Cluster vs Grid vs Cloud

There are two questions:

1) What is the difference between a cluster and a grid 2) What is a cloud

I'm not looking for conceptual definitions, I found a lot of things thanks to googling, but the problem is that I still do not understand this. therefore, I believe that the answer I'm looking for is different. From what I could search again on the Internet, I begin to think that many article authors who try to explain this either do not understand it deeply enough or are unable to explain their knowledge to such an average guy like me (this is a common problem with very technical people).

Just so that you know my level: I am a .NET and LAMP programmer, I can do a basic administrator on both Linux and Windows, I have experience with Hyper-V and are now learning Xen and XCP to install a test cloud on two computers for training .

Below information you do not need to read is just my current understanding of cluster, grid and cloud. just to support my two questions, because I thought it would help to understand what a mess in my head right now and what answers I am looking for.

Thanks.


The two computers used for links in my applications are "A" and "B"

for A: 2-processor Intel processor, 8 GB memory, 500 GB disk

for "B": 2 core Intel processors, 8 GB memory, 500 GB disk,


Now I would like to see the roles of A and B from the cluster, grid and cloud.

General definitions between grid and cloud

1) a cluster or grid consists of two or more computers connected together, at the hardware level they are connected through network cards and at the software level it uses some kind of program that implements a message transfer interface so that commands can be sent between nodes.

2) a cluster or grid does NOT combine processor power or memory between nodes, which means that in this simulation, the FireFox browser running on A still has only one processor with two cores, 8 GB of memory and 500 GB.

Differences between grid and cloud:

1) The cluster provides only the fault-tolerant part if A node breaks while FireFox is running; cluster software will restart the FireFox process on node B.

2) However, the Grid can simultaneously run software on multiple nodes, provided that the software is MPI-encoded. It can also dine with any software on any node on demand (even if it is not written for MPI)

3) The grid can also combine different types of nodes, Linux Server, Windows XP, Xbox and Playstation into one grid.

Cloud Definition:

1) The cloud is not a technical term at all, it's just a short convenient word to describe a computer with unlimited resources, it can also be called a supercomputer, beast, ocean or the Universe, but someone first said β€œCloud”, and we are here.

2) The cloud can be based on grids or clusters

3) From a technical point of view, Cloud is a software for combining hardware resources into one, that if I install Cloud software on a Grid or Cluster, then it will combine A and B and I will get one cloud, like this: 4 processor cores, 16 GB of memory and 1000 GB of disk.

edited: 2013.04.02 clause 3) was completely nonsense, the cloud would not combine resources from many nodes into one huge resource, so in this case there would be 4 main processors, 16 GB of memory and 1000 GB of cloud.

+7
source share
4 answers

Grid computing is designed to distribute large workloads for many participating grid members β€” through software for each member who expects to hear this query for computation or data, and answer it with a small piece of the general puzzle. Applications must be written specifically for this problem-solving approach. It can be heterogeneous because it is not an OS, but it is important that the software expects to hear problem solving requests.

The expectation of the cluster is that it can run the same executable image through any member of the node β€” any node can execute this code, which makes it require uniformity. You can write code that supports the cluster, which distributes the workload across the cluster, but again you need to write your code to be aware of the cluster in order to take advantage of the cluster backup features. Since most application providers do not write cluster-compatible code, a simple backup function is all that is commonly used when deploying clusters, but this does not limit the architecture. Clusters can share their resources and share them at the same time.

The cloud, as it is usually defined, is neither one nor the other, but it also does not exclude them. Cloud computing involves the ability to deploy an application without additional knowledge about this operating system or even manage this operating system in combination with the ability to expand or reduce the processing and memory for this application without having to destroy and recreate this environment - everything is done with sufficient isolation that the application will not know and will not be able to find out what other applications can be installed or run on this common infrastructure, if only this one stupid is not approved by both managers applications.
+6
source

I would like to answer my question before it is closed as a duplicate, because I believe that it can be very difficult to find the right information about clusters, meshes and clouds, and I think this post can save time for many. If someone wants to challenge, please do it, otherwise I will mark it as an answer in 1 week.

1) There are many differences, and there are none, it really depends on the technical context, but usually you can connect several nodes and call it a grid, or you can call it a cluster. I would say that Grid is a cluster with advanced features, for example, the ability to connect heterogeneous nodes. Both Grid and Cluster will serve as a scalable platform equally well. From the point of view of a network engineer and a programmer, the difference in implementation or coding will be quite large if Gird connects heterogeneous nodes.

2) Now the first question was actually a prelude to the second, and I believe that it is best to answer Matt Joyce in this post: https://stackoverflow.com/a/314618/

0
source

I will hack it. I have been collecting and saving my notes, scripts and programs since 2002 AD This has been shredding and pasting my applications over the years. Here is a list of memorable mental abilities:

  • Grid are hardware and hardware specifications.

    but. You connect to the router or switch and configure IP addresses and top-level domains over the Internet (which is also known as ICANN).

    b. This is similar to OSI layer 1, 2, and 3.

  • A cluster is the core (software ring 0 or 1, if it occurs in a virtual type).

    but. The kernel is configured (compiled) to start the network stack, which can handle sessions, permission and account.

    b. You set up a messaging port, usually over TCP / IP (for example, in the OSI model).

    from. You install iptables , pf , arp and other OS level applications or shared objects.

    e. You can configure ssh, kerberos, ldap, or some other PKI database and protocol socket.

    e. This is similar to OSI layer 4, 5, and 6.

  • The cloud is user space applications.

    but. Application processes interact with other application processes within the cluster.

    b. You set process level permissions (through files, groups, and / or user groups).

    from. You install mysql, redis, riak, Message Brokers, hadoop, apache, nginx, cron, java, haskell, erlang, etc.

    e. This is similar to OSI layer 7.

A cloud floats over a cluster that grows out of a grid. And actually visually think of clouds in the air, a cluster in a tree, and a grid on the ground. Most of us creative types (which all these technologies make) are visual thinkers who can support it with mathematical data and code. Therefore, always see if you can answer the riddle and compare technological facsimiles with our physical world here on Earth.

Introduction

A grid, cluster, and cloud are three different words that mark their specific time in history. Their definitions have overlapping features, and they are interchangeable. You just need to know when to apply the correct or related word. For example, I talked with some senior doctors (doctors), and they wanted to know what a cloud is. So I told them that the cloud is a computer cluster that you rent through the Internet. And Bingo, they got the idea in 10 seconds.

I will use a little history in chronological prose.

Grid

The term β€œmesh” is first used to represent a single resource that repeats in a terrestrial landscape or space. This term is often used when distributing telegraphs, where repeaters had to be placed on the poles for each N radii (multiple radius) to amplify the signal. Another example is the electrical grid, which Thomas Edison and Nikola Tesla have begun to compete on the Earth. Computers became very popular, and soon they were expanded through a grid to replace the operators of telegraph (telephone) telephones. The grid is now a bunch of computers that can connect and interrupt communication channels. Grid is an infrastructure of computers that work for one purpose, which is assembly code (or binary). [/ P>

Cluster

Overcoming the power of computers and actually observing computers, wars wars (Turing machine), DARPA (or ARPA, which is the US Armed States). DARPA has launched the introduction of universities and colleges to use the Grid to multiplex communication methods (which use baud and protocols). Universities and colleges began to create protocols to separate the different tasks that they wanted to perform on a grid and target computers. This is the beginning of the modern internet. The laboratories have created their own test clusters to simulate the mesh. Clusters are great for orchestrating. Work can be divided into all or some of the followers in the cluster. The military used college and university results and applied the SOFTWARE to the grid. There were some gotchas with clusters:

  • It should be the same (or almost the same) equipment
  • Must have the same operating system

The rules were strict because all instruction sets had to go through processors equally. Typically, clusters have a relationship between a master and a sub type. The cluster typically ran one unic (or unix) job at a time. Clusters had task schedulers. Clusters then became more complex because hardware manufacturers began to create parallel chip architectures (on top of the von Neumann arc). Clusters are becoming more powerful. Clusters inherited greater complexity, and people did more creative things. Now the cluster can perform various tasks, tasks, processes, asynchronous processes, synchronized processes and much more. One box (or computer node) can run more tasks. Grid can now be used for several purposes. Cluster software update rates were faster than the actual grid. Clusters were deployed locally on campuses. Clusters began replacing the grid because you could directly create a public stack that did not execute the (national) grid.

My experience

I went to college in the late 1990s and 2000s, and the cluster was the word for the physical laboratory of several computers working as one virtual computer. Clusters were used for testing. After your software has been running in a cluster, you can mv (move) it to the production grid class. Then I see that network worms and computer viruses control zombie computers. These swarms of zombies can be used as one giant virtual cluster used to run commands. Well, programmers started to do DIY (do it themselves) protocols and software such as bit-torrent and Napster. Therefore, moving on to the future, cluster software testing begins to be replaced by Solaris prisons, FreeBSD prisons, Linux containers, QEMU, hypervisors, VMWare, VirtualBox, Vagrant and Docker.

Cloud

Cloud is a marketing term used to probe the hardware of various networks and the software of these clusters. The cloud is one big ubiquitous word used to advertise, promote and use all of this cluster technology to make money profit. The cloud is also an attempt to wrap all of these technologies under one single word. The cloud allows multi-story processes to share a giant grid. Cloud maximizes efficiency through the separation of electricity, CPU, RAM, DISK, electricity and broadband, which are distributed and paid by consumers. A side effect is that these consumer subscriptions and / or tariffs have begun to make a profit. Cloud also allows multiple users to install multiple operating systems that perform many processes in software. So now we have abbreviations like IaaS, PaaS and SasS. The cloud can replace the initial cost, which was once so difficult to finance and download. Cloud is a great solution for testing your software and creating a consumer base for your business.

On the other hand, Cloud makes the brain of non-programmers think in a certain way. For example, the human resources department can understand and isolate what is presented to them.

So, if you have money, you can gain your share of the cloud experience and have easy support along with it. But if you have a set of skills, time, fast know-how and the ability to install your own servers in the field, then do it because it is cheaper in the long run.

This is my story about Grid vs Cluster vs Cloud.

0
source

I think this link compared the cluster and the grid well.

As I know, in the case of clusters there are some exceptions. YARN (Yahoo!) is trying to handle mutli-renting and distributed planning. Also Corona (Facebook) has distributed planning.

-one
source

All Articles