Grids and Grid Technologies for Wide-Area Distributed Computing Mark Baker, Rajkumar Buyya and Domenico Laforenza.

Grids and Grid Technologies for Wide-Area Distributed Computing

Mark Baker, Rajkumar Buyya and Domenico Laforenza

Agenda

• Overview of Grid

• Concerning issues in Grid

• Components and Services in Grid

• Current Projects and Future

• Grid in Multimedia Applications

What is Grid?

• In the last decade, the availability of cheap computational resources and high-speed networks leads to an opportunity of using distributed computers as a single, unified computing resources.

• This is known as Grid Computing

What is Grid?

• The term Grid is chosen as analogy to power Grid that provides consistent and transparent access to electricity irrespective to its sources

• A Grid can be viewed as seamless, integrated computational and collaborative environment

What is Grid?

• Some well-known name of Grid Computing– Meta Computing– Scalable Computing– Global Computing– Internet Computing– Peer-to-Peer Computing

Construction of Grid

• The four main issues in designing and constructing a Grid– Multiple Administrative Domains and

Autonomy– Heterogeneity– Scalability– Dynamicity and Adaptability

Construction of Grid

• Multiple Administrative Domains and Autonomy– A Grid Infrastructure is supposed to be

geographically distributed– It is owned by different individuals and

organizations.

• Heterogeneity– Computational resources are

heterogeneous in nature

Construction of Grid

• Scalability– Ranging from a few to million– Raising a problem of performance

degradation as population increases

• Dynamicity and Adaptability– Failure is a rule, not exception– Services or applications must use

resources dynamically, effectively and efficiently.

Grid Components

• Conceptually, a grid is designed based on four layers– Grid Fabric– Core Grid Middleware– User level Middleware– Applications

Grid Components

• Grid Fabric– It manages the network resources, like

computers, data storage, routers and input devices etc. across the network geographically.

• Core Grid Middleware– It manages how the network resources

co-operating together. It provides the services like remote process management and QoS etc.

Grid Components

• User Level Middleware– It provides an application development

environment offering for high-level application.

• Applications

Assumptions

• No intervention with existing technologies– Do not interfere with the existing site

administration or autonomy.– Do not compromise existing security of users

and remote sites– Do not need to replace the existing network

protocol, operating system or services. Use existing standard

– Use standard, existing protocols and is able to interact with the legacy applications

Assumptions

• Support reliable and fault tolerant system

• Support for heterogeneity

• Allow computational resources join or leave at any time

Services of Grid

• Administrative Hierarchy– This determine how the administrative

information flows through the Grid

• Communication Services– This requires to the Grid to support

protocol for different data transfer, like streaming, group communication and QoS etc.

Services of Grid

• Information Services– A major goal is to make all resources

accessible to any process in the Grid.– Grid is a dynamic environment– This services must provide a

mechanism for registering and obtaining information of the Grid structure, resources, services and status

Services of Grid

• Naming Services– This services must provide a uniform

name space across the complete Grid environment

• Distributed File Systems and Caching– Same as the Naming Services, this

provides a uniform name space for retrieving file.

– Support I/O protocols, performance enhancement etc.

Services of Grid

• Security and Authorization– This provides a security infrastructure

for managing and utilizing the Grid resources.

• System Status and Fault Tolerant Services– Failure is a rule in the Grid, not

exception– This provides a reliable and robust

environment

Services of Grid

• Resource Management and Scheduling– This is responsible for scheduling all the

processes within the Grid to utilizing the resources, like processing power, storage etc.

• Programming Tools– This provides an interface for

development

• User Administration and GUI

Grid Computing Projects

• Main Initiatives– Global Grid Forum

• http://www.gridforum.org

– Peer-to-Peer Working Group (P2PWG)• http://www.p2pwg.org

• Three important technologies in Grid Computing nowadays– Java Technology– Common Object Request Broker Architecture

(CORBA)– Distributed Common Object Model (DCOM)

Current Projects

• Active Sheets• Compute Power Market• DISC World• GridSim• Nimrod/G & GRACE• Virtual Lab• World Wide Grid (WWG)• More ……

Future

• Grid resources can be used to solve grand challenges in a wide scope of fields, like physics etc.

Grid in Multimedia Applications• Administrative Hierarchy• Communication Services• Information Services• Naming Services• Distributed File Systems and Caching• System Status and Fault Tolerant Services• Security and Authorization• Resource Management and Scheduling• Programming Tools• User Administration and GUI

Grid in Multimedia Applications

• Communication Services– It is a problem of how a group of clients

streaming data so as to minimize their costs under the current system status

• Distributed File Systems and Caching– It is a problem of how data are stored in

a group of clients under the current system status

Grid in Multimedia Applications

• System Status and Fault Tolerant Services• Resource Management and Scheduling• Information Services

– It is a problem of how to manage the join/leave action of clients in the multimedia applications in which it changes the system status

– How such changes affect the decision made by the previous 2 services

Thank you