UNIT-I

Introduction

Grid Computing :
• Networks numerous computing devices in a unified "grid."
• Functions like a utility power grid delivering computing resources.
Key Features:
• Adds infinite computing devices to the grid.
• Enhances computing capabilities and problem-solving tasks.
Simple use cases of Grid Computing, are as follows:
Financial Organization Use Case:
• Pools computational resources for wealth management applications.
• Achieves faster task execution and real-time data access.
• Enhances customer satisfaction through reduced turnaround time.
Scientific Research Use Case:
• Scientists studying the ozone layer manage massive experimental data.
• Utilizes geographically dispersed, efficient storage systems.
• Facilitates effective and efficient scientific research processes.
Massive Multiplayer Gaming Use Case:
• Requires numerous servers for real-time, global interactions.
• Dynamically allocates and manages resources based on workload.
• Supports larger gaming communities with scalable infrastructures.
• Enhances profits and customer satisfaction for gaming corporations.
Government Disaster Management Use Case:
• Simulates models for events like chemical spills to assess impacts.
• Collaborates across departments to manage and plan responses.
• Focuses on public safety, wildlife protection, and ecosystem preservation.
Foundations of Grid Computing:
• Coordinated Resources:
• Decentralized systems with coordination based on policies and SLAs.
• Open Standards:
• Ensures interoperability for resource discovery, access, and coordination.
• Quality of Service (QoS):
• Key features include:
• Response time and performance.
• Security and scalability.
• Availability and failover mechanisms.
• Autonomic features like event correlation and configuration management.
Computational and Data Grids

Data Management in Grid Computing:

• Integration: Support multiple distributed, heterogeneous, and independent data
sources.
• Efficient Transfer: Ensure data transfer to computation locations for scalability.
• Caching/Replication: Minimize network traffic via data caching or replication.
• Data Discovery: Enable users to locate data based on specific characteristics.
• Security: Implement encryption and integrity checks for secure data transport.
• Backup/Restore: Provide mechanisms to prevent data loss and minimize
downtime.
Computation in Grid Computing:
• Independent Management: Allow independent management of computing
resources.
• Resource Selection: Provide intelligent and transparent resource selection
mechanisms.
• Resource Awareness: Monitor current/predicted loads, availability, and dynamic
configurations.
• Failure Management: Implement failure detection and failover mechanisms.
• Security: Ensure secure resource management, access, and integrity.
Evolution of Computing Power:
•Home-based PCs are now capable of complex computations once unimaginable.
•Business applications demand more complex computing for tasks like simulations,
scientific experiments, and real-time portfolio management.
Challenges in Computational Power:
•Some applications exceed the computational capacity of a single organization.
•Similar to the early 1900s' electric power needs, where users had to operate their
own generators.
The Concept of Computational Grids:
•The shift to computational grids provides demand-driven, reliable, and cost-
effective computing power.
•Just like the electric grid revolutionized power utilization, computational grids
transform the accessibility
• of computing resources.
 THE GRID PROBLEM

Grid Computing and its Evolution:

• Differentiates from distributed computing with a focus on resource sharing,
coordination, and high performance.
• Solves problems of resource sharing among individuals/groups.
• Challenges increase when Grid Computing is applied to utility computing, where
industrial resources are shared.
Example: IBM's Business On Demand:
• Demonstrates resource implementation in Grid Computing.
• Introduces challenges such as:
• Service-level management.
• Accounting and utilization metering.
• Flexible pricing and federated security.
• Scalability and complex networking services.
Key Issues in Resource Sharing:
• Resource discovery.
• Event correlation.
• Authentication, authorization, and access control.
 The Concept of Virtual Organizations

Virtual Organizations in Grid Computing:

• Defined as dynamic groups of individuals/institutions around resource-sharing
rules.
• Share common concerns but vary in size, scope, duration, and structure.
• Members negotiate resource sharing based on agreed-upon rules and conditions.
Key Challenges in Virtual Organizations:
• Assigning users, resources, and organizations across global domains.
• Issues include:
• Defining resource discovery mechanisms.
• Determining resource sharing methods and rules.
• Ensuring security, federation, delegation, and access control.
Examples of Virtual Organizations:
• CERN Data Grid
• Thousands of physicists collaborate across laboratories.
• Create, design, and analyze data from a major detector at CERN (European
high-energy physics lab).
• Form a "data grid" for intensive resource sharing:
• Computing power
• Data storage
• Network services
• Analyze petabytes of data generated by the detector.
Financial Modeling Organization
• Dynamic virtual organization formed for financial forecasting.
• Utilizes data from internal and external sources.
• Provides:
• Real-time financial market analysis.
• Forecasting and advisory capabilities for investment portfolios.
• Benefits:
• Reduced customer wait times.
• Increased reliability using real-time data and advanced computational
techniques.
 Grid Architecture
•A new architecture developed for managing cross-organizational resource sharing
•within a virtual organization.
•Focuses on defining, establishing, and managing interoperability among resources.
Key Features
•Basic Components: Identifies components of a grid system.
•Purpose & Functions: Defines their roles and interactions.
•Interoperability:
•Emphasis on shared relationships between resource providers and users.
•Uses common protocols at each layer for seamless integration.
Grid Protocol Architecture
•A structured model that ensures standardized communication and interaction.
•Facilitates efficient and unified resource sharing.
Significance
•Enhances coordination across multiple organizations.
•Ensures scalable and flexible resource management.
Figure 1

This protocol architecture defines common mechanisms, interfaces, schema,

and protocols at each layer, by which users and resources can negotiate,
establish, manage, and share resources.
Fabric Layer: Interface to Local Resources
Definition
•Specifies the resources available for sharing in a grid system.
Types of Resources
•Physical Resources: Computational resources, data storage, networks, etc.
•Logical Resources:
•Distributed file systems (e.g., NFS).
•Computer clusters and pools.
•Software applications.
•Advanced networking services.
Implementation
•Logical resources rely on internal protocols:
•NFS (Network File Systems) for distributed systems.
•LFS (Logical File Systems) for clusters.
Integration
•Logical resources are interconnected to form a network of physical resources
Connectivity Layer: Manages Communications
Purpose
•Defines core communication and authentication protocols for grid-specific
networking services.
Communication Protocols
•Facilitates data exchange between Fabric layer resources.
•Works with networking layer protocols (e.g., TCP/IP Internet protocol stack).
•Includes aspects like transport, routing, and naming in networking.
Authentication Protocols
•Provides secure authentication and data exchange between users and resources.
Key Features:
•Single Sign-On (SSO): Authenticate once, access all resources without
further intervention.
•Delegation: Resources can relay user credentials (or subsets) for additional
resource access.
Integration with Local Security Solutions
•Supports diverse local security methods, including: Kerberos, Windows, Linux,
and UNIX security.
•Ensures seamless compatibility with local environments.
User-Based Trust Relationships
•Critical for establishing trust between users and multiple service providers.
•Eliminates the need for provider-to-provider interaction for resource access.
Data Security
•Ensures data integrity and confidentiality during transactions.
•Employs cryptographic and encryption mechanisms to secure data.
 Resource Layer: Sharing of a Single Resource
• The Resource Layer controls secure operations (negotiation, initiation, monitoring, metering,
accounting, and payment) involving resource sharing.
• Uses communication and security protocols defined by the Networking Communications Layer.
• Calls Fabric Layer functions to manage individual local resources.
• Handles only individual resources, excluding global states or atomic actions across resource
collections (handled by the Collective Layer).
Key Components
• Information Protocols:
• Functionality: Retrieve information about a resource's structure and operational state.
• Data: Configuration, usage policies, service-level agreements (SLAs), and current
resource state.
• Purpose: Monitor resource capabilities and identify availability constraints .
Management Protocols:
Core Functions:
• Negotiation:
• Manages access agreements for shared resources.
• Considers quality of service (QoS), reservations, and scheduling.
• Operations:
• Executes processes (e.g., process creation, data access).
• Acts as a policy enforcement point, ensuring alignment with
service/resource policies.
• Accounting and Payment:
• Tracks usage and handles payment for resource sharing.
• Monitoring and Control:
• Tracks operation statuses.
• Provides controls (e.g., termination, asynchronous notifications) to
ensure operational integrity.
The Collective Layer: Coordinating Multiple Resources
• Manages global resource interactions and coordination within a collection of
resources.
• Implements advanced sharing behaviors using protocols from the Resource and
Connectivity layers.
• Focuses on collective operations that span multiple resources or systems within a
virtual organization.
Key Services in the Collective Layer
• Discovery Services:
Purpose: Allows participants to locate and understand available resources
within a virtual organization.
Capabilities: Identifies resource existence and their properties.
• Coallocation, Scheduling, and Brokering Services:
Purpose: Allocates multiple resources for specific tasks during designated time
periods.
Capabilities:
• Coordinates resource allocation.
• Schedules tasks on the appropriate resources.
• Monitoring and Diagnostic Services:
Purpose: Ensures resilience and reliability of resources through failure recovery
and diagnostic insights.
Capabilities:
• Resource failure recovery.
• Monitoring of networking and device services.
• Event logging and intrusion detection.
• Data Replication Services:
Purpose: Optimizes storage resource usage for enhanced data access
performance.
Capabilities:
• Balances response time, reliability, and cost.
• Grid-Enabled Programming Systems:
Purpose: Facilitates the use of familiar programming models within Grid
Computing environments.
Capabilities:
• Supports grid-specific networking services like resource discovery,
allocation, and problem resolution.
• Workload Management Systems and Collaborative Frameworks:
Purpose: Manages complex, multi-step workflows across distributed systems.
Capabilities:
• Handles asynchronous and multi-component workflows.
• Ensures optimal performance and system integrity.
• Software Discovery Services:
Purpose: Identifies and selects the most suitable software for solving specific
problems.
Capabilities:
• Matches software implementations to platform requirements.
• Community Authorization Servers:
Purpose: Enforces resource access policies for virtual organization participants.
Capabilities:
• Acts as policy enforcement agents.
• Controls access based on community utilization policies.
• Community Accounting and Payment Services:
Purpose: Tracks resource usage and generates payment requirements for
participants.
Capabilities:
• Provides usage metrics.
• Facilitates payment processes for resource utilization.
• The Collective Layer handles coordination and global management for virtual
organizations, enabling complex operations and workflows.
• It relies on and extends lower-layer protocols to implement robust, scalable
services critical to Grid Computing.
• Its wide-ranging services ensure resource optimization, operational integrity, and
collaborative efficiency.
 Application Layer: User-Defined Grid Applications
• These applications leverage the services and functionalities provided by the
underlying layers of the grid architecture.
• They are typically domain-specific and tailored to address specific user needs
or problems.
Key Features
• Access to Resources:
• Applications can interact with resources in two ways:
• Direct Access: Bypassing collective services and working directly with the
resource layer.
• Through Collective Services: Utilizing high-level APIs provided by the
Collective Layer for tasks like resource discovery, scheduling, and
coordination.
• APIs and SDKs for Integration:
• Each layer in the grid architecture exposes a set of Application
Programming Interfaces (APIs) and Software Development Kits (SDKs) to
facilitate application development and integration.
• Developers can choose the most appropriate level of abstraction based on their
requirements.
• Developer Flexibility:
• Developers can decide whether to use collective services for high-level
operations or directly manage resources for granular control.
• This flexibility allows for:
• Customization based on specific operational needs.
• Optimization for domain-specific applications.
• Domain-Specific Focus:
• Most grid applications are domain-specific, meaning they are designed to
solve particular problems in fields like healthcare, finance, engineering, or
scientific research.
Grid Architecture and Relationship to Other Distributed Technologies

1. Grid Computing and the World Wide Web

• Strengths of the Web for Virtual Organizations:
• Relies on open and ubiquitous technologies like TCP, HTTP, SOAP, and XML.
• These technologies make the Web a viable platform for building virtual organizations.
• Current Limitations:
• The Web primarily operates on a browser-server messaging model, which lacks:
• Single sign-on mechanisms.
• Delegation of authority.
• Advanced authentication mechanisms.
• Event correlation.
• Once these issues are addressed, the Web could evolve into a robust platform for grid portals,
enabling multiple virtual organizations.
• Continuity in Foundations:
• Core layers like platforms, fabric, and networking technologies will remain consistent,
ensuring compatibility.
2. Grid Computing and Distributed Computing Systems
• Distributed Technologies:
• Examples: CORBA, J2EE, DCOM.
• Strengths: Well-suited for distributed computing applications.
• Drawbacks for Grid Computing:
• Limited support for:
• Resource discovery across virtual organizations.
• Collaborative security and dynamic virtual organization construction.
• Scalability for resource sharing.
• Lack of interoperability among protocols.
• Promising Developments:
• Java JINI has drawn Grid Computing research attention due to its platform-
independent infrastructure for service discovery, negotiation, and leasing.
3. Application and Storage Service Providers (ASP/SSP)
• Current Model:
• Providers outsource business, scientific applications, and high-speed storage.
• Customers negotiate QoS (quality of service) and pricing.
• Typically executed over VPNs or dedicated lines, which are static and limited
in scope.
• Limitations:
• Lack of dynamic resource sharing among heterogeneous systems.
• Limited inter-domain networking interactions.
• Grid Computing Enhancements:
• Introduces flexible resource sharing across virtual organizations.
• Enhances the utility model of ASP/SSP, improving flexibility and value
proposition.
4. Grid Computing and Peer-to-Peer (P2P) Systems
• Similarities:
• Both focus on resource sharing.
• Widely utilized in home, commercial, and scientific markets.
• Differences:
• Target Communities:
• Grid Computing: Smaller, application-focused communities with higher
security and application integrity.
• P2P Systems: Larger communities with simpler security and resource-
sharing requirements.
• Resource Characteristics:
• Grid systems manage more complex, diverse, and interconnected
resources.
• Convergence Potential:
• Efforts like the Global Grid Forum (GGF) aim to merge these technologies to
integrate a broader audience.
5. Grid Computing and Clusters
• Clusters:
• Localized systems designed to pool computational resources for enhanced
power through parallel workload execution.
• Strengths: Centrally controlled nodes with dedicated functionality.
• Limitations:
• Limited to local domains.
• Unsuitable for inter-domain resource sharing.
AUTONOMIC COMPUTING
The term “autonomic” comes from an analogy to the autonomic central nervous
system in the human body, which adjusts to many situations automatically without
any external help.
• With the increasing complexity in dealing with distributed systems, solutions, and
shared resources in grid environments, we require a significant amount of
autonomic functions to manage the grid.
• As detailed in Figure 3, basic autonomic computing systems must follow the four
basic principles:
• Self-configuring (able to adapt to the changes in the system)
• Self-optimizing (able to improve performance)
• Self-healing (able to recover from mistakes)
• Self-protecting (able to anticipate and cure intrusions)
Figure 3
Examples of the grid computing efforts (IBM):
IBM's Business On Demand Initiative highlights a comprehensive approach to
transforming business practices and operational models. It encompasses the
integration of advanced technologies, process optimization, and adaptability to real-
time market conditions. Below is a detailed breakdown of the initiative's principles,
characteristics, and application in various industries.
Core Principles of Business On Demand
• Not Limited to Utility Computing:
• Goes beyond providing computational resources on-demand, focusing on
overall business transformation.
• Real-Time Responsiveness:
• Enables companies to adapt dynamically to fluctuating market demands and
customer needs.
• Business and Technology Integration:
• Emphasizes seamless collaboration between business processes and
technological infrastructure.
Essential Characteristics of On-Demand Businesses

• Responsiveness:
• Ability to sense and respond to dynamic market conditions in real-time.
• Adaptability:
• Leverages variable cost structures to optimize resource utilization.
• Core Competency Focus:
• Concentrates resources on core business areas while outsourcing or automating non-core activities.
• Resiliency:
• Ensures consistent availability and robust operations under varying conditions.
• Seamless Integration:
• Connects customers, partners, and internal systems for unified operations.
• Virtualization of Resources:
• Creates a flexible and scalable infrastructure to meet diverse demands.
• Autonomic Resources:
• Introduces self-managing and dependable systems to minimize manual intervention.
• Open Standards: Promotes interoperability across platforms and systems.
Benefits
• Acceleration of Implementation:
• Faster deployment of solutions to align with business goals.
• Enhanced Collaboration:
• Facilitates improved productivity among virtual organizations and distributed teams.
• Virtual Organizations:
• Enables dispersed departments and businesses to share data and resources efficiently.
• Robust Infrastructure:
• Supports highly flexible, resilient, and scalable operational models.
• Optimal Resource Utilization:
• Prevents overprovisioning and reduces costs by leveraging existing capital and operational
investments.
• Instantaneous Access:
• Provides immediate availability of vast computing and data resources.
Business Application Areas
IBM's initiative recognizes specific industries where Grid Computing can deliver transformative results:
1. Life Sciences
• Use Case:
• Analyze and decode complex biological and chemical data.
• Impact:
• Accelerates drug discovery, genomics research, and molecular simulations.
2. Financial Services
• Use Case:
• Run complex financial models to enhance decision-making accuracy.
• Impact:
• Optimizes risk analysis, portfolio management, and fraud detection.
3. Higher Education
• Use Case:
• Support advanced research requiring intensive data and computation.
• Impact:
• Enables breakthroughs in scientific research and collaborative learning.
4. Engineering Services
• Use Case:
• Facilitate collaborative design and testing in automotive and aerospace industries.
• Impact:
• Enhances innovation and efficiency in product development and testing.
5. Government
• Use Case:
• Seamless interdepartmental collaboration and resource sharing.
• Impact:
• Improves agility and operational efficiency in civil and military projects.
6. Collaborative Games
• Use Case:
• Replace single-server games with massively multiplayer online games.
• Impact:
• Introduces parallel, scalable gaming environments for a better user experience.
Industry-Specific Requirements
Each sector requires unique Grid Computing capabilities tailored to its operational
needs.
Examples include:
Life Sciences: High-performance computing for simulations and data analysis.
Financial Services: Secure and scalable platforms for real-time computations.
Government: Interoperable systems for multi-agency coordination.
Thank you