirements from resource owners, we also have all sorts of diverging >> requirements from users. >> The least common denominator is 1: meaning "ping", because even >> "hello world" is ambiguous >> - where do you send the output, to a file or to a standard output? >> If it is standard output, do you write it to a log? >> Who said "hello world", the individual user or the whole experiment? >> Do we have a right to log individual activities at all? And so on. >> >> There are several basic kinds of typical jobs by LHC experiments, in >> general they can be separated in 5 categories: >> >> 1. Monte Carlo generation (no input, small resource consumption, small >> output) >> 2. Detector simulation (one input file, moderate resource consumption, >> moderate output) >> 3. Signal reconstruction (multiple input files, moderate resource >> consumption, large output) >> 4. Data merging (very many input files - like, 400, large resource >> consumption, large output) >> 5. Data analysis (huge amount of input files not necessarily known in >> advance, small resource consumption, small output) >> >> The job can have a number of states, e.g.: >> >> 1. Job is defined (may require authorisation) >> 2. Job matched to a site (requires authorisation, detailed site >> information, maybe data availability information) >> 3. Data are made available to a job (authorisation, probably delegation >> of data staging rights to a staging service) >> 4. Job processes data (CPU, I/O, network access to external databases >> requiring authorisation) >> 5. Job places data at [multiple] pre-defined destinations >> (authorisation, contacting external databases, probably delegation to a >> staging service) >> 6. Job is finished >> 7. Job has failed >> 8. Job is terminated (requires authorisation) >> 9. Job is resubmitted (authorisation, information) >> >> Each state may have a number of sub-states, depending on the >> experiment-specific framework. >> >> Authorisation may be per Virtual Organisation (file reading), per Role >> and/or Group within a VO (job types 1-4), per person (job type 5), or >> even per service (some frameworks accept services as VO members). >> >> Delegation of rights in general is very much needed, because of the >> large number of auxiliary services, distributed environment and general >> complexity of the workflow. No-one really knows how to achieve the goals >> without delegation. >> >> Each experiment has their own framework. Most such frameworks circumvent >> Grid services because they are too generic. This means that jobs are >> highly non-trivial as they attempt to re-implement Grid services such as >> information gathering and publishing, data movement and registration, >> resubmission etc.; they also tend to tweak authorisation by executing >> payload of users not authorised by Grid means. This complicates even >> further the job state picture. >> >> If PGI's outcome will make any of the above mentioned jobs impossible, >> most key ARC and gLite customers will not use PGI specs, and they will >> have only academic value. This was not the PGI goal, as "P" stands for >> "Production". >> >> Cheers, >> Oxana >> >> >> 13.05.2010 15:20, Andrew Grimshaw пишет: >>> Aleksandr, >>> Referring to your sentence/paragraph >>> >>> "Such "simple" job is very far from being "typical". At least in >>> NorduGrid world AFAIK." >>> >>> Could you elaborate. I see in my work basically two "types" of jobs >>> that dominate - sets of HTC "parameter space" jobs, and true parallel >>> MPI jobs. >>> In both cases the "job" is basically a command line - either an >>> mpiexec/run, and application with parameters, or a script with >>> parameters. >>> The job has known inputs and outputs, or a directory tree location >>> where it needs to run. >>> The jobs runs to completion, or it fails, in either case there are >>> output files and result codes. >>> Sometimes the job is a workflow, but when you pick that apart it turns >>> into jobs that have inputs and outputs along with a workflow engine >>> orchestrating it all. >>> >>> What is a typical job that you see? >>> When I say "typical" I mean covers 80% of the jobs. >>> >>> A >>> >>> -----Original Message----- >>> From: Aleksandr Konstantinov [mailto:aleksandr.konstantinov@fys.uio.no] >>> Sent: Sunday, May 02, 2010 3:36 PM >>> To: pgi-wg(a)ogf.org >>> Cc: Andrew Grimshaw; 'Oxana Smirnova'; 'Etienne URBAH'; >>> 'David SNELLING'; lodygens(a)lal.in2p3.fr >>> Subject: Re: [Pgi-wg] OGF PGI Requirements - Flexibility and clarity >>> versus Rigidity and confusion >>> >>> Hello, >>> >>> I agree that problem is to difficult to solve. >>> One should take into account that initially task was different. >>> Originally AFAIR it was an attempt of few grid project to make a >>> common interface suitable for them. Later those were forced to OGF >>> and problem upgraded to almost unsolvable. >>> >>> Andrew Grimshaw at Saturday 01 May 2010 15:36 wrote: >>>> Oxana, >>>> Well said. >>>> >>>> I would add that I fear we may be trying too much to solve all >>>> problems the first time around - to "boil the ocean". >>>> To completely solve the whole problem is a daunting task indeed >>>> as there are so many issues. >>>> >>>> I personally believe we will make more progress if we solve the >>>> minimum problem first, e.g., securely run a simple job from >>>> infrastructure/sw-stack A on infrastructure/sw-stack B. >>> >>> This problem is already solved. And it was done in few ways. >>> 1. Client stacks supporting multiple service stacks >>> 2. BES + GSI >>> 3. Other combinations currently in use >>> And none is fully suitable for real production. >>> So unless task of PGI is considered to be purely theoretical this >>> approach would become equal to one more delay. >>> >>>> >>>> "Infrastructure/sw-stack A" means a set of resources (e.g., true >>>> parallel-Jugene, clusters, sets of desktops) running a middleware >>>> stack (e.g., Unicore 6 or Arc) configured a particular way. >>>> In the European context this might mean an NGI such as D-Grid with >>>> Unicore 6 running a job on NorduGrid running Arc. >>>> (Please forgive me if I have the particulars of the NGIs wrong.) >>>> >>>> "Simple job" means a job that is typical, not special. >>>> This is not to say that its resource requirements are simple, it may >>>> have very particular requirements (cores per socket, interconnect, >>>> memory), rather I mean that the job processing required is simple: >>>> run w/o staging, simple staging, >>> >>> Such "simple" job is very far from being "typical". At least in >>> NorduGrid world AFAIK. >>> >>>> perhaps client interaction with the session directory pre, post, and >>>> during execution. >>>> Try to avoid complex job state models that will be hard to agree on, >>>> and difficult to implement in some environments. >>>> >>>> "Securely" means sufficient authentication information required at B >>>> is provided to B in a form it will accept from a policy perspective. >>>> Further, that we try as much as possible to avoid a delegation >>>> definition that extends inwards beyond the outer boundary of a >>>> particular infrastructure/sw-stack. >>> >>> I'm lost. Is is delegation or definition which extends? >>> >>>> (The last sentence is a bit awkward, I personally think that we will >>>> need to have two models of authentication and delegation >>>> - a legacy transport layer mechanism, >>>> and a message layer mechanism based on SAML, and that inside of a >>>> software stack we cannot expect sw-stacks to change their internal >>>> delegation mechanism.) >>>> >>>> I believe authentication/delegation is the most critical item: >>>> if we cannot get the authentication/delegation issues solved, >>>> the rest is moot with respect to a PRODUCTION environment. >>>> We may be able to do demo's and stunts while punting on >>>> authentication/delegation, but we will not integrate production >>>> systems.) >>> >>> Wasn't delegation voted no during last review? >>> >>> >>> A.K.

1 0

stepping down as a PGI co-chair
by Balazs Konya 08 Jun '10

08 Jun '10

Dear All, Some of you may know that since 1st May i work as the technical director of European Middleware Initiative project (www.eu-emi.eu) in addition to my existing NorduGrid/ARC roles. Due to my largely increased non-PGI duties i am not able to devote the work required to be a PGI co-chair any longer. After quite some thinking i have decided to step down as a PGI co-chair. In the future i intend to carry on with my PGI work as an ordinary group member representing both ARC and EMI. With my decision the PGI group will be chaired by Morris, who will get help from Johannes, the group's secretary. Thanks for everyone who helped and contributed to the PGI group's progress in the past months/years! regards, Balazs Konya

1 0

no meeting Today (3/06/2010)
by Balazs Konya 03 Jun '10

03 Jun '10

Dear All, In view of the recent mails i propose to cancel Today's PGI meeting. bye, Balazs On 2010-06-02 09:30, Morris Riedel wrote: > Hi Balazs, > > I'm not available tomorrow (public holiday) and we lost our phone > conference system again. On 2010-06-03 10:59, Morris Riedel wrote: > Hi, > > I already informed Balazs that I'm unfortunately not available today due to a > public holiday in Germany. > > In addition, nobody from UNICORE will be available since standardization > expertise is located in JUELICH (Germany) within the UNICORE consortium. > However, this is not a problem I believe since we also discussed the last > telcon w/o ARC presence. Topics for today might be preparation and agenda > plans for OGF+workshop - if not - we do it next week. > > I leave it open to Balazs whether there will be a PGI call today. > > Take care, Morris Tim Parkinson: > Balazs, > > Steve Crouch is off sick today. > > I have a conflicting telecon this afternoon that I have to attend. > > So neither of us will be able to join the call if you elect to hold it. > > Hope that helps your decision > > regards > > Tim Parkinson OMII-UK Project Manager University of Southampton Oxana Smirnova: > Hi, > > I discovered there'll be a clash with another important meeting today, so I > can only attend the first 30 minutes of the PGI call. > > Cheers, Oxana

1 0

PGI-WG Telecon 2010-06-03 Apologies
by Tim Parkinson 03 Jun '10

03 Jun '10

Balazs, Steve Crouch is off sick today. I have a conflicting telecon this afternoon that I have to attend. So neither of us will be able to join the call if you elect to hold it. Hope that helps your decision regards Tim Parkinson OMII-UK Project Manager University of Southampton

2 1

Not available today
by Morris Riedel 03 Jun '10

03 Jun '10

Hi, I already informed Balazs that I'm unfortunately not available today due to a public holiday in Germany. In addition, nobody from UNICORE will be available since standardization expertise is located in JUELICH (Germany) within the UNICORE consortium. However, this is not a problem I believe since we also discussed the last telcon w/o ARC presence. Topics for today might be preparation and agenda plans for OGF+workshop - if not - we do it next week. I leave it open to Balazs whether there will be a PGI call today. Take care, Morris ------------------------------------------------------------ Morris Riedel Distributed Systems and Grid Computing Division Jülich Supercomputing Centre (JSC) Info: http://www.fz-juelich.de/jsc/JSCPeople/riedel

1 0

Re: [Pgi-wg] OGF PGI Use Cases and Scope
by Etienne URBAH 01 Jun '10

01 Jun '10

Balazs, Morris, Luigi, Johannes and all PGI members, Concerning OGF PGI Uses Cases and Scope : UML diagrams ------------ I have updated inside http://forge.gridforum.org/sf/docman/do/listDocuments/projects.pgi-wg/docma… the source file and the pictures of UML Class Diagrams (designed with the ArgoUML tool) : - showing INTERFACES of Distributed Data Processing which are necessary or useful to standardize, - permitting to assess the impact of SCOPE and architecture on the list of INTERFACES which are absolutely required to be standardized to permit minimum interoperability. For each interface, I have added the known relevant standard(s) inside square brackets. Arrows in RED depict relationships NOT using interfaces. These relationships hinder interoperability and should be avoided when possible. Terminology / Vocabulary ------------------------ - INSTRUMENT : Anything generating large sets of scientific data (HEP experiment, Medical imaging, Observation instrument, Petaflopic computer, ...) as taken into account by the DORII project - PAYLOAD : Any Grid or Cloud Entity directly useful for a scientist (Data, Activity, Instrument, ...) - SUPPORT : Any other Grid or Cloud Entity (Security, Info, Application, License, VM image, Log, Accounting, ...) Security, Info, Log and Accounting are absolutely needed to operate a Production Grid. Files ----- http://forge.gridforum.org/sf/go/doc15977?nav=1 ArgoUML source file http://forge.gridforum.org/sf/go/doc15979?nav=1 Abstract Class Diagram http://forge.gridforum.org/sf/go/doc15978?nav=1 Detailed Class Diagram with Abstract Dependencies http://forge.gridforum.org/sf/go/doc15980?nav=1 Detailed Class Diagram with Detailed Dependencies http://forge.gridforum.org/sf/go/doc15981?nav=1 Scope with BES + JSDL http://forge.gridforum.org/sf/go/doc15982?nav=1 Interoperability with a Monolithic Execution Service Standardization priorities -------------------------- The collaboration diagrams show in particular that : - The interfaces for Activity Management are only connected to the Instrument Manager, the Activity Managers and the Computing Resources. So, it is easy to build a gateway to bridge Activity Management between 2 completely different Grid or Cloud infrastructures (The EDGeS 3G bridge is in full operation between gLite- and BOINC-powered infrastructures). Therefore, Activity Management alone does NOT require urgent standardization. - The interfaces for Security, Info, Log and Accounting are directly connected to most functionalities. Therefore, interoperability between Production Grids absolutely require urgent standardization of these interfaces. Criticism, remarks and suggestions are welcome. Best regards. ----------------------------------------------------- Etienne URBAH LAL, Univ Paris-Sud, IN2P3/CNRS Bat 200 91898 ORSAY France Tel: +33 1 64 46 84 87 Skype: etienne.urbah Mob: +33 6 22 30 53 27 mailto:urbah@lal.in2p3.fr ----------------------------------------------------- On Fri, 14/05/2010 18:03, Etienne URBAH wrote: > Balazs, Morris, Luigi, Johannes and all PGI members, > > > Lot of thanks to Oxana, Aleksandr and Andrew for their contribution to > OGF PGI : > > We finally come to what we all should have begun with, that is the USE > CASES and the SCOPE. > > > +-------------+ > | USE CASES | > +-------------+ > I suggest that : > > - We all provide clear descriptions of Use Cases inside > http://forge.gridforum.org/sf/docman/do/listDocuments/projects.pgi-wg/docma… > > In particular, Use Cases for the five LHC categories described by Oxana > are welcome. > Could the already published GROMACS Use Case be included inside one LHC > category ? > > - Each Use Case (in particular the GROMACS Use Case) must clearly > indicate if the required file stagings should be performed : > - automatically by the Execution Service (from locations specified > inside the Job Description), or > - manually by the Submitter of the Activity, which then requires to be > in a 'Hold' state, and its 'session directory' (or alike) published. > > - We describe these Use Cases with a graphical UML tool, such as ArgoUML > (which runs on MS-Windows, MAC OS X and Linux), and publish the > corresponding source files inside GridForge (open collaborative process). > > - Simple example of Use Case : > The Submitter of an Activity MAY be a scientist using a Scientific Grid > Portal permitting to submit predefined Applications. > Therefore, this Activity Submitter MAY have very little knowledge of > Grids, and MAY have very little knowledge of the Application executed as > Payload on the Computing Resource. > > > +---------+ > | SCOPE | > +---------+ > Execution of Activities or Jobs is only a little part of the > 'Distributed Data Processing' functionalities that a Production Grid is > required to provide. > > In particular, we MUST clearly understand the difference of 'Quality of > Service' required by : > - Transient entities (such as Activities or Jobs), which MAY fail at any > time for any reason, > - Persistent entities (such as Grid resource descriptors, Security > descriptors, Data sets, Accounting Records, Log records, ...), which > SHOULD be securely kept. > > I will publish very soon the source file and the pictures of UML > Collaboration Diagrams (designed with the above mentioned ArgoUML tool) > showing : > > - INTERFACES of Distributed Data Processing which are necessary or > useful to standardize > > - The impact of ARCHITECTURE on the list of INTERFACES which absolutely > require to be standardized to permit minimum interoperability. > > > +------------------------+ > | FOUNDATION STANDARDS | > +------------------------+ > In order to ease mutual understanding and general agreement, we need > Foundation Standards as sound basis. > > Requirement NF6 (162) : JSPG (Security Policies) > ------------------------------------------------- > Without agreement on AUTHN and AUTHZ, there is simply NO interoperability. > > Requirement IS1 (1) : GLUE model > --------------------------------- > - At the Amsterdam meeting, someone said about GLUE : 'Information model > does not concretely say anything'. > - I have the totally opposite opinion : I strongly suggest to use the > GLUE model (currently GLUE 2.0) as one of these Foundation Standards, > and to describe as much concepts as possible using GLUE entities. > > > +----------------------------+ > | TERMINOLOGY / VOCABULARY | > +----------------------------+ > For clarity, I suggest following definitions : > > Client > ------ > Holder of credentials belonging to a member of a GLUE UserDomain. > For example, a Client MAY submit an Activity to an Execution Service, or > query the Status of an Activity. > > Activity = Job (fully synonyms) > -------------- > Remote processing which a Client describes in a 'Job Description', which > the Client then submits to an Execution Service. > > Payload > ------- > Anything (Application, Script, Pilot Job, ...) executed by a Computing > Resource on request of the Activity. The payload MAY completely ignore > that is is executed inside a grid Activity. > > Simple Activity > --------------- > Simple Job Description containing only ONE local job executed by only > ONE batch system, WITHOUT 'Hold' states NOR manual staging. > - This is a suggestion of restrictive evolution of requirement JM5 (55). > - An Activity requiring 'Hold' states and/or manual staging is then > called 'Interactive Activity'. > > > Criticism, remarks and suggestions are welcome. > > Best regards. > > ----------------------------------------------------- > Etienne URBAH LAL, Univ Paris-Sud, IN2P3/CNRS > Bat 200 91898 ORSAY France > Tel: +33 1 64 46 84 87 Skype: etienne.urbah > Mob: +33 6 22 30 53 27 mailto:urbah@lal.in2p3.fr > ----------------------------------------------------- > > > On 13/05/2010 21:00, Oxana Smirnova wrote: >> Hi Andrew, all, >> >> allow me to start from very starters, to explain "typical" workload >> Aleksandr referred to. >> >> Both ARC and gLite "grew" from the requirements of the High Energy >> Physics community, more specifically - those of LHC experiments. I'll >> come back later to what it means in practice. >> >> The basic difference between gLite and ARC starting requirements is that >> gLite is designed for resources owned or controlled to a large extent by >> their users, while ARC is designed to support resources shared between >> different users and controlled by fairly independent resource providers. >> >> The immediate difference is policies: while gLite community is largely >> expected to comply with policies devised by the LHC Grid (WLCG) Joint >> Security and Policy Group, ARC community has no unique set of policies. >> ARC sites that contribute to LHC computing in general tend to respect >> the WLCG policies, but not too close, giving priority to local policies >> of resource owners. Needless to say, this introduces extra complexity >> into requirements, and reduces the rate of simple use cases. >> >> Now, LHC experiments are huge communities, between 500 and 3000 members >> in each. All well-educated computer-savvy people who never hesitate >> coming with their own brilliant solutions. Even within one experiment >> divergence is huge, and there are sites that support 4 experiments. >> This adds to the complexity: not only we have diverging or contradictory >> requirements from resource owners, we also have all sorts of diverging >> requirements from users. >> The least common denominator is 1: meaning "ping", because even >> "hello world" is ambiguous >> - where do you send the output, to a file or to a standard output? >> If it is standard output, do you write it to a log? >> Who said "hello world", the individual user or the whole experiment? >> Do we have a right to log individual activities at all? And so on. >> >> There are several basic kinds of typical jobs by LHC experiments, in >> general they can be separated in 5 categories: >> >> 1. Monte Carlo generation (no input, small resource consumption, small >> output) >> 2. Detector simulation (one input file, moderate resource consumption, >> moderate output) >> 3. Signal reconstruction (multiple input files, moderate resource >> consumption, large output) >> 4. Data merging (very many input files - like, 400, large resource >> consumption, large output) >> 5. Data analysis (huge amount of input files not necessarily known in >> advance, small resource consumption, small output) >> >> The job can have a number of states, e.g.: >> >> 1. Job is defined (may require authorisation) >> 2. Job matched to a site (requires authorisation, detailed site >> information, maybe data availability information) >> 3. Data are made available to a job (authorisation, probably delegation >> of data staging rights to a staging service) >> 4. Job processes data (CPU, I/O, network access to external databases >> requiring authorisation) >> 5. Job places data at [multiple] pre-defined destinations >> (authorisation, contacting external databases, probably delegation to a >> staging service) >> 6. Job is finished >> 7. Job has failed >> 8. Job is terminated (requires authorisation) >> 9. Job is resubmitted (authorisation, information) >> >> Each state may have a number of sub-states, depending on the >> experiment-specific framework. >> >> Authorisation may be per Virtual Organisation (file reading), per Role >> and/or Group within a VO (job types 1-4), per person (job type 5), or >> even per service (some frameworks accept services as VO members). >> >> Delegation of rights in general is very much needed, because of the >> large number of auxiliary services, distributed environment and general >> complexity of the workflow. No-one really knows how to achieve the goals >> without delegation. >> >> Each experiment has their own framework. Most such frameworks circumvent >> Grid services because they are too generic. This means that jobs are >> highly non-trivial as they attempt to re-implement Grid services such as >> information gathering and publishing, data movement and registration, >> resubmission etc.; they also tend to tweak authorisation by executing >> payload of users not authorised by Grid means. This complicates even >> further the job state picture. >> >> If PGI's outcome will make any of the above mentioned jobs impossible, >> most key ARC and gLite customers will not use PGI specs, and they will >> have only academic value. This was not the PGI goal, as "P" stands for >> "Production". >> >> Cheers, >> Oxana >> >> >> 13.05.2010 15:20, Andrew Grimshaw пишет: >>> Aleksandr, >>> Referring to your sentence/paragraph >>> >>> "Such "simple" job is very far from being "typical". At least in >>> NorduGrid world AFAIK." >>> >>> Could you elaborate. I see in my work basically two "types" of jobs >>> that dominate - sets of HTC "parameter space" jobs, and true parallel >>> MPI jobs. >>> In both cases the "job" is basically a command line - either an >>> mpiexec/run, and application with parameters, or a script with >>> parameters. >>> The job has known inputs and outputs, or a directory tree location >>> where it needs to run. >>> The jobs runs to completion, or it fails, in either case there are >>> output files and result codes. >>> Sometimes the job is a workflow, but when you pick that apart it turns >>> into jobs that have inputs and outputs along with a workflow engine >>> orchestrating it all. >>> >>> What is a typical job that you see? >>> When I say "typical" I mean covers 80% of the jobs. >>> >>> A >>> >>> -----Original Message----- >>> From: Aleksandr Konstantinov [mailto:aleksandr.konstantinov@fys.uio.no] >>> Sent: Sunday, May 02, 2010 3:36 PM >>> To: pgi-wg(a)ogf.org >>> Cc: Andrew Grimshaw; 'Oxana Smirnova'; 'Etienne URBAH'; >>> 'David SNELLING'; lodygens(a)lal.in2p3.fr >>> Subject: Re: [Pgi-wg] OGF PGI Requirements - Flexibility and clarity >>> versus Rigidity and confusion >>> >>> Hello, >>> >>> I agree that problem is to difficult to solve. >>> One should take into account that initially task was different. >>> Originally AFAIR it was an attempt of few grid project to make a >>> common interface suitable for them. Later those were forced to OGF >>> and problem upgraded to almost unsolvable. >>> >>> Andrew Grimshaw at Saturday 01 May 2010 15:36 wrote: >>>> Oxana, >>>> Well said. >>>> >>>> I would add that I fear we may be trying too much to solve all >>>> problems the first time around - to "boil the ocean". >>>> To completely solve the whole problem is a daunting task indeed >>>> as there are so many issues. >>>> >>>> I personally believe we will make more progress if we solve the >>>> minimum problem first, e.g., securely run a simple job from >>>> infrastructure/sw-stack A on infrastructure/sw-stack B. >>> >>> This problem is already solved. And it was done in few ways. >>> 1. Client stacks supporting multiple service stacks >>> 2. BES + GSI >>> 3. Other combinations currently in use >>> And none is fully suitable for real production. >>> So unless task of PGI is considered to be purely theoretical this >>> approach would become equal to one more delay. >>> >>>> >>>> "Infrastructure/sw-stack A" means a set of resources (e.g., true >>>> parallel-Jugene, clusters, sets of desktops) running a middleware >>>> stack (e.g., Unicore 6 or Arc) configured a particular way. >>>> In the European context this might mean an NGI such as D-Grid with >>>> Unicore 6 running a job on NorduGrid running Arc. >>>> (Please forgive me if I have the particulars of the NGIs wrong.) >>>> >>>> "Simple job" means a job that is typical, not special. >>>> This is not to say that its resource requirements are simple, it may >>>> have very particular requirements (cores per socket, interconnect, >>>> memory), rather I mean that the job processing required is simple: >>>> run w/o staging, simple staging, >>> >>> Such "simple" job is very far from being "typical". At least in >>> NorduGrid world AFAIK. >>> >>>> perhaps client interaction with the session directory pre, post, and >>>> during execution. >>>> Try to avoid complex job state models that will be hard to agree on, >>>> and difficult to implement in some environments. >>>> >>>> "Securely" means sufficient authentication information required at B >>>> is provided to B in a form it will accept from a policy perspective. >>>> Further, that we try as much as possible to avoid a delegation >>>> definition that extends inwards beyond the outer boundary of a >>>> particular infrastructure/sw-stack. >>> >>> I'm lost. Is is delegation or definition which extends? >>> >>>> (The last sentence is a bit awkward, I personally think that we will >>>> need to have two models of authentication and delegation >>>> - a legacy transport layer mechanism, >>>> and a message layer mechanism based on SAML, and that inside of a >>>> software stack we cannot expect sw-stacks to change their internal >>>> delegation mechanism.) >>>> >>>> I believe authentication/delegation is the most critical item: >>>> if we cannot get the authentication/delegation issues solved, >>>> the rest is moot with respect to a PRODUCTION environment. >>>> We may be able to do demo's and stunts while punting on >>>> authentication/delegation, but we will not integrate production >>>> systems.) >>> >>> Wasn't delegation voted no during last review? >>> >>> >>> A.K.

1 0