CHEP as an AMS Regional Center G N

CHEP as an AMS Regional Center G. N. Kim, J. W. Shin, N. Tasneem, M. W. Lee, D. Son (Center for High Energy Physics, KNU) H. Park (Supercomputing Center, KISTI) The Third International Workshop on HEP Data Grid CHEP, KNU 2004. 8. 26 -28

Outline • AMS Experiment • Data flow from ISS to AMS Remote Center • Data Size for AMS Experiment for 3 years • AMS Science Operating Center • Status of CHEP as an AMS Regional Center • Status of MC Data Production • bb. FTP test • Summary

AMS (Alpha Magnetic Spectrometer) Experiment • A Particle Physics Experiment on the International Space Station for 3 or 4 Years • Will Collect ~1010 Cosmic Rays in Near-Earth Orbit from 300 Me. V to 3 Te. V PHYSICS GOALS : To search for Antimatter (He, C) in space with a sensitivity of 103 to 104 better than current limits. n To search for Dark matter High statistics precision measurements of e , and p spectrum. n To study Astrophysics. High statistics precision measurements of D, 3 He, 4 He, B, C, 9 Be, 10 Be spectrum «B/C: to understand CR propagation in the Galaxy (parameters of galactic wind). « 10 Be/9 Be: to determine CR confinement time in the Galaxy. n

International Collaboration ~200 scientists + dozens of contractors from 14 countries Spokesperson: TING, Samuel C. C.

AMS-02 (Alpha Magnetic Spectrometer)

Data Flow from ISS to Remote AMS Centers Commercial Satellite Service Space Station p ku etr o ide Fe s& ed ac y. B m ele T k. V n nli ive w Do Remote User Facility F White Sands, NM MSFC POIC JSC Telemetry Voice Planning Commanding Internet Or A Dedicated Service/Circuit 0. 9 meter dish Distribute Locally & Conduct Science

ACOP: AMS Crew Operations Post, POIC: Payload Operation Integration Center, GSC: Ground Support Computers

Table. AMS 02 data transmitted to SOC from POIC Stream Band Width High Rate 3 -4 Mbit/s Slow Rate NASA 16 kbit/s Data Category Volume (TB/year) Scientific Calibration Housing Keeping Auxillary Data 11 – 15 0. 01 0. 06 0. 01 Table. AMS 02 Data Volumes Origin Beam Calibrations Preflight Tests 3 years flight Data Summary Files (DST) Catalogs Event Tages TDV files Data Category Calibration Scientific Calibration House Keeping Ntuples or ROOT files Flat files or ROOT files or ORACLE Volume (TB) 0. 3 0. 4 33 – 45 0. 03 0. 18 165 -260 0. 05 0. 2 0. 5 Total AMS 02 data volume is about 200 TB

AMS Science Operations Center (SOC) Data processing and Science Analysis • • • receive the complete copy of data science analysis primary data storage data archiving data distribution to AMS Universities and Laboratories MC production The SOC will provide all functioning and will give the possibility to process and analyze all data. SOC computing facilities should be enough to provide data access and analysis for all members of the collaboration.

Science Operations Center Computing Facilities Analysis Facilities (linux cluster) N AMS Physics Services Data Servers, Central Data Services Shared Tape Servers Interactive and Batch physics analysis tape robots tape drives LTO, DLT 10 -20 dual processor PCs 5 PC servers Production Facilities, 40 -50 Linux dual-CPU computers Shared Disk Servers 25 Tera. Byte disk 6 PC based servers interactive physics analysis AMS regional Centers Home directories & registry consoles & monitors CERN/AMS Network Linux, Intel and AMD batch data processing Engineering Cluster 5 dual processor PCs

AMS Science Operation Center Computing Facilities Production Farm PC Linux 3. 4+GHz Cell #7 PC Linux 3. 4+GHz Archiving and Staging Gigabit Switch (CERN CASTOR) AFS Server PC Linux Server 2 x 3. 4+GHz, RAID 5 , 10 TB Cell #1 Gigabit Switch (1 Gbit/sec) MC Data Server Disk Server Disk Server PC Linux Server 2 x 3. 4+GHz Data Server Tested, prototype in production Not tested and no prototype yet Simulated data AMS data NASA data metadata Web, News Production, DB servers PC Linux Server 2 x 3. 4+GHz, RAID 5 , 10 TB Analysis Facilities

AMS Regional Center(s) • Monte-Carlo Production • Data Storage and data access for the remote stations AMS Regional Stations(s) and Center(s) • Access to the SOC data storage * for the detector verification studies * for detector calibration purposes * for alignment * event visualization • Access to the SOC to get the detector and production status • Access to SOC computing facilities for Science Analysis • Science Analysis using local computing facilities of Universities and Laboratories.

Connectivity from ISS to CHEP (RC) International Space Station (ISS) 무궁화위성 Commercial Satellite Service CHEP ISP-1 White Sands, NM & eds y etr em l Ba Te Chicago NAP e o. F ide e Fiv JSC ISP-2 p cku V nk nli w Do CERN (SOC) v. BNS MSFC POIC/ POCC Telemetry Voice Planning Commanding NINS B 4207 v. BNS MIT LANs ISP : Internet Service Provider NAP - Network Access Point v. BNS - very high Broadband Network Service NINS - NASA Integrated Network Services POCC - Payload Operations Control Center POIC- Payload Operations Integration Center ISP-3 RC RS

CHEP as an AMS Regional Center 무궁화위성 Internet Server Ewha Data Storage Display Facility AMS RC Disk Storage (20 TB) Analysis Facility Gigabit Ethernet Tape Library (~200 TB) DB Server Hub Linux Clusters Cluster Servers AMS RS

AMS Analysis Facility at CHEP 1) CHEP AMS Cluster Server: 1 CPU: AMD Athlon MP 2000+ dual Disk Space: 80 GB OS: Redhat 9. 0 Gigabit Ethernet Linux Clusters AMS Cluster Servers Hub Linux Clusters : 12 CPU: AMD Athlon MP 2000+ dual Disk Space: 80 GB x 12 =960 GB OS : Redhat 7. 3 2) KT AMS Cluster Server: 1 CPU: Intel XEON 2 GHz dual Disk Space: 80 GB OS: Redhat 9. 0 Linux Clusters : 4 CPU: Intel XEON 2 GHz dual Disk Space: 80 GB x 4 =320 GB OS: Redhat 7. 3

Data Storage IBM TAPE LIBRARY SYSTEM – 43. 7 TB 3494 -L 12 8. 4 TB 3494 -S 10 13. 537 TB 3494 -L 12 7. 36 TB 3494 -S 10 43. 7 TB 14. 4 TB L 12 S 10 Raid Disks Fast T 200: 1 TB (Raid 0: striping) experimental group C C Login Machin e C C cluster C C Fast. T 200 RAID Disks CDF, CMS, NFS AMS, BELLE, PHENIX file system Intermediate DISKPOOL Migration Tape. Library

Network Configuration CERN GEANT-TEIN(EU) Research Traffics 34 Mbps APII(Japan) 8 Mbps APII/KREONET 2 (US) Fermilab KEK 622 Mbps x 2 1 Gbps x 2 KREONET KOREN 2. 5 Gbps other Traffics (total 145 Mbps) Gigabit Ethernet KORNET Boranet C 6509 Gigabit Ethernet/ATM 155 (Physics L 3 Switch Department) Gigabit Switches Gigabit Ethernet (CHEP) IBM 8271 PCs … … Servers

Connectivity to the Outside from CHEP KOREN Topology APII China USA Star. Tap, ESNET Seoul (IHEP) APII USA (Chicago) TEIN (CERN) Daejeon 2. 5 Gbps APII JAPAN (KEK) Daegu CHEP (Regional Center) Daegu ■ Singapore (1) SIng. AREN through APII (2 Mbps) ■ China (1, Preplanned) CSTNET (APII) thru APII

Connectivity to the Outside from Korea Star. Tap APII-Trans. Pac Europe CERN TEIN Hyunhai/Genkai US FNAL

MC Production Center and their Capacity

MC Production Procedure by Remote Client Step: 1 Write here registered info. Step: 2 Click here! Step: 3 Choose any one from datasets. Step: 4 Choose appropriate dataset.

MC Production Procedure by Remote Client Step: 7 Put total number of jobs requested Step: 5 Choose appropriate CPU Type & CPU clock. Step: 6 Put appropriate CPU time limit. Step: 9 Choose MC production Mode. Step: 10 Click On ‘Submit Request’ Step: 8 Put ‘Total Real Time Required’.

MC Database Query Form

MC Data Production Status MC Center Responsible GB % CIEMAT J. Casuas 2045 24. 3 CERN V. Choutko, A. Eline, A. Klimentov 1438 17. 1 Yale E. Finch 1268 15. 1 Academia Sinica Z. Ren, Y. Lei 1162 13. 8 LAPP/Lyon C. Goy, J. Jacquemier 825 9. 8 INFN Milano M. Boschini, D. Grandi 528 6. 2 CNAF & INFN Bologna D. Casadei 441 5. 2 UMD A. Malinine 210 2. 5 EKP, Karlsruhe V. Zhukov 202 2. 4 GAM, Montpellier J. Bolmont, M. Sapinski 141 1. 6 INFN Siena&Perugia, ITEP, LIP, IAC, SEU, KNU P. Zuccon, P. Maestro, Y. Lyublev, F. Barao, C. Delgado, Ye Wei, J. Shin 135 1. 6

MC Production Statistics 185 days, 1196 computers 8. 4 TB, 250 PIII 1 GHz/day Particle % of Total protons 7630 99. 9 helium 3750 99. 6 electrons 1280 99. 7 positrons 1280 100 deuterons 250 100 antiprotons 352. 5 100 carbon 291. 5 97. 2 photons 128 100 Nuclei (Z 3… 28) URL: pcamss 0. cern. ch/mm. html Million Events 856. 2 85

MC Events Production at CHEP Total Number of Requested Job : 229 Completed Job : 218 Failed Job : 1 Unchecked Job (may be error) : 10 Particle Requested Completed Failed DST size[GB] Positron 15 15 0 3. 3 Electron 40 40 0 7. 1 Proton 99 94 5 29. 5 Nuclear 40 35 5 4. 7 C 20 19 1 5. 5 He 15 15 0 5. 0 Total 229 218 11 55. 1

Data Handling Program: bb. FTP • bb. FTP is a file transfer software developed by Gilles Farrache in Lyon ü Encoded username and password at connection ü SSH and Certificate authentification modules üMulti-stream transfer ü Big windows as defined in RFC 1323 ü On-the-fly data compression üAutomatic retry ü Customizable time-outs ü Transfer simulation ü AFS authentification integration ü RFIO interface http: //ccweb. in 2 p 3. fr/bbftp/

Data Transmission Test using bb. FTP • Using TEIN in Dec. 2002 - for a number of TCP/IP streams - for a file size

• Using APII/KREONET 2 -Startap in Aug. 2004 - for a number of TCP/IP streams - for a file size

Summary CHEP as AMS Regional Center - Prepared an Analysis Facility and Data Storage - Producing MC events - Progress in MC production with GRID Tools