Grid Computing Hakan ÜNLÜ CMPE 511 Presentation Fall 2004.

Slides:



Advertisements
Similar presentations
Distributed Data Processing
Advertisements

The Anatomy of the Grid: An Integrated View of Grid Architecture Carl Kesselman USC/Information Sciences Institute Ian Foster, Steve Tuecke Argonne National.
FP7-INFRA Enabling Grids for E-sciencE EGEE Induction Grid training for users, Institute of Physics Belgrade, Serbia Sep. 19, 2008.
Grid Resource Allocation Management (GRAM) GRAM provides the user to access the grid in order to run, terminate and monitor jobs remotely. The job request.
Grid Networks. What is Grids? Cluster of clusters – geographically distributed and connected with high-speed MAN and WAN links. Made up of tens to thousands.
High Performance Computing Course Notes Grid Computing.
A Computation Management Agent for Multi-Institutional Grids
1 Software & Grid Middleware for Tier 2 Centers Rob Gardner Indiana University DOE/NSF Review of U.S. ATLAS and CMS Computing Projects Brookhaven National.
USING THE GLOBUS TOOLKIT This summary by: Asad Samar / CALTECH/CMS Ben Segal / CERN-IT FULL INFO AT:
Workload Management Workpackage Massimo Sgaravatto INFN Padova.
Slides for Grid Computing: Techniques and Applications by Barry Wilkinson, Chapman & Hall/CRC press, © Chapter 1, pp For educational use only.
6/2/20071 Grid Computing Sun Grid Engine (SGE) Manoj Katwal.
Office of Science U.S. Department of Energy Grids and Portals at NERSC Presented by Steve Chan.
1-2.1 Grid computing infrastructure software Brief introduction to Globus © 2010 B. Wilkinson/Clayton Ferner. Spring 2010 Grid computing course. Modification.
Milos Kobliha Alejandro Cimadevilla Luis de Alba Parallel Computing Seminar GROUP 12.
UNICORE UNiform Interface to COmputing REsources Olga Alexandrova, TITE 3 Daniela Grudinschi, TITE 3.
Workload Management Massimo Sgaravatto INFN Padova.
1 GRID D. Royo, O. Ardaiz, L. Díaz de Cerio, R. Meseguer, A. Gallardo, K. Sanjeevan Computer Architecture Department Universitat Politècnica de Catalunya.
Grids and Globus at BNL Presented by John Scott Leita.
Grid Computing Net 535.
SPRING 2011 CLOUD COMPUTING Cloud Computing San José State University Computer Architecture (CS 147) Professor Sin-Min Lee Presentation by Vladimir Serdyukov.
Resource Management Reading: “A Resource Management Architecture for Metacomputing Systems”
Ali YILDIRIM Emre UZUNCAKARA
Microsoft Active Directory(AD) A presentation by Robert, Jasmine, Val and Scott IMT546 December 11, 2004.
Dynamic Firewalls and Service Deployment Models for Grid Environments Gian Luca Volpato, Christian Grimm RRZN – Leibniz Universität Hannover Cracow Grid.
DISTRIBUTED COMPUTING
Grid Computing - AAU 14/ Grid Computing Josva Kleist Danish Center for Grid Computing
GT Components. Globus Toolkit A “toolkit” of services and packages for creating the basic grid computing infrastructure Higher level tools added to this.
Computational grids and grids projects DSS,
Grid Workload Management & Condor Massimo Sgaravatto INFN Padova.
CSF4 Meta-Scheduler Name: Zhaohui Ding, Xiaohui Wei
1 Introduction to Microsoft Windows 2000 Windows 2000 Overview Windows 2000 Architecture Overview Windows 2000 Directory Services Overview Logging On to.
Evaluation of Agent Teamwork High Performance Distributed Computing Middleware. Solomon Lane Agent Teamwork Research Assistant October 2006 – March 2007.
The Globus Project: A Status Report Ian Foster Carl Kesselman
Resource Brokering in the PROGRESS Project Juliusz Pukacki Grid Resource Management Workshop, October 2003.
Tool Integration with Data and Computation Grid GWE - “Grid Wizard Enterprise”
Ames Research CenterDivision 1 Information Power Grid (IPG) Overview Anthony Lisotta Computer Sciences Corporation NASA Ames May 2,
NA-MIC National Alliance for Medical Image Computing UCSD: Engineering Core 2 Portal and Grid Infrastructure.
Grid Security: Authentication Most Grids rely on a Public Key Infrastructure system for issuing credentials. Users are issued long term public and private.
GVis: Grid-enabled Interactive Visualization State Key Laboratory. of CAD&CG Zhejiang University, Hangzhou
SOA-9: Implementing SOA in Financial Services Banco Comafi a Real Leading Case Hernan Aymard Sr Solution Architect Javier Betancourt Sr. Project Manager.
GRIDS Center Middleware Overview Sandra Redman Information Technology and Systems Center and Information Technology Research Center National Space Science.
GRID Overview Internet2 Member Meeting Spring 2003 Sandra Redman Information Technology and Systems Center and Information Technology Research Center National.
Campus grids: e-Infrastructure within a University Mike Mineter National e-Science Centre 14 February 2006.
Introduction to Grids By: Fetahi Z. Wuhib [CSD2004-Team19]
Internet2 AdvCollab Apps 1 Access Grid Vision To create virtual spaces where distributed people can work together. Challenges:
International Symposium on Grid Computing (ISGC-07), Taipei - March 26-29, 2007 Of 16 1 A Novel Grid Resource Broker Cum Meta Scheduler - Asvija B System.
Development of e-Science Application Portal on GAP WeiLong Ueng Academia Sinica Grid Computing
Globus Grid Tutorial Part 2: Running Programs Across Multiple Resources.
Super Computing 2000 DOE SCIENCE ON THE GRID Storage Resource Management For the Earth Science Grid Scientific Data Management Research Group NERSC, LBNL.
Tool Integration with Data and Computation Grid “Grid Wizard 2”
Background Computer System Architectures Computer System Software.
The Globus Toolkit The Globus project was started by Ian Foster and Carl Kesselman from Argonne National Labs and USC respectively. The Globus toolkit.
Intro to Distributed Systems Hank Levy. 23/20/2016 Distributed Systems Nearly all systems today are distributed in some way, e.g.: –they use –they.
PARALLEL AND DISTRIBUTED PROGRAMMING MODELS U. Jhashuva 1 Asst. Prof Dept. of CSE om.
INTRODUCTION TO GRID & CLOUD COMPUTING U. Jhashuva 1 Asst. Professor Dept. of CSE.
The EPIKH Project (Exchange Programme to advance e-Infrastructure Know-How) gLite Grid Introduction Salma Saber Electronic.
Workload Management Workpackage
DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING CLOUD COMPUTING
Introduction to Distributed Platforms
Duncan MacMichael & Galen Deal CSS 534 – Autumn 2016
Peter Kacsuk – Sipos Gergely MTA SZTAKI
GWE Core Grid Wizard Enterprise (
Grid Computing.
University of Technology
Cloud Computing.
Grid Computing Software Interface
Chapter 21 Successfully Implementing The Information System
Presentation transcript:

Grid Computing Hakan ÜNLÜ CMPE 511 Presentation Fall 2004

Overview  General Introduction to Grid Computing Introduction: Why Grids? Applications for Grids Basic Grid Architecture Grid Platforms & Standarts  Issues in Grid Computing Hardware: Blade Computers System Management : Globus Toolkit Software: Scheduling

What is Grid Computing?  Computational and Networking Infrastructure that is designed to provide pervasive, uniform and reliable access to data, computational and human resources distributed over wide area environments

Grids Are By Definition Heterogeneous  It’s about legacy resources, infrastructure, applications, policies, and procedures  The grid and its administrators must integrate in stealth mode…with Firewalls Filesystems Queuing systems Grumpy systems administrators Tried and true applications

A Grid Example

Challenges in Grid Computing  Reliable performance  Trust relationships between multiple security domains  Deployment and maintenance of grid middleware across hundreds or thousands of nodes  Access to data across WAN’s  Access to state information of remote processes  Workflow / dependency management  Distributed software and license management  Accounting and billing

Applications for a Grid  Generally, apps that work well on clusters can work well on grids  Non-interactive / batch jobs  Parallel computations with minimal interprocess communication and workflow dependencies  Reasonable data transfer requirements  Sensible economics Productivity Gains > Cost of Building Grid + Opportunity Costs of Resources

Non-Interactive / Batch Jobs  Difficult to get a real-time UI for jobs running on the grid A possible interactive application: spreadsheet computation  Want to take advantage of off-peak free cycles Jobs run for several days, weeks or months The user might prefer to be sleeping while the job runs!  Running processes might need to be interrupted or re-prioritized based on the current load on a grid compute engine Idle thread / “screensaver” computing

Parallel Computations  Application needs to be able to run as multiple, mostly independent pieces Can’t depend on the network’s Quality of Service Can’t rely upon the order of execution and completion Apps that need these things are better suited for tightly coupled compute platforms (e.g. SMP systems) Grid can still be useful as a meta-scheduler and data source for such apps  e.g. the user submits the job to the grid queue and asks for the best available SMP resource

Some Costs and Benefits Costs:  Grid Middleware  Architects and Developers  User Training  Infrastructure Hardware  Opportunity Costs Would a big SMP box return better results for your problem? Benefits:  Better Utilization of Existing Capital Resources  More Efficient Users  Ability to complete more work in the same amount of time Performance near or sometimes as good as the big SMP box

Basic Grid Architecture  Clusters and how grids are different than clusters  Departmental Grid Model  Enterprise Grid Model  Global Grid Model

What Makes a Cluster a Cluster?  Uses a Distributed Resource Manager (DRM) to manager job scheduling  Tightly coupled - High speed, low latency interconnect network  Fairly homogenous - Configuration management is important!  Single administrative domain

The Cluster Model RDPM3ADMMP Operating System StorageCompute Cluster DRM RDPM3ADMMP Operating System StorageCompute Cluster DRM RDPM3ADMMP Operating System StorageCompute Cluster DRM RDPM3ADMMP Operating System StorageCompute Cluster DRM RDPM3ADMMP User Interface/API Cluster DRM Cluster Node High Speed Interconnect Master Node Shared Storage Configuration Management

How is an Enterprise Grid Different from a Cluster?  Heterogeneous - Clusters, SMP, even workstations of dissimilar configurations, but all are tied together through a grid middleware layer  Lightly coupled - Connected via 100 or 1000Mbps Ethernet  Introduces a resource registry and grid security service But usually only a single registry and security service for the grid  Not necessarily a single administrative domain

The Enterprise Grid Model RDPMAADMMP Operating System StorageCompute Cluster Interface RDPMAADMMP Operating System StorageCompute Cluster Interface RDPMAADMMP Operating System StorageCompute Cluster Interface RDPM3ADMMP Operating System StorageCompute Grid Interface RDPM3ADMMP Operating System StorageCompute Grid Interface RDPM3ADMMP User Interface/API Grid Interface SMP Enterprise LAN or WAN Security Infrastructure Resource Registry Grid Interface Cluster DRM RDPMAADMMP Operating System StorageCompute Cluster Interface RDPMAADMMP Operating System StorageCompute Cluster Interface RDPMAADMMP Operating System StorageCompute Cluster Interface Grid Interface Cluster DRM RDPM3ADMMPRDPM3ADMMP

How is a Global Grid Different from an Enterprise Grid?  "Grid of Grids" - Collection of enterprise grids  Loosely coupled between sites - Not much control over Quality of Service  Mutually distrustful administrative domains  Multiple grid resource registries and grid security services

The Global Grid Model Grid WAN RRSI Cluster Grid SMP Grid SMP Grid Cluster UI/API Grid LAN Grid RRSI SMP Grid SMP Grid SMP Grid Cluster RRSI ClusterSMP Grid Cluster Grid LAN Site A Site B Site C UI/API Grid UI/API Grid LAN

Grid Platforms & Standards  The Global Grid Forum  Globus Toolkit  DCML (Data Center Markup Language)

Globus Toolkit V2 “Pillars” Information Services (MDS) Data Management (GASS) Resource Management (GRAM) Grid Security Infrastructure (GSI)

Globus Toolkit V2 Stack MDSGASS/GridFTPGRAM GSI HTTPLDAPFTP TLS/SSL TCP/IP

Globus Toolkit V2 Key Components: GRAM, MDS and GASS  Grid Resource Allocation Manager (GRAM) Server-side: “gatekeeper” process that controls execution of job managers Client-side: “globusrun” UI to launch jobs  Monitoring and Directory Service (MDS) GRIS: Grid Resource Information Service collects local info GIIS: Grid Index Information Service collects GRIS info  Global Access to Secondary Storage (GASS) GridFTP, implemented through “in.ftpd” daemon and “globus-url-copy” command Files accessed through a URI, e.g. gsiftp://node1.ncbiogrid.org/data/ncbi/ecoli.nt

Globus Toolkit V2 Additional Components  Grid Packaging Tools (GPT) Used to build (“gpt-build”), install (“gpt- install”) and localize (“gpt-postinstall”) Globus components  MPICH-G2 A Globus V2 enabled version of MPI (Message Passing Interface) Based on MPICH Utilizes GSI, MDS and GRAM

Globus Toolkit V2 Network Services Certificate Authority GIIS Server GRIS gatekeeper in.ftpd Grid Node GRAM Client Client Node GRIS gatekeeper in.ftpd Grid Node GRIS gatekeeper in.ftpd Grid Node GRIS gatekeeper in.ftpd Grid Node Network

GRAM, MDS and GASS Interactions resource process job manager gatekeeper process GRAM GRIS resource GIIS MDS GridFTP in.ftpd GASS job allocation job management resource discovery data transfer data control user / proxy Client RSL/DUROC/HTTP 1.1LDAP gsiftp

Globus Toolkit V2 Strengths and Weaknesses Strengths:  Mindshare and collaboration in both industry & academia  Open source  Standards-based underpinnings (e.g. SSL, LDAP)  Flexibility and CoG API's  Driving OGSA with heavy resource commitment from IBM Weaknesses:  Significant effort required to get applications working on a grid  Not production quality at this time  No “metascheduler” -- user has to explicitly tell their jobs where to run

Issues in Grid Computing Hardware : Blades

Hardware Trends  HW Trends that enable Grids and Distributed Processing There is a lot of idle computing power Computers are now better connected There are many different brands and configurations in any environment  And Distributed Computing that give rise to new HW architectures Blade Computers

What is a blade?  Inclusive chassis-based modular computing system that includes processors, memory, network interface cards and local storage on a single board. Blade Blade Chasis & Blades Blade Farm

Anatomy of a blade

How far it can go?

Advantages & Disadvantages  Low Cost (power, heat, data center space)  Physical Server Consolidation (Save space, eliminate cables)  High Availability  Integrated Systems Management  Not suitable in small numbers  Need for standardization (for network connection and management)

Blades & Grid  Each blade is a server that can run jobs.  Blades can be used to form clusters or grids.  With efficient management different configurations of blades can be used in a single grid computer. Easy to expand Protects investment

Issues in Grid Computing System Management : Globus Toolkit

Globus Toolkit V2 “Pillars” Information Services (MDS) Data Management (GASS) Resource Management (GRAM) Grid Security Infrastructure (GSI)

Globus Toolkit V2 Stack MDSGASS/GridFTPGRAM GSI HTTPLDAPFTP TLS/SSL TCP/IP

Globus Toolkit V2 Key Components: GRAM, MDS and GASS  Grid Resource Allocation Manager (GRAM) Server-side: “gatekeeper” process that controls execution of job managers Client-side: “globusrun” UI to launch jobs  Monitoring and Directory Service (MDS) GRIS: Grid Resource Information Service collects local info GIIS: Grid Index Information Service collects GRIS info  Global Access to Secondary Storage (GASS) GridFTP, implemented through “in.ftpd” daemon and “globus-url-copy” command Files accessed through a URI, e.g. gsiftp://node1.ncbiogrid.org/data/ncbi/ecoli.nt

Globus Toolkit V2 Additional Components  Grid Packaging Tools (GPT) Used to build (“gpt-build”), install (“gpt- install”) and localize (“gpt-postinstall”) Globus components  MPICH-G2 A Globus V2 enabled version of MPI (Message Passing Interface) Based on MPICH Utilizes GSI, MDS and GRAM

Globus Toolkit V2 Network Services Certificate Authority GIIS Server GRIS gatekeeper in.ftpd Grid Node GRAM Client Client Node GRIS gatekeeper in.ftpd Grid Node GRIS gatekeeper in.ftpd Grid Node GRIS gatekeeper in.ftpd Grid Node Network

GRAM, MDS and GASS Interactions resource process job manager gatekeeper process GRAM GRIS resource GIIS MDS GridFTP in.ftpd GASS job allocation job management resource discovery data transfer data control user / proxy Client RSL/DUROC/HTTP 1.1LDAP gsiftp

Globus Toolkit V2 Strengths and Weaknesses Strengths:  Mindshare and collaboration in both industry & academia  Open source  Standards-based underpinnings (e.g. SSL, LDAP)  Flexibility and CoG API's  Driving OGSA with heavy resource commitment from IBM Weaknesses:  Significant effort required to get applications working on a grid  Not production quality at this time  No “metascheduler” -- user has to explicitly tell their jobs where to run

Issues in Grid Computing Software : Scheduling

Superscheduling  Superscheduling means scheduling resources in multiple administrative domains.  Various models Submiting a job to a specific single machine Submiting a job to single machines at multiple sites (With cancellation option) Scheduling a single job to use multiple resources  Most common superscheduler : USERS

Phases Of Superscheduling  Resource Discovery Authorisation Filtering Application Requirement Definition Minimal Requirement Filtering  System Selection Gathering Information (Query) Select Systems to run on  Run the Job Make an Advance Reservation (Optional) Submit Job to Resources Preperation Tasks Monitor Progress Job Completion Completion Tasks Source : Global Grid Forum, Scheduling Working Group, 10 Actions When Scheduling, Schopf, 2001

Scheduling Framework (Ranganathan & Foster 2003)  External Scheduler  Local Scheduler  Dataset Scheduler

Scheduling And Replication Algorithms  External Scheduler JobRandom JobLeastLoaded JobDataPresent JobLocal  Dataset Scheduler DataDoNothing: No Active Replitication. Everything is on demand DataRandom: Popular Datasets are replicated to Random Sites DataLeastLoaded: Popular Datasets are snet to the least loaded sites.

Simulation Results Average Response TimesAverage Data Transfered

Grid Computing Thank You and Questions?