What are the difference between cluster computing. grid computing and cloud computing.
|
Cluster Computing |
Grid Computing |
Cloud Computing |
|
Characteristics of Cluster computing 1:Tightly coupled systems 2: Single system image 3: Centralized Job management & scheduling system |
Characteristics of Grid Computing 1: Loosely coupled (Decentralization) 2: Diversity and Dynamism 3: Distributed Job Management & Scheduling |
Characteristic of cloud computing 1: Dynamic computing infrastructure 2: IT service-centric approach 3: Self-service-based usage model 4: Minimally or self-managed platform 5: Consumption-based billing |
|
In cluster computing, a bunch of similar (or identical) computers are hooked up locally (in the same physical location, directly connected with very high speed connections) to operate as a single computer |
In grid computing, the computers do not have to be in the same physical location and can be operated independently. As far as other computers are concerned each computer on the grid is a distinct computer. |
In cloud computing, the computers need not to be in the same physical location. |
|
The cluster computers all have the same hardware and OS. |
The computers that are part of a grid can run different operating systems and have different hardware |
The memory, storage device and network communication are managed by the operating system of the basic physical cloud units. Open source software such as LINUX can support the basic physical unit management and virtualization computing. |
|
The whole system (all nodes) |
Every node is autonomous i.e. it |
Every node acts as an independent |
|
behaves like a single system view and resources are managed by centralized resource manager. |
has its own resource manager and behaves like an independent entity |
Entity |
|
The computers in the cluster are normally contained in a single location or complex. |
Grid are inherently distributed by its nature over a LAN, metropolitan or WAN |
Clouds are mainly distributed over MAN |
|
More than 2 computers are connected to solve a problem |
A large project is divided among multiple computers to make use of their resources. |
It does just the opposite. It allows multiple smaller applications to run at the same time. |
|
Areas of cluster computing 1. Educational resources 2.Commercial sectors for industrial promotion 3.Medical research |
Areas of Grid Computing 1.Predictive Modeling and Simulations 2.Engineering Design and Automation 3.Energy Resources Exploration 4.Medical, Military and Basic Research 5.Visualization |
Areas of cloud Computing 1.Banking 2. Insurance 3. Weather Forecasting 4.Space Exploration 5.Software as a service 6.PaaS 7.Infrastructure- as -a-Service |
|
Commodity computers |
High-end computers (servers, clusters) |
Commodity computers and high-end servers and network attached storage |
|
Size or scalability is 100s |
Size or scalability is 1000s |
Size or scalability is 100s to 1000s |
|
One of the standard OSs (Linux, Windows) |
Any standard OS (dominated by Unix) |
A hypervisor (VM) on which multiple OSs run |
|
Single Ownership |
Multiple Ownership |
Single Ownership |
|
Dedicated, high-end with low latency and high bandwidth Interconnection Network |
Mostly Internet with high latency and low Bandwidth Interconnection Network |
Dedicated, high-end with low latency and high Bandwidth Interconnection Network |
|
Traditional login/password- based. Medium level of privacy depends on user privileges. |
Public/private key pair based authentication and mapping a user to an account. Limited support for privacy. |
Each user/application is provided with a virtual machine. High security/privacy is guaranteed. Support for setting per-file access control list (ACL). |
|
Membership services discovery |
Centralized indexing and decentralized info services discovery |
Membership services discovery |
|
Limited service negotiation |
Yes, SLA based service negotiation |
SLA based service negotiation |
|
User management is centralized |
User management is decentralized and also virtual organization (VO)-based |
User management is centralized or can be delegated to third party |
|
Resource management is centralized |
Resource management is distributed |
Resource management is centralized/distributed |
|
Virtual Interface Architecture (VIA)-based standards |
Some Open Grid Forum standards |
Web Services (SOAP and REST) standards |
|
Single system image |
No single system image |
Yes, but optionally include Single system image |
|
Stable and guarantee capacity |
Varies, but high capacity |
Provisioned on demand capacity |
|
Failure management (Self- healing) is limited (often failed tasks/applications are restarted). |
Failure management (Self- healing) is limited (often failed tasks/applications are restarted). |
Strong support for failover and content replication. VMs can be easily migrated from one node to other. |
|
Limited pricing of services but not open market |
Pricing of services is dominated by public good or privately assigned |
Utility pricing, discounted for larger customers |
|
Multi-clustering within an Organization for internetworking |
Limited adoption for internetworking, but being explored through research efforts such as Gridbus InterGrid |
High potential, third party solution providers can loosely tie together services of different Clouds for internetworking. |
|
Potential for building 3rd party or value-added solutions is limited due to rigid architecture |
Potential for building 3rd party or value-added solutions is limited due to strong orientation for scientific Computing |
High potential - can create new services by dynamically provisioning of compute, storage, and application services and offer as their own isolated or composite Cloud services to users |
1. Cluster Computing:
Cluster computing it’s a group of computers connected to each other and work together as a single computer. These computers are often linked through a LAN. The clusters came to existence for the high need for them, because the computing requirements are increasing in a high rate and there’s more data to process, so the cluster has been used widely to improve performance.The cluster is a tightly coupled systems, and from its characteristics that it’s a centralized job management and scheduling system. All the computers in the cluster use the same hardware and operating system, and the computers are the same physical location and connected with a very high speed connection to perform as a single computer. The resources of the cluster are managed by centralized resource manager. the cluster is single owned, to only one organization. Its interconnection network is a high-end with low latency and high bandwidth, the security in the cluster is a login/password-based, and it has a medium level of privacy depends on users privileges. it has a stable and guaranteed capacity. The self healing in the cluster is Limited, it’s often restarts the failed tasks and applications. its service negotiations are limited, and the user management is centralized. The cluster computing is usually used in educational resources, commercial sectors for industrial promotion & Medical research.
Architecture:
The architecture of cluster computing contains some main components and they are:
- Multiple stand alone computers.
- Operating system.
- High performance interconnects
- Communication software
- Different application platforms
Advantages:
- In the cluster software is automatically installed and configured, and the nodes of the cluster can be added and managed easily, so it’s very easy to deploy, it’s an open system, and very cost effective to acquire and manage, clusters have many sources of support and supply, it’s fast and very flexible, the system is optimized for performance as well as simplicity and it can change software configurations at any time, also it saves the time of searching the net for latest drivers, The cluster system is very supportive as it includes software updates.
Disadvantages:
- Cluster computing contains some disadvantages such as that it’s hard to be managed without experience, also when the size of cluster is large, it’ll be difficult to find out something has failed, the programming environment is hard to be improved when software on some node is different from the other.
2. Grid Computing:
Grid computing is a combination of resources from multiple administrative domains to reach a common target, and this group of computers can distributed on several location and each a group of grids can be connected to each other. The need of access to additional resources and the collaborating between organizations leads to the need for grid computing. Grid environments are extremely well suited to run jobs that can be split into smaller chunks and run concurrently on many nodes. The grid is a loosely coupled systems and from its characteristics that it’s a distributed Job Management and scheduling, the computers in the grid are not required to be in the same physical location and can be operated independently, so each computer on the grid is concerned a distinct computer, the computers in the grid are not tied to only on operating system and can run different OSs and different hardware, when it comes to a large project, the grid divides it to multiple computers to easily use their resources. The grid is multiple owned, it could be owned by several companies. Interconnection network is mostly internet with high latency and low bandwidth. The security in the grid is public and private based on authentication and mapping user to an account. And it has limited support privacy, its capacity is not stable, it varies, but it’s high. The self healing in the cluster is Limited, it often restarts the failed tasks and applications. its service negotiations are based on service level agreements, and the user management is decentralized. The grid computing is usually used in predictive modeling and simulations, engineering design and automation, energy resources exploration, medical, military, basic Research and visualization.
Architecture:
- Fabric layer to provide the resources which shared access is mediated by grid computing.
- Connectivity layer and it means the core communication and authentication protocols required for grid specific network functions.
- Resource layer and it defines the protocols, APIs and SDK for secure negotiations, imitations, monitoring control, accounting and payment of sharing operations on individual resources.
- Collective layer which it contains protocols and services that capture interactions among a collection of resources.
- Finally the Application layer, and it’s user applications that operate within VO environment.
Advantages:
One of the advantages of grid computing that you don’t need to buy large servers for applications that can be split up and farmed out to smaller commodity type servers, secondly it’s more efficient in use of resources. Also the grid environments are much more modular and don't have much points of failure. About policies in the grid it can be managed by the grid software, beside that upgrading can be done without scheduling downtime, and jobs can be executed in parallel speeding performance.
Disadvantages:
It needs fast interconnect between computers resources, and some applications may need to be pushed to take full advantage of the new model, and licensing across many servers may make it forbidden for some applications, and the grid environments include many smaller servers across various administrative domains. also political challenges associated with sharing resources especially across different admin domains.
3. Cloud Computing:
Cloud computing is a term used when we are not talking about local devices which it does all the hard work when you run an application, but the term used when we’re talking about all the devices that run remotely on a network owned by another company which it would provide all the possible services from e-mail to complex data analysis programs. This method will decrease the users’ demands for software and super hardware. The only thing the user will need is running the cloud computing system software on any device that can access to the Internet . Cloud computing is useful for the small business companies to make their resources from external sources as well as for medium companies, the large companies have obtained the largest storage without the need to build internal storage centers, thus, the cloud computing has given for both small and large companies the ability to reduce the cost clearly. In return for these services, the providing companies for cloud computing requires a financial gain determined by use. The cloud is a dynamic computing infrastructure, IT service-centric approach; also it’s a self-service based usage model and self-managed platform and its consumption based on billing, the computers in cloud computing are not required to be in the same physical place, wherever you are you will be served. The operating system of the basic physical cloud units manages the memory, the storage device and network communication. In the cloud you can use multiple operating systems at the same time. Every node in the cloud is an independent entity. It allows multiple smaller applications to run at the same time. The cloud is owned by only one company and it provides its services to the users, it interconnection network is a high-end with low latency and high bandwidth. The security in the cloud is high and the privacy is guaranteed, each user/application is provided with a virtual machine. Its capacity is provided on demand. The self healing in the cloud has a strong support for failover and content replication, and virtual machines can be easily migrated from one node to other. its service negotiations are based on service level agreements, and the user management is centralized or can be delegated to third party. The cloud computing is usually used in banking, insurance, weather forecasting, space exploration, software as a service, platform as a service, infrastructure as a service.
