Today I explain to you some of my understanding of a cluster technology, if wrong or bad can be a lot of talk made, I will make the appropriate changes, first thank you for the comments in advance! ! !
Cluster (Cluster) technology: this can be obtained by using lower cost high performance, cost reduction, increased reliability and scalability. The task scheduler is the core technology Cluster systems ---- (task scheduling) load balancing
Cluster is composed of a set of network computers, and managed in a single system. Cluster can be compared to a separate server, a network set of computers can imagine many computers together to meet the computing processing capability (which may be more or may be dozens)
Cluster talk about improving performance, reduce costs, increase scalability, reliability has it
Improve performance: like the now popular big data, it uses the conventional tools of data collection, processing, and mining profits key to further implement the product. But the operation on time, you need a strong processing and computing capacity of the existing technology and large computer is very difficult to do it well.
So Cluster technology can be integrated with dozens or even more computers to meet and improve the processing and computing power, which is one of the important research goals Cluster technology.
Note: large data mentioned technique uses a distributed framework. Feature is the mining of vast amounts of data, but it relies on cloud computing distributed processing, distributed databases and distributed storage and virtualization technologies. I am here just for example using a hot term to give you a better understanding and share some of our knowledge and understanding
Cluster improve the performance of reading, a lot of that before they think and are willing to dig deeper curious baby who will ask me, then distributed cluster are multi-processing computer is running what some of the differences? After then I understand, I will be saying what I understand the difference between them before. There are curious to see the babies, enjoy looking oh
He said improved performance, then talk about reducing the cost.
Reduced costs: Generally, we configured a good Cluster hardware and software it costs around $ 100,000, see here just out of society babies are not feel shocked, I began is the same, but when you know when a supercomputer and its performance year on year terms, its cost is much lower, a supercomputer generally around $ 1 million.
To talk about scalability:
Scalability: how to reduce the cost of increased scalability, curious babies like to spend the same price to buy a notebook, A curious baby can be for expansion of the memory location after notebook, curious baby B notebook of its memory when he bought fixed position, and that our notebook configuration to keep up with the speed of our time, curious baby a memory can be added to increase the operating speed of the computer, but our curious position because the memory of the baby B filled, so if he were to run faster on the need for a computer to upgrade the computer to buy a new laptop.
After using Cluster technology we only need to build a Cluster, then we need to be extended, as long as the new node is added to the computer to go inside to the Cluster. (It is by a few corresponding load balancing cluster to handle the load balancing cluster - is one of the three classification clustering technology)
Finally, talk about reliability
Reliability: time to say reliability, we will be a little story.
There are curious baby A, B, C one day they also write a blog post, but suddenly blog system malfunctioned, curious baby A, curious baby B written blog post can be uploaded successfully, and curious blog post baby C's but can not upload, normal computer, if our system fails, all operations all canceled, but the cluster because it is a cluster, it can happen to continue to work after a fault in the system, it will greatly reduce the loss of failure, which is the cluster the benefits of
Curious babies, pay attention to the attention of the attention, it is important to say three times, and finally talk about the technical sub-Cluster hardware and software load tool, load software tools commonly used Apache and nginx, and nginx is now used more tool.