David Adams ATLAS DIAL Distributed Interactive Analysis of Large datasets David Adams BNL July 15, 2003 LCG Analysis RTAG CERN.

Slides:



Advertisements
Similar presentations
David Adams ATLAS DIAL Distributed Interactive Analysis of Large datasets David Adams BNL March 25, 2003 CHEP 2003 Data Analysis Environment and Visualization.
Advertisements

David Adams ATLAS DIAL Distributed Interactive Analysis of Large datasets David Adams BNL June 23, 2003 GAE workshop Caltech.
Resource Management Reading: “A Resource Management Architecture for Metacomputing Systems”
The ATLAS Production System. The Architecture ATLAS Production Database Eowyn Lexor Lexor-CondorG Oracle SQL queries Dulcinea NorduGrid Panda OSGLCG The.
5 November 2001F Harris GridPP Edinburgh 1 WP8 status for validating Testbed1 and middleware F Harris(LHCb/Oxford)
Workload Management WP Status and next steps Massimo Sgaravatto INFN Padova.
Copyright © 2007, Oracle. All rights reserved. Managing Concurrent Requests.
David Adams ATLAS ATLAS Distributed Analysis David Adams BNL March 18, 2004 ATLAS Software Workshop Grid session.
9 Chapter Nine Compiled Web Server Programs. 9 Chapter Objectives Learn about Common Gateway Interface (CGI) Create CGI programs that generate dynamic.
K. Harrison CERN, 20th April 2004 AJDL interface and LCG submission - Overview of AJDL - Using AJDL from Python - LCG submission.
03/27/2003CHEP20031 Remote Operation of a Monte Carlo Production Farm Using Globus Dirk Hufnagel, Teela Pulliam, Thomas Allmendinger, Klaus Honscheid (Ohio.
Grid Workload Management & Condor Massimo Sgaravatto INFN Padova.
Scalable Systems Software Center Resource Management and Accounting Working Group Face-to-Face Meeting October 10-11, 2002.
David Adams ATLAS AJDL: Analysis Job Description Language David Adams BNL December 15, 2003 PPDG Collaboration Meeting LBL.
Nick Brook Current status Future Collaboration Plans Future UK plans.
ATLAS DIAL: Distributed Interactive Analysis of Large Datasets David Adams – BNL September 16, 2005 DOSAR meeting.
David Adams ATLAS DIAL status David Adams BNL July 16, 2003 ATLAS GRID meeting CERN.
David Adams ATLAS ATLAS Distributed Analysis Plans David Adams BNL December 2, 2003 ATLAS software workshop CERN.
Grid Workload Management Massimo Sgaravatto INFN Padova.
Event Data History David Adams BNL Atlas Software Week December 2001.
Stuart Wakefield Imperial College London Evolution of BOSS, a tool for job submission and tracking W. Bacchi, G. Codispoti, C. Grandi, INFN Bologna D.
Datasets on the GRID David Adams PPDG All Hands Meeting Catalogs and Datasets session June 11, 2003 BNL.
David Adams ATLAS ADA, ARDA and PPDG David Adams BNL June 28, 2004 PPDG Collaboration Meeting Williams Bay, Wisconsin.
Tool Integration with Data and Computation Grid GWE - “Grid Wizard Enterprise”
INFSO-RI Enabling Grids for E-sciencE ATLAS Distributed Analysis A. Zalite / PNPI.
David Adams ATLAS Architecture for ATLAS Distributed Analysis David Adams BNL March 25, 2004 ATLAS Distributed Analysis Meeting.
Giuseppe Codispoti INFN - Bologna Egee User ForumMarch 2th BOSS: the CMS interface for job summission, monitoring and bookkeeping W. Bacchi, P.
CPT Demo May Build on SC03 Demo and extend it. Phase 1: Doing Root Analysis and add BOSS, Rendezvous, and Pool RLS catalog to analysis workflow.
David Adams ATLAS DIAL status David Adams BNL November 21, 2002 ATLAS software meeting GRID session.
An RTAG View of Event Collections, and Early Implementations David Malon ATLAS Database Group LHC Persistence Workshop 5 June 2002.
Enabling Grids for E-sciencE System Analysis Working Group and Experiment Dashboard Julia Andreeva CERN Grid Operations Workshop – June, Stockholm.
David Adams ATLAS DIAL/ADA JDL and catalogs David Adams BNL December 4, 2003 ATLAS software workshop Production session CERN.
David Adams ATLAS ADA: ATLAS Distributed Analysis David Adams BNL June 7, 2004 BNL Technology Meeting.
GDB Meeting - 10 June 2003 ATLAS Offline Software David R. Quarrie Lawrence Berkeley National Laboratory
David Adams ATLAS Virtual Data in ATLAS David Adams BNL May 5, 2002 US ATLAS core/grid software meeting.
David Adams ATLAS ATLAS Distributed Analysis David Adams BNL September 30, 2004 CHEP2004 Track 5: Distributed Computing Systems and Experiences.
D. Adams, D. Liko, K...Harrison, C. L. Tan ATLAS ATLAS Distributed Analysis: Current roadmap David Adams – DIAL/PPDG/BNL Dietrich Liko – ARDA/EGEE/CERN.
David Adams ATLAS DIAL: Distributed Interactive Analysis of Large datasets David Adams BNL August 5, 2002 BNL OMEGA talk.
David Adams ATLAS DIAL Distributed Interactive Analysis of Large datasets David Adams BNL November 17, 2003 SC2003 Phoenix.
K. Harrison CERN, 3rd March 2004 GANGA CONTRIBUTIONS TO ADA RELEASE IN MAY - Outline of Ganga project - Python support for AJDL - LCG analysis service.
David Adams ATLAS ATLAS distributed data management David Adams BNL February 22, 2005 Database working group ATLAS software workshop.
Tool Integration with Data and Computation Grid “Grid Wizard 2”
K. Harrison CERN, 22nd September 2004 GANGA: ADA USER INTERFACE - Ganga release status - Job-Options Editor - Python support for AJDL - Job Builder - Python.
David Adams ATLAS ATLAS Distributed Analysis: Overview David Adams BNL December 8, 2004 Distributed Analysis working group ATLAS software workshop.
David Adams ATLAS ATLAS-ARDA strategy and priorities David Adams BNL October 21, 2004 ARDA Workshop.
ATLAS-specific functionality in Ganga - Requirements for distributed analysis - ATLAS considerations - DIAL submission from Ganga - Graphical interfaces.
David Adams ATLAS Datasets for the Grid and for ATLAS David Adams BNL September 24, 2003 ATLAS Software Workshop Database Session CERN.
STAR Scheduling status Gabriele Carcassi 9 September 2002.
INFSO-RI Enabling Grids for E-sciencE Using of GANGA interface for Athena applications A. Zalite / PNPI.
Grid Workload Management (WP 1) Massimo Sgaravatto INFN Padova.
ATLAS Distributed Analysis Dietrich Liko IT/GD. Overview  Some problems trying to analyze Rome data on the grid Basics Metadata Data  Activities AMI.
David Adams ATLAS ATLAS Distributed Analysis (ADA) David Adams BNL December 5, 2003 ATLAS software workshop CERN.
STAR Scheduler Gabriele Carcassi STAR Collaboration.
D.Spiga, L.Servoli, L.Faina INFN & University of Perugia CRAB WorkFlow : CRAB: CMS Remote Analysis Builder A CMS specific tool written in python and developed.
David Adams ATLAS ATLAS Distributed Analysis and proposal for ATLAS-LHCb system David Adams BNL March 22, 2004 ATLAS-LHCb-GANGA Meeting.
INFSO-RI Enabling Grids for E-sciencE Ganga 4 Technical Overview Jakub T. Moscicki, CERN.
ATLAS Distributed Analysis DISTRIBUTED ANALYSIS JOBS WITH THE ATLAS PRODUCTION SYSTEM S. González D. Liko
David Adams ATLAS AJDL: Abstract Job Description Language David Adams BNL June 29, 2004 PPDG Collaboration Meeting Williams Bay.
David Adams ATLAS ADA: ATLAS Distributed Analysis David Adams BNL December 15, 2003 PPDG Collaboration Meeting LBL.
ATLAS DIAL: Distributed Interactive Analysis of Large Datasets David Adams Brookhaven National Laboratory February 13, 2006 CHEP06 Distributed Data Analysis.
Ganga/Dirac Data Management meeting October 2003 Gennady Kuznetsov Production Manager Tools and Ganga (New Architecture)
Application architectures Advisor : Dr. Moneer Al_Mekhlafi By : Ahmed AbdAllah Al_Homaidi.
Seven things you should know about Ganga K. Harrison (University of Cambridge) Distributed Analysis Tutorial ATLAS Software & Computing Workshop, CERN,
David Adams ATLAS DIAL Distributed Interactive Analysis of Large datasets David Adams BNL May 19, 2003 BNL Technology Meeting.
U.S. ATLAS Grid Production Experience
BOSS: the CMS interface for job summission, monitoring and bookkeeping
BOSS: the CMS interface for job summission, monitoring and bookkeeping
BOSS: the CMS interface for job summission, monitoring and bookkeeping
The Ganga User Interface for Physics Analysis on Distributed Resources
Presentation transcript:

David Adams ATLAS DIAL Distributed Interactive Analysis of Large datasets David Adams BNL July 15, 2003 LCG Analysis RTAG CERN

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG2 Contents Goals of DIAL What is DIAL? Design Dataset Application Task Job Result Scheduler Exchange format Implementation Status Development plans Batch production Interactivity issues Other projects RTAG architecture

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG3 Goals of DIAL 1. Demonstrate the feasibility of interactive analysis of large datasets How much data can we analyze interactively? 2. Set requirements for GRID services Datasets, schedulers, jobs, results, resource discovery, authentication, allocation, Provide ATLAS with a useful analysis tool For current and upcoming data challenges Real world deliverable Like to add another experiment would show generality

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG4 What is DIAL? Distributed Data and processing Interactive Iterative processing with prompt response – (seconds rather than hours) Analysis of Fill histograms, select events, … Large datasets Any event data (not just ntuples or tag)

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG5 What is DIAL? (cont) DIAL provides a connection between Interactive analysis framework –Fitting, presentation graphics, … –E.g. ROOT, JAS, … and Data processing application –Natural to the data of interest –E.g. athena for ATLAS DIAL distributes processing Among sites, farms, nodes To provide user with desired response time Look to other projects to provide most infrastructure

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG6

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG7 Design DIAL has the following major components Dataset describing the data of interest Application defined by experiment/site Task is user extension to the application Job uses application and task to process a dataset Result is the output of a job Scheduler creates and manages jobs Together these define a high-level JDL (job definition language) Figure shows how these components interact →

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG8 User Analysis Job 1 Job 2 ApplicationTask Dataset 1 Scheduler 1. Create or locate 2. select3. Create or select 4. select 8. run(app,tsk,ds1) 5. submit(app,tsk,ds) 8. run(app,tsk,ds2) 6. split Dataset Dataset 2 7. create e.g. ROOT e.g. athena Result 9. fill 10. gather Result 9. fill ResultCode

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG9 Dataset Dataset specifies a collection of data Used as input for job Of special interest is event dataset –Each piece of data associated with one “event” –Events can be processed independently Not just a collection of (logical files) Data may appear in multiple logical files, Some data in databases, … However collection of LF’s is most interesting case Datasets can be split into sub-datasets Each can be processed independently Each used to define a sub-job for distributed processing

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG10 Dataset (cont) Splitting datasets is the key to distributed processing Hard problem –Especially for interactive analysis Locate data that can be accessed quickly Match CPU and data Starting to identify splitting as a separate component Datasets are subject for another talk

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG11 Application User specifies application with Name and version –E.g. dial_cbnt 0.3 Physical application Installed at sites where data is processed Receives task and dataset as input Creates result with output Typically a wrapper around an existing data-processing executable –PAW for PAW files (e.g. dial_cbnt) –ROOT for ROOT files –Athena for ATLAS event data

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG12 Task Task allows users to extend an application Run time configuration –Parameters, –Sequence of algorithms, … Code for processing the data Empty result to specify output –E.g. empty histograms (name, ID, min, max, # bins), –Policy for naming and placing data in output files, … Application Defines the task syntax Provides means to build and install task –E.g. compile and dynamic load

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG13 Job Job created by specifying Dataset, application and task Submitting these to a scheduler Job has a user interface providing Status –Dataset, application and task –Running, done, failed, … –Start and stop times –Result Means to update the status –Check schedulers, queues, etc…

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG14 Result Result is filled during processing Examples –Histogram –Event list –New dataset –Combination of the above Accessible to the user through job Partial results may be available during processing Should be small Dataset or logical file identifiers rather than the data in those files

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG15 Scheduler A DIAL scheduler provides means to Submit a job Terminate a job Monitor a job –Access user interface for the job Verify availability of an application Handle tasks –Installation –Verification of installation –Verify consistency with application DIAL philosophy is scheduler should be generic I.e. not restricted to a single application

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG16 Scheduler (cont) Schedulers may form a hierarchy Corresponding to that of compute nodes –Grid, site, farm, node Each scheduler splits job into sub-jobs and distributes these over lower-level schedulers Lowest level LocalScheduler starts processes to carry out the sub-jobs Scheduler concatenates results for its sub-jobs User may enter the hierarchy at any level Client-server communication

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG17

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG18 Exchange format DIAL data components are exchanged (Data means all but scheduler) Between –User and scheduler –Scheduler and scheduler –Scheduler and application executable Exchange can be between –Different process –Different nodes or sites –Different languages (C++, Java, Python, …) Each component has an XML representation –Should XML define the JDL?

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG19 Implementation DIAL provides in C++ Generic implementations for –Application, Task and Job Interfaces (base classes) for –Dataset, Result, Job and scheduler ATLAS implementations of these interfaces –Dataset: AthenaRoot file, combined ntuple (hbook file) –Result: hbook file –Job: child process, LSF (soon) –Scheduler: Single job, farm (soon) Utilities and registries to support these components Some of the implementation lags behind the design

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG20 Implementation (cont) Applications dial_cbnt is wrapper around PAW for processing ATLAS combined ntuple files ROOT can be used as front end to DIAL All classes imported with ACLiC Demo shows how to –Create application specification for dial_cbnt –Create empty result from an hbook file with empty histograms –Create a task with this result and code to fill histograms –Create a combined ntuple dataset from an XML file –Submit the application, task and dataset to a scheduler –Fetch the job description –Query the job for status and result

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG21 Status Preceding is in DIAL release 0.3 last month Demo runs at BNL (ACF and RCF) and CERN (lxplus) Scheduler is very simple (no job splitting) Anticipate release 0.4 in August Magda files accessible Local distributed processing More information DIAL home page – CHEP paper – –Design and implementation have evolved

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG22 Development plans Farm Scheduler Distribute processing over a single farm (BNL) –First LSF (BNL and CERN) –Then Condor (local, COD, master-worker, Condor-G?) Makes DIAL a useful tool for distributed processing Remote access to scheduler Job submission from anywhere Web service Add policy to scheduler interface Response time How to split dataset for distributed processing

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG23 Development plans (cont) Grid schedulers Distribute data and processing over multiple sites Interact with dataset, file and replica catalogs Authentication, authorization, resource location and allocation, … ATLAS POOL dataset After ATLAS incorporates POOL ATLAS athena as application

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG24 Batch production Original goal of DIAL was to provide means for interactive data analysis. However interactive analysis and batch production are not so easily separated DIAL JDL is appropriate for either Schedulers will need to operate over a continuum of –Produced data size –Acceptable response time Some users will want to interactively browse a small sample and then submit larger sample to batch

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG25 Interactivity issues Response time is critical Interactive system provides means for user to specify maximum acceptable response time All actions must take place within this time –Locate data and resources –Splitting and matchmaking –Job submission –Gathering of results Longer latency for first pass over a dataset –Record state for later passes –Still must be able to adjust to changing conditions

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG26 Interactivity issues (cont) Interactive and batch must share resources Share implies more available resources for both Interactive use varies significantly –Time of day –Time to the next conference –Discovery of interesting events Interactive request must be able to preempt long- running batch jobs –But allocation determined by sites, experiments, …

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG27 Other projects JDL (job definition language) DIAL may be thought of as a proposal for a user-level (high-level) JDL Like to come to agreement on this JDL with other projects so components can be shared Language issues: C++, Java, Python, XML, … Scheduler Look to other projects to deliver robust and efficient implementations of the DIAL scheduler In the mean time DIAL will –Try to wrap analogous components from other projects –Make relatively simple implementations to meet our project goals

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG28 Other projects (cont) Candidates to provide DIAL schedulers: Look to Condor to deliver a grid scheduler Scheduler can be implemented using Chimera Scheduler restricted to ROOT as an application could be implemented using PROOF Scheduler for athena jobs can come from GANGA Tech-X has SBIR to develop JDAP scheduler for use with JAS A java-based scheduler is being developed by STAR Last (but not least?), DIAL will provide schedulers

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG29 Other projects (cont) Build web services to create and access DIAL schedulers Using Clarens and/or OGSA Environments that might use DIAL schedulers: A Python implementation of the JDL would enable GANGA and PI/SEAL to use DIAL schedulers DIAL is already imported into ROOT A Java version of the JDL would enable JAS to use DIAL schedulers Web or script based user-level production for any experiment could benefit from the scheduler interface

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG30

David Adams ATLAS July 15, 2003 DIAL Analysis RTAG31 RTAG architecture How does DIAL connect to the RTAG architecture? Pictured above DIAL addresses the lines –What is exchanged (dataset, app, task, job, …) –Interfaces DIAL neither requires nor excludes persistent workers –Choice made by the scheduler