d C a c h e

1. M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 d C a c h e d C a c h e Michael Ernst Patrick Fuhrmann Tigran Mkrtchyan Chep2003 UCSD, California

2. M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 Overview Overview dCache basics Modes of operation Attraction Model Dataset Location Management dCache Components (Overview) Native dCache access method (dCap) The GRID Storage Resource Manager (SRM) The Goal of the SRM Collaboration The Storage System Abstraction An SRM initiated transfer example Major SRM Topics

3. M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 dCache : Modes of Operation dCache : Modes of Operation Used as an HSM Frontend, the dCache provides standard caching mechanisms to optimize tape accesses: Transfer speed adaption Tunable deferred HSM stores (space , time) Automatic staging Continuous garbage collection (no thresholds) Fetch ahead (from HSM) [in preparation]

4. Distributed Pool Architecture M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 externally enforced attraction Super Cluster Caches (Experiment) Cluster Caches (Working Group) Central Cache Tertiary Storage Host Caches Enstore, OSM, … Topic Caches destination determined attraction

5. M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 dCache : Modes of Operation dCache : Modes of Operation - dCache Pools without HSM backend can hold: Precious datasets Files are never automatically removed. System can run out of disk space. Volatile datasets Unused files are automatically removed. System won't run out of disk space. - The dCache can be operated in hybrid mode, running HSM and NON - HSM pools.

6. Typical Tier2 Today … Head/Storage Node GridFTP, bbcp Worker Nodes

7. Tier2 w/dCache Admin Node GridFTP dCache Worker Nodes

8. Tier2 w/dCache UCSD GridFTP Catalog dCache/ Enstore Catalog GridFTP dCache FNAL Florida, … GridFTP GridFTP Catalog CERN Catalog dCache CASTOR

9. M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 dCache : The Attraction Model dCache : The Attraction Model Files are attracted (through store/retrieve requests) by Pools based on : Statically configured parameters, e.g. : Client host IP or subnet numbers HSM groups Subdirectory Trees Dynamically taken parameters from system in operation, e.g. : Pool CPU cost (number of active movers) Pool Space costs (space left, age of datasets)

10. M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 dCache : Dataset Location Management dCache : Dataset Location Management With increasing load on pool nodes, the cache creates dataset replicas on less loaded nodes to smoothen hot spots, featuring Replica Redirection Services. Decreasing load marks dataset replicas for removal in case space is running short. Already staged files can be pinned to disk, overriding dynamic location management. Datasets can be defined 'sticky', independently of their status (CACHED, REPLICATED, ON TAPE ONLY)

11. GFtp Admin SRM Pool Nodes HSMs M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 dCache Components dCache Components I/O Door Nodes Admin Doors (Krb5,ssl) (K)Ftp Http dCap File Name Space File Name Space NFS Server Pool Manager Provider Database pnfs Enstore TSM ? OSM

12. M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 dCache : Native Access Method (dCap) dCache : Native Access Method (dCap) Besides supporting FTP, GridFTP, KerberosFTP and http, dCache has a native access protocol (dCap), allowing POSIX like file operations. dCache provides an implementation of the dCap protocol in C As shared object or preload library For Linux, Solaris and Irix Supporting automatic reconnect on network or server problems Providing security tunnels for Kerberos and ssl. Interfacing ROOT Win dCap .dll support e.t.c.

13. M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 The Storage Resource Manager Collaboration The Storage Resource Manager Collaboration In order to make local storage resources (e.g. disk space, tertiary storage space and large quantities of HEP information) globally available on Data Grids, a collaboration has been setup by LBNL, FermiLab, JLAB and CERN to define an abstract storage system interface, covering methods for : Storing and retrieving datasets Obtaining status information about datasets Pinning datasets (guarantee availability) Negotiating data transfer protocols. Defining dataset lifetimes and possible transitions. HEP sites intending to access remote/provide their storage resources on Data Grids can do so by integrating the SRM protocol into their local storage system. FermiLab is doing this for dCache.

14. SRM Client SRM SRM SRM SRM HPSS M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 Storage System Abstraction Storage System Abstraction Store/Retrieve Data Transfer (e.g. GridFTP) Transfer Protocol Negotiation Client System Pre-staging, Space allocation, Pinning System System Virtual Storage Layer J Cache dCache a s m i HSM OSM Enstore Tsm ? n e

15. SRM Initiated Transfer SRM Initiated Transfer map LFN -> SURL CERN-dataset/user/p/file1 Any SRM Replica Catalog REPLY (SURL) srm://srm.fnal.gov/cern/user/p/file1 GET (SURL) srm://srm.fnal.gov/cern/user/p/file1 d REPLY (Request ID) E FNAL SRM C #-12345 n STAGE a s STATUS (Request ID) c #-12345 t h o REPLY (TURL) e gridftp://pnfs/fnal.gov/cern/user/p/file1 r e GET (TURL) gridftp://pnfs/fnal.gov/cern/user/p/file1

16. M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 SRM Topics (SRM v2.1) SRM Topics (SRM v2.1) Storing and retrieving datasets Transfer Protocol Negotiation (includes direct I/O methods rfio,dCap, etc.) Obtaining dataset status information Pinning datasets (guarantee availability of data to job) File Space Allocation / Reservation Dataset / space lifetime definitions (volatile, durable, permanent) SRM SRM third party transfers Directory support (mkdir / rmdir ) Security ( srm protocol will support gsi over http)

17. References M. Ernst, P. Fuhrmann, T. Mkrtchyan Chep 2003 dCache www.dCache.org SRM http://sdm.lbl.gov/srm-wg/