Distributed Computing Concepts

Study Notes

Resource sharing enables users to utilize the computing power and storage of a system collectively.
Load sharing distributes various tasks across nodes, optimizing overall system performance.
Easy node expansion is anticipated to be quick and efficient with minimal time investment.
Parallel computing is a subset within the broader category of distributed computing.

Computing Intensive: Tasks that require extensive computation time, such as calculating Pi using Monte Carlo methods.
Data Intensive: Involves handling large volumes of data, exemplified by platforms like Facebook and LHC data processing.
Robustness: Absence of a Single Point Of Failure (SPOF); if one node fails, other nodes can continue the same task without interruption.

Comprise processes dispersed across a network, working collaboratively to solve common issues.
Historically centered around "client-server" models, which centralized resource management.
The shift towards "Peer to Peer" computing promotes genuinely distributed applications.

A networking form that leverages unused processing power across computers for complex problem-solving.
Virtualizes computing and data resources, providing a collective single-system image.
Essential for scientific research, enabling computer simulations, data analysis, and enhanced data visualization.
Facilitates efficient use of underutilized resources.

Computational Grid: Provides shared processing power for high-throughput applications.
Data Grid: Supports storage, discovery, handling, and manipulation of large datasets from heterogeneous sources.
Collaboration Grid: Enhances collaborative projects across organizations without revealing proprietary information.
Network Grid: Offers high-performance, fault-tolerant communication services.
Utility Grid: Shares not only data and computation but also software and resources, allowing applications to run on singular machines while serving multiple users.

Consists of interconnected standalone computers working collaboratively as a single resource.
Improves speed and reliability compared to a single computer, typically being more cost-effective.
Employs high-performance interconnects and low-latency protocols to enhance communication.

High Availability/Failover Clusters: Ensure continuous system availability through redundancy.
Load Balancing Clusters: Distribute workloads evenly across multiple computers.
Parallel/Distributed Processing Clusters: Enhance processing capabilities by dividing tasks among several nodes.

High system availability due to hardware redundancy.
Supports varied business models such as pay-for-use and optimal resource utilization.
Enables data center virtualization and provisioning.
Streamlines web services delivery and automation processes.

On-demand cyber infrastructure providing flexible services to consumers and businesses.
Pricing models mimic traditional utility services, including flat rates, tiered pricing, and pay-as-you-go systems.
Risks include data backup, security, partner competency, and defining Service Level Agreements (SLAs).

Defined by the US National Institute of Standards and Technology as a model for ubiquitous network access to shared configurable computing resources.
Focuses on rapid provisioning with minimal management effort or service interactions, emphasizing convenience and scalability.