Grid Computing Frameworks

33
Seminar: Parallel Computing Grid Computing Frameworks Sabbir Ahmmed

Transcript of Grid Computing Frameworks

Page 1: Grid Computing Frameworks

Seminar: Parallel Computing

Grid Computing Frameworks

Sabbir Ahmmed

Page 2: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 2

Outline

➔ Introduction to Grid Computing

➔ Grid Construction

➔ Grid Frameworks

➔ Globus Toolkit

➔ Gridbus Toolkit

➔ UNICORE

➔ Legion

Page 3: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 3

Outline

➔ Comparison

➔ Other emerging frameworks

➔ Conclusion

Page 4: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 4

Introduction to Grid Computing (I)

End of 1998 the concept of "Grid computing" was introduced in the monograph "The Grid: Blueprint for a New Computing Infrastructure" by I. Foster and C. Kesselman.

Page 5: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 5

Introduction to Grid Computing (II)

➔ The notion of grid computing:➔ The term grid is chosen as an analogy to a power grid ! ➔ Grid computing is a special type of parallel computing➔ How it differs from supercomputing?

➔ Few essential concepts related to grid computing:➔ Utility computing➔ Volunteer computing➔ CPU scavenging➔ Loosely coupled system

➔ Virtual supercomputers

Page 6: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 6

Introduction to Grid Computing (III)

➔ grids typically share at least some of the following characteristics:➔ They are numerous. ➔ They are owned and managed by different, potentially mutually-distrustful organizations and individuals. ➔ They are potentially faulty. ➔ They have different security requirements and policies. ➔ They are heterogeneous, e.g., they have different CPU architectures, run different operating systems, and have different amounts of memory and disk. ➔ They are connected by heterogeneous, multi-level networks. ➔ They have different resource management policies. ➔ They are likely to be geographically-separated (on a campus, in an enterprise, on a continent).

Page 7: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 7

Compute Grids Data Grids

Access Grids Knowledge Grids

Bio Grids Commodity Grids

Campus Grid Tera Grids

Science Grids Sensor Grids Cluster Grids

Introduction to Grid Computing (IV)

Page 8: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 8

➔ Grid Computing Initiatives World-wide ( source: http://www.gridguide.org/ )

Introduction to Grid Computing (V)

Page 9: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 9

➔General Principles (four main aspects characterize a grid)➔ Multiple administrative domain and autonomy➔ Heterogeneity➔ Scalability➔ Dynamicity or adaptability

Grid Construction (I)

Page 10: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 10

➔ The steps necessary to realize a grid➔ Integration of individual software and hardware components➔ Deployment of

➔ Low level middleware➔ User level middleware

➔ Development and optimization of distributedapplications.

Grid Construction (II)

Page 11: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 11

➔ A Layered Grid Architecture

Grid Construction (III)

Page 12: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 12

Grid Construction (IV)

Page 13: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 13

➔ A software framework (or middleware)➔ contains executables or tools➔ provides inversion of control➔ has a default behavior➔ extensibility➔ non-modifiable framework code

Grid Frameworks (I)

Page 14: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 14

● Grid framework (or middleware), is a software stack that facilitates ● writing grid applications ● and manages the underlying grid infrastructure

➔ Grid frameworks can be categorized by the grid layers.

➔ Core middleware and toolkit:➔ Globus, Gridbus (Alchemi, GridSim), UNICORE, Legion, GridGain, gLite

➔ User-level middleware and toolkit:➔ SAGA, MetaMPI, Cactus, GrADS, Gridport, WebFlow, XtremeWeb

Grid Frameworks (II)

Page 15: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 15

➔ Open source software toolkit for developing Grid applications➔ The de facto standard for open source grid computing infrastructure➔ Supported by industry leaders such as IBM, Intel, HP with others (The Globus Consortium)➔ R&D project conducted by the “Globus Alliance”➔ Work on the toolkit first began in 1996. Historically, the Globus Toolkit was used widely by three groups of people

➔ Grid builders➔ Application developers➔ Application framework developers

Globus Toolkit (I)

Page 16: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 16

➔ Provides three main groups of services accessible through a security layer :1. Resource Management2. Data Management3. Information Services

Globus Toolkit (II)

Page 17: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 17

Globus Toolkit (III)

Impact: Globus Toolkit have enabled many exciting new scientific and business grids. (Source: http://www.globus.org/alliance/impact/)

Page 18: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 18

➔ Originated from Gridbus (GRIDcomputing andBUSiness) project.➔ Toolkit for Service Oriented Grid and Utility Computing➔ Supports development of grid infrastructure for eScience and eBusiness applications.➔ Uses economic models (supply and demand) for efficient management of shared resources. ➔ Promotes commoditization of grid services at various levels:

➔ Raw resources level➔ Application level➔ Aggregated service level

Gridbus Toolkit (I)

Page 19: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 19

Gridbus Toolkit (II)

Page 20: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 20

Gridbus Toolkit (III)

Impact: Gridbus Toolkit have enabled several exciting scientific and business grids. (Source: http://www.cloudbus.org/applications.html)

● High Energy Physics and Grid Networks (BelleDataGrid): Melbourne School of Physics

● NeuroGrid: Brain Activity Analysis on the Grid : Osaka University, Japan

● KidneyGrid - Distributed Kidney Models Integration: Melbourne Medical School

● Austronomy: Australian Virtual Observatory

Page 21: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 21

➔ UNICORE (Uniform Interface to Computing Resources) ➔is a ready-to-run Grid system including client and server software

➔ is part of the European Middleware Initiative.➔Project was initially funded by the Federal Ministry of Education and Research (BMBF)

➔ UNICORE was started before "Grid computing"➔ developed by several European partners under the leadership of Jülich Supercomputing Centre. ➔ platform-independent, based on open standards and technologies such as Web Services➔ mostly written in Java and is available as open source under BSD license and available at SourceForge. Current version is UNICORE 6

UNICORE (I)

Page 22: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 22

UNICORE (II)

➔The architecture of UNICORE 6 is three-layered in ➔ client layer, ➔ service layer ➔ and system layer

UNICORE 6 Architecture

Page 23: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 23

UNICORE (III)

Impact: UNICORE6 has been the middleware of choice in numerous grids in EU. (Source: http://www.unicore.eu/community/projects/)

Page 24: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 24

➔ Legion is an object-based metasystem developed at the University of Virginia. ➔ The software developed under the Legion project has been commercialized by a spin-off company called Avaki Corporation➔ The Legion system uses an object-oriented approach. In the Legion system the following apply

➔ Everything is an object.➔ Classes manage their instances➔ Users can define their own classes

➔ The Legion interfaces are described in an Interface Definition Language (IDL).

Legion (I)

Page 25: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 25

➔ Legion core objects support the basic services needed by the metasystem.➔ Legion objects are independent, active, and capable of communicating with each other via unordered non-blocking calls. ➔ Some core objects in Legion are:

➔ Host objects: represent processors in Legion. ➔ Vault objects: represent persistent storage.➔ Context objects: Context objects map context names to Legion object IDs➔ Binding agents: A binding agent maps object IDs to physical addresses➔ Implementation object: hides the storage details of object implementations➔ Class object : is used to define and manage its corresponding Legion object. Class objects are given system-level responsibility.

Legion (II)

Page 26: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 26

Legion (III)

Page 27: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 27

Comparison (I)

Page 28: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 28

Comparison (II)

Page 29: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 29

Comparison (II)

Page 30: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 30

➔ Alchemi: a .NET-based grid computing framework. For more information on Alchemi please visit http://www.alchemi.net/➔ Gridgain 2.0: Java Grid Computing Framework Released by GridGain Systems.

➔ Since its release in August 2007 GridGain became the fastest growing Java grid computing infrastructure with over 10,000 downloads ➔ more than 500 unique projects utilizing it ➔ and deployed in a dozen production systems.

➔ gLite: a framework for building applications tapping into distributed computing and storage resources across the Internet

➔ used in the CERN LHC experiments and in other scientific domains➔ adopted by more than 250 computing centres and used by more than 15000 researchers in Europe and around the world.

Other Emerging Frameworks

Page 31: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 31

➔ Opinion!➔ Critical assessment !!

Conclusion

Page 32: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 32

➔ [01]: Parvin Asadzadeh et. al. “Global Grids and Software Toolkits: A Study of Four Grid Middleware Technologies”.➔ [02]: Mark Baker et. al. “Grids and Grid technologies for wide-area distributed computing”

References

Page 33: Grid Computing Frameworks

30 September 2012 Grid Computing Frameworks 33

Questions