1. Kento Aida, Tokyo Institute of Technology 1 Tutorial: Technology of the Grid 1. Definition 2. Components 3. Infrastructure Kento Aida Tokyo Institute of Technology

2. Kento Aida, Tokyo Institute of Technology 2 Goal of the Tutorial What is the grid? definition What technology is needed to create the grid? component technology How is the grid environment is constructed? infrastructure

3. Kento Aida, Tokyo Institute of Technology 3 1. Definition

4. Kento Aida, Tokyo Institute of Technology 4 Definition of the Grid Definition [http://www.jpgrid.org/about/index.html] The grid is an infrastructure to dynamically organize a virtual organization (or a virtual computer) on demand by virtualizing and integrating resources such as computers, data, experimental devices, sensors, people. (The original definition is written by Japanese.) What is the grid? A three point checklist [http://www.gridtoday.com/02/0722/100136.html] coordinates resources that are not subject to centralized control using standard, open, general-purpose protocols and interfaces to deliver nontrivial qualities of service

5. Kento Aida, Tokyo Institute of Technology 5 What can we do using the grid? We can use information resources (services) on network securely (to guarantee security), stably (to use required resources on demand), and easily (without knowledge of network, computers, …). Internet virtual organization

6. Kento Aida, Tokyo Institute of Technology 6 Examples of Virtual Organizations Members in a collaborative research project Researchers in a collaborative research project share resources distributed over their sites, e.g. universities, institutes, laboratories, ….  large-scale scientific computing  large-scale distributed database Project team in a company Members in a project team share resources distributed over multiple branches in a company.  business  transaction

7. Kento Aida, Tokyo Institute of Technology 7 Definition of the Grid (again) Definition The grid is an infrastructure to dynamically organize a virtual organization (or a virtual computer) on demand by virtualizing and integrating resources …. What is the grid? A three point checklist coordinates resources that are not subject to centralized control  dynamic organization of VO using standard, open, general-purpose protocols and interfaces  access to resources by standardized protocols to deliver nontrivial qualities of service  Users do not have to have knowledge about network, computers, etc.

8. Kento Aida, Tokyo Institute of Technology 8 Grid? Grid = supercomputer + network? Grid = idle PCs + network? Grid = large-scale parallel processing on the internet? If we connect our resources to the grid, anonymous users’ jobs will run on our resources without owners’ knowledge? If we submit jobs to the grid, our job will run on resources in unknown sites?

9. Kento Aida, Tokyo Institute of Technology 9 Classification of the Grid Computing Grid （ high-performance computing ） Data Grid （ high-performance data processing ） Sensor Grid （ advanced sensing ） Access Grid （ support for collaboration ） Business Grid （ advanced web service ） PC Grid （ utilization of idle PCs ） science business

10. Kento Aida, Tokyo Institute of Technology 10 Computing Grid Grid computing high-performance computing service to utilize computers on the grid Merit of users reducing computation time expanding problem size receiving computation service Component technology security, resource management, job management, programming, problem solve environment (PSE), …

11. Kento Aida, Tokyo Institute of Technology 11 Data Grid Large-scale data processing/computing large-scale distributed database on the internet data processing service to access distributed data Merit of users high-speed access to distributed data high-performance and reliable processing using large- scale data Component technology security, high-speed data transfer, replica management, scheduling

12. Kento Aida, Tokyo Institute of Technology 12 Access Grid Communication support on the grid Example remote conference virtual laboratory remote medical service  SARS Grid (NCHC) entertainment  “KARAOKE” Grid (AIST)

13. Kento Aida, Tokyo Institute of Technology 13 Sensor Grid Advanced Monitoring coordination of autonomous sensors connected by network  wired network, wireless network, satellite, … advanced sensing, analysis, forecasting Example meteorology (weather forecast), ecology, agriculture, …

14. Kento Aida, Tokyo Institute of Technology 14 Component technology security, information service, resource management job management, scheduling data management programming problem solve environment (PSE) Infrastructure production grid Application applying to big science applying business Technical Issues of the Grid

15. Kento Aida, Tokyo Institute of Technology 15 2. Components

16. Kento Aida, Tokyo Institute of Technology 16 Component Technology of the Grid resource management job management programming problem solve environment security application data management infrastructure (computer, network, experimental device, …) information service

17. Kento Aida, Tokyo Institute of Technology 17 Security Issues authentication, encryption of communication Single sign on user authentication on one host Authentication on other hosts is automatically performed. internet user authentication authentication is automatically performed. authentication Org. A Org. B Org. C

18. Kento Aida, Tokyo Institute of Technology 18 Resource Management internet user GW OS A GW OS B GW OS C com. a com. c com. b common command Org. A Org. B Org. C Common interfaces to the grid wrapping differences of commands/operations among different machines

19. Kento Aida, Tokyo Institute of Technology 19 Information Service internet GW Org. A Org. B Org. C CPU: … memory: … OS: … CPU: … memory: … OS: … info. service network monitoring Information about resources on the Grid

20. Kento Aida, Tokyo Institute of Technology 20 Big picture of the GT2 Client Site C Site B CA GRIS GIIS gatekeeper User Cert. Proxy Cert. Proxy Cert. Proxy Cert. grid-proxy-init Query Resource Status GridFTP Server Data Transfer process Process invocation Return result [source: Yoshio Tanaka, AIST]

21. Kento Aida, Tokyo Institute of Technology 21 Job Management user (0) (2) (1,3,4) (4) internet GW Org. A Org. B Org. C info. service resource broker Resource selection, Scheduling, Job control

22. Kento Aida, Tokyo Institute of Technology 22 Condor High Throughput Computing matching jobs and resources by ClassAds mechanism fault tolerance by check pointing Implementation on the Globus Tool Kit Condor-G Schedd Match maker Startd Client job [ http://www.cs.wisc.edu/condor/ ] owner: aaa CPU: 2GHz 以上 Memory: 512MB 以上 Disk: 10GB 以上 :

23. Kento Aida, Tokyo Institute of Technology 23 Scheduling Application scheduling Scheduling of a single application (job) on resources  How do we decompose an application program into tasks?  Where do we allocate tasks?  When do we start execution of tasks? Job scheduling Scheduling of multiple jobs on resources  Where do we dispatch jobs on resources?  When do we start execution of jobs? Goal minimizing the execution time, meeting the deadline, minimizing the cost, preserving fairness, …

24. Kento Aida, Tokyo Institute of Technology 24 Nimrod Job management system for parameter-survey applications computational economy deadline scheduling Implementation on the Globus Tool Kit Nimrod/G [ http://www.csse.monash.edu.au/~davida/nimrod.html/ ] [source: D. Abramson, et.al., “High Performance Parametric Modeling with Nimrod/G: Killer Application for the Global Grid?,” IPDPS2000, 2000 ]

25. Kento Aida, Tokyo Institute of Technology 25 Data Management user GW Org. A Org. B Org. C replication high-speed file transfer file internet GW data management Distributed file management, High-speed file transfer, Replica management

26. Kento Aida, Tokyo Institute of Technology 26 Data Grid Applications High Energy Physics Earth Science, Astronomical Observation Bio informatics [source: Osamu Tatebe, AIST]

27. Kento Aida, Tokyo Institute of Technology 27 Grid Datafarm Peta-to-Exascale Global Filesystem on unified CPU/storage cluster Parallel I/O and parallel processing with local I/O scalability [source: Osamu Tatebe, AIST]

28. Kento Aida, Tokyo Institute of Technology 28 Trans-Pacific Gfarm Datafarm testbed: Network and cluster configuration 2.4G 10G 1G 2.4G(1G) 1G SuperSINET APAN/TransPAC Los Angeles 622M AIST Titech Maffin 10G APAN Tokyo XP SuperSINET Tsukuba WAN 10G 2.4G New York OC-12 ATM SC2003 Phoenix 32 nodes 23.3 TBytes 2 GB/sec 5G 16 nodes 11.7 TBytes 1 GB/sec 16 nodes 11.7 TBytes 1 GB/sec 7 nodes 3.7 TBytes 200 MB/sec 10 nodes 1 TBytes 300 MB/sec 147 nodes 16 TBytes 4 GB/sec Indiana Univ Kasetsert Univ, Thailand SDSC Trans-Pacific thoretical peak 3.9 Gbps Gfarm disk capacity 70 TBytes disk read/write 13 GB/sec Chicago Abilene KEK Univ Tsukuba NII 1G [2.34 Gbps] [950 Mbps] [500 Mbps] [source: Osamu Tatebe, AIST]

29. Kento Aida, Tokyo Institute of Technology 29 Programming MPI programming with Message Passing Interface  MPICH-G2 ， GridMPI ， … GridRPC programming with Remote Procedure Call (RPC) mechanism  Ninf-G ， OmniRPC ， NetSolve ， … Master Worker Template template to develop master-worker programs  MW ， AMWAT ， …

30. Kento Aida, Tokyo Institute of Technology 30 GridRPC ------ for (…) { grpc_call_async( ) } ------ for (…) { grpc_call_async( ) } ------ library program input data output data user program internet master worker

31. Kento Aida, Tokyo Institute of Technology 31 GridRPC （ cont’d ） for (i = start; i <= end; i++) { SDP_search(argv[1], i, &value[i]); } grpc_function_handle_init(&hdl, …, “SDP/search”); for (i = start; i <= end; i++) { grpc_call_async(&hdl, argv[1], i, &value[i]); } for (i = start; i <= end; i++) { SDP_search(argv[1], i, &value[i]); } grpc_function_handle_init(&hdl, …, “SDP/search”); for (i = start; i <= end; i++) { grpc_call_async(&hdl, argv[1], i, &value[i]); } Ninf-G [ http://ninf.apgrid.org/ ] reference implementation of GridRPC implementation on the Globus Toolkit  using security functions on the Globus (authentication, encrypted communication).

32. Kento Aida, Tokyo Institute of Technology 32 Problem Solve Environment (PSE) Portal frontend to search, run, monitor, and control applications on the grid  Web page cooperation with a workflow system Workflow mechanism to run multiple applications following their dependencies  representing dependencies among applications by a graph  initiation of applications following the workflow by the workflow engine

33. Kento Aida, Tokyo Institute of Technology 33 Example of PSE (UNICORE) [source: http://www.unicore.org/unicore.htm]

34. Kento Aida, Tokyo Institute of Technology 34 3. Infrastructure

35. Kento Aida, Tokyo Institute of Technology 35 Resources in Grid Infrastructure Computer PC, PC cluster, supercomputer, … Storage HDD, RAID, … [source: http://www.gsic.titech.ac.jp/Japanese/Service /R_System/Overview/index.html] [source: Matsuoka Lab, TITECH]

36. Kento Aida, Tokyo Institute of Technology 36 Resources in Grid Infrastructure (cont’d) Experimental device microscope, accelerator, … Sensor thermometer, camera, … Ultra-High Voltage Electron Microscope, Osaka University [source: http://www.biogrid.jp/] Large Hadron Collider, CERN [source: Osamu Tatebe, AIST] EcoGrid, NCHC [source: Fang Pang Lin, NCHC]

37. Kento Aida, Tokyo Institute of Technology 37 Resources in Grid Infrastructure (cont’d) Network LAN, WAN, internet, … [ source: http://www.apan.net/] [source: http://www.noc.titech.ac.jp/titanet/ supertitanet/index.ja.shtml]

38. Kento Aida, Tokyo Institute of Technology 38 Grid Infrastructure Classification by objectives test bed the grid environment construct to perform experiment.  temporally available production grid the grid environment for production use, or to run practical applications  permanently available.  Resources are fully operated for 24hrs. Classification by geographic sites department grid, campus grid, national grid, international grid

39. Kento Aida, Tokyo Institute of Technology 39 ACT-JST Testbed Grid testbed for running applications to solve large-scale optimization problem construction of 1000CPU scale testbed application development collaboration among Grid researchers and application scientists TITECH AIST TDU Tokushima U.

40. Kento Aida, Tokyo Institute of Technology 40 Grid Challenge Federation (GCF) Test bed constructed for the Grid Challenge event, programming contest on the grid Resources Grid Technology Research Center, AIST HPCS Lab., U. Tsukuba Yuba-Honda Lab., UEC Matsuoka Lab., TITECH Aida Lab., TITECH Ono Lab., Tokushima U. Hiraki Lab., U. Tokyo Chikayama-Taura Lab., U. Tokyo

41. Kento Aida, Tokyo Institute of Technology 41 ApGrid ／ PRAGMA [ source: http://www.apgrid.org/] Grid Partnership among Asia-Pacific region

42. Kento Aida, Tokyo Institute of Technology 42 Titech Grid [source: http://www.gsic.titech.ac.jp/index-j.html]

43. Kento Aida, Tokyo Institute of Technology 43 NAREGI [source: http://www.naregi.org/ ]

44. Kento Aida, Tokyo Institute of Technology 44 TeraGrid [source: http://www.teragrid.org/] The 40Gbps network connects sites. 20TeraFlops ， 1PB resources CalTech, ANL, SDSC, NCSA, PSC

45. Kento Aida, Tokyo Institute of Technology 45 Operation of Infrastructure Objectives An organization/staff is required to stably provide a grid infrastructure to users.  The current internet is operated by experts (organizations) for network operation. Network Operation Center (NOC) Grid Operation Center organization to operate a grid infrastructure providing information of grid resources  resources in VO  load on computing resources, traffic on networks, … user support  accounting, documents archives, help desk, trouble shooting, …

46. Kento Aida, Tokyo Institute of Technology 46 PRAGMA GOC

47. Kento Aida, Tokyo Institute of Technology 47 Network Weather Map http://mrtg.koganei.itrc.net/mmap/grid.html Thanks: Dr. Hirabaru and APAN Tokyo NOC team

48. Kento Aida, Tokyo Institute of Technology 48