1. Introduction to Storage Element Hsin-Wei Wu Academia Sinica Grid Computing Center, Taiwan

2.  SRM  A union form standard for SEs  Storage Elements  BeStMan  Castor  dCash  DPM  StoRM  LFC (LHC File Catalog)  Demonstration  Data Management Commands 2 Outline

3. 3 API Access Job Mgmt. Services Computing Element Workload Management Metadata Catalog Data Services Storage Element Data Movement File & Replica Catalog Authorization Security Services Authentication Information & Monitoring Information & Monitoring Services Service Discovering Accounting Auditing Job Provenance Package Manager CLI Network Monitoring Overview of gLite Middleware

4. Data service in gLtie Ref gLite-3.2 User Guide 4  Storage(SE): The files are physical located  Storage URL or SURL  srm://as-ds01.euasiagrid.org/dpm/euasiagrid.org/data  Catalogs: High level hierarchical namespace, maps the “physical” files to a virtual “logical” filename  Logical File Name or LFN  lfn:/grid/twgrid/data  Movement: put/get files into grid SE's, move/replicate files between SE's  File Transfer Service or FTS  Transport URL or TRUL  gsiftp://sc003.grid.sinica.edu.tw:2811/data

5. 5  A Storage Element (SE) is an aggregate of Grid services that allows Grid users to store and manage files together with the space assigned to them  Data are stored on disk pool servers or Mass Storage Systems  Storage resource management needs to take into account  Transparent access to files (migration to/from disk pool)  File pinning  Space reservation  File status notification  Life time management  The SRM (Storage Resource Manager) takes care of all these details  The SRM is a single interface that takes care of local storage interaction and provides a Grid interface to the outside world Storage Resource Management

6. Berkeley Storage Manager (BeStMan) LBNL  Java implementation  Designed to work with unix- based disk systems  As well as MSS to stage/archive from/to its own disk (currently HPSS)  Adaptable to other file systems and storages (e.g. NCAR MSS, VU L-Store, TTU Lustre, NERSC GFS)  Uses in-memory database (BerkeleyDB) Multiple transfer protocolsMultiple transfer protocols Space reservationSpace reservation Directory management (no ACLs)Directory management (no ACLs) Can copy files from/to remote SRMs or GridFTP ServersCan copy files from/to remote SRMs or GridFTP Servers Can copy entire directory recursivelyCan copy entire directory recursively Large scale data movement of thousands of files Recovers from transient failures (e.g. MSS maintenance, network down) Local PolicyLocal Policy Fair request processing File replacement in disk Garbage collection

7. Castor-SRM CERN and Rutherford Appleton Laboratory  CASTOR is the HSM in production at CERN  Support for multiple tape robots  Support for Disk-only storage recently added  Designed to meet Large Hadron Collider Computing requirements  Maximize throughput from clients to tape (e.g. LHC experiments data taking) C++ ImplementationC++ Implementation Reuse of CASTOR software infrastructureReuse of CASTOR software infrastructure Derived SRM specific classes Configurable number of thread pools for both front- and back- endsConfigurable number of thread pools for both front- and back- ends ORACLE centricORACLE centric Front and back ends can be distributed on multiple hostsFront and back ends can be distributed on multiple hosts

8. dCache-SRM FNAL and DESY  Strict name space and data storage separation  Automatic file replication based on access patterns  HSM Connectivity (Enstore, OSM, TSM, HPSS, DMF)  Automated HSM migration and restore  Scales to Peta-byte range on 1000’s of disks  Supported protocols:  (gsi/krb)FTP, (gsi/krb)dCap, xRoot, NFS 2/3  Separate I/O queues per protocol  Resilient dataset management  Command line and graphical admin interface  Variety of Authorization mechanisms including VOMS  Deployed in a large number of institutions worldwide Support SRM 1.1 and SRM 2.2Support SRM 1.1 and SRM 2.2 Dynamic Space ManagementDynamic Space Management Request queuing and schedulingRequest queuing and scheduling Load balancingLoad balancing Robust replication using srmCopy functionality via SRM, (gsi)FTP and http protocolsRobust replication using srmCopy functionality via SRM, (gsi)FTP and http protocols

9. Disk Pool Manager (DPM) CERN  Provide a reliable, secure and robust storage system  Manages storage on disks only  Security  GSI for authentication  VOMS for authorization  Standard POSIX permissions + ACLs based on user’s DN and VOMS roles  Virtual ids  Accounts created on the fly  Full SRMv2.2 implementation  Standard disk pool manager capabilities  Garbage collector  Replication of hot files  Transfer protocols  GridFTP (v1 and v2)  Secure RFIO  https  Xroot  Works on Linux 32/64 bits machines  Direct data transfer from/to disk server (no bottleneck)  Support DICOM backend  Requirement from Biomed VO  Storage of encrypted files in DPM on the fly + local decryption  Use of GFAL/srm to get TURLs and decrypt the file Supported database backendsSupported database backends MySQL Oracle High availabilityHigh availability All servers can be load balanced (except the DPM one) Resilient: all states are kept in the DB at all times

10. Storage Resource Manager (StoRM) INFN/CNAF - ICTP/EGRID  It's designed to leverage the advantages of high performing parallel file systems in Grid.  Different file systems supported through a driver mechanism:  generic POSIX FS  GPFS  Lustre  XFS  It provides the capability to perform local and secure access to storage resources (file:// access protocol + ACLs on data).file:// StoRM architecture: FrontendsFrontends: C/C++ based, expose the SRM interface BackendsBackends: Java based, execute SRM requests. DBDB: based on MySQL DBMS, stores requests data and StoRM metadata. replicated and instantiated on a dedicated machine.Each component can be replicated and instantiated on a dedicated machine.

11. LHC File Catalog 5th NRENs and Grids Workshop, Paris - 11-12 June 2007 11  LFC maps LFNs to SURLs  Logical File Name (LFN): user file name  in VO namespace, aliases supported  Glbally Unique IDentifier (GUID)  unique string assigned by the system to the file  Site URL (SURL): identifies a replica  A Storage Element and the logical name of the file inside it  GSI security: ACLs (based on VOMS)  To each VOMS group/role corresponds a virtual group identifier  Support for secondary groups  Web Service query interface: Data Location Interface (DLI)  Hierarchical Namespace  Supports sessions and bulk operations LFC GUID SURL 1 SURL 2 ACL LFN 1 LFN 2 lfc-ls –l /grid/vo/ /grid /vo /data file LFC DLI lfc-getacl /grid/vo/data

12. Data Management Command Ref gLite-3.2 User Guide 12 SRMLFCDPM listlcg-ls SRMlfc-ls LFNdpns-ls DPNS upload file lcg-cp LOCAL SRM globus-url-copy LOCAL SRM lcg-cr -d SE -l LFN LOCALrfcp LOCAL DPNS download file lcg-cp SRM LOCAL globus-url-copy SRM LOCAL lcg-cp LFN LOCALrfcp DPNS LOCAL delete file lcg-del --nolfc SRMlcg-del -a LFNrfrm DPNS environment variable LFC_HOST LFC_HOME DPM_HOST DPNS_HOST DPMS_HOME

13. DPM Demonstration Ref gLite-3.2 User Guide 13  Initial Proxy  Check DPM_HOST and DPNS_HOST

14. DPM Demonstration Ref gLite-3.2 User Guide 14  dpns-ls  Upload file to SE

15. DPM Demonstration Ref gLite-3.2 User Guide 15  Remove file

16. SRM Demonstration Ref gLite-3.2 User Guide 16  Initial Proxy  Check SE

17. SRM Demonstration Ref gLite-3.2 User Guide 17  Ls  Upload fire to SE

18. SRM Demonstration Ref gLite-3.2 User Guide 18  Remove file

19. Thank you for your listening and patience Ref gLite-3.2 User Guide 19