Shulou(Shulou.com)05/31 Report--

This article mainly explains "what are the knowledge points of server cluster technology". The content of the explanation in this article is simple and clear, and it is easy to learn and understand. Please follow the editor's train of thought. Let's study and learn what are the knowledge points of server cluster technology.

Load balancing technology is used to solve the problem of how to distribute many unrelated small or medium-sized tasks to different servers. Unrelated small or medium tasks are tasks that are not related to each other and can be accomplished with only one server. The tasks of the vast majority of individual tenants fall into this category.

For large tasks, because one server can not complete them on time, it is necessary to split the large tasks into many small and medium-sized tasks, and then assign them to multiple servers to complete them together. This is the problem to be solved by computer cluster technology.

For tenants, a cluster system consisting of many servers is like a supercomputer, no matter how complex and large-scale tasks are run, you can get the results immediately. However, supercomputers with the same computing power are extremely expensive, so when the price of PC servers drops sharply, people are keen to use cheap cluster systems to complete a variety of high-performance computing tasks. A more obvious example is that many large companies use tens of thousands of servers to build search cluster systems, and the servers are assembled by themselves.

Next, we will explain that the cluster involves two major technologies: one is the splitting of tasks, and the other is the scheduling of tasks.

1. Split of tasks

One of the principles of task splitting is to reduce the correlation between subtasks as much as possible, so as to improve the parallelism of processing tasks. The most common correlation is the time sequence of task processing, for example, subtask A can not be processed until subtask B is completed, that is, the output of task B is used as the input of task A. There is a class of large tasks with unrelated subtasks, and it is easy to split such tasks.

For example, to search for the keyword "cloud computing" in hundreds of millions of web pages around the world, one of the ways to split this task is to divide this task by country: a subtask searches Chinese websites, B subtasks search American websites, C subtasks search British websites, and so on. these subtasks are not related and can be processed in parallel, and finally the search results of the whole subtasks can be combined. Another example is China's weather forecast, which can be split by province.

two。 Task scheduling

Task scheduling is how to reasonably schedule the split subtasks to the cloud server and let them work together. The "reasonable" here is closely related to the requirements of the tenants. It can be defined as follows: the scheduling method that can best meet the requirements of tenants is reasonable.

Tenants have a variety of requirements, some want the task to be completed as soon as possible, some want to complete the task before a certain time in the future, some want the results to be as accurate as possible, and some want to calculate the cost as little as possible. According to the different requirements of tenants, people put forward the corresponding task scheduling methods. For example, on the premise of meeting the tenant's expected goals, assign tasks to as few computers as possible, so as to meet the requirements of "the least used computers to complete tasks on time and quality", because the fewer computers involved, the lower the cost of communication.

For example, there is a 100-person-day task that Party A requires to be completed within 10 days. In order to complete the task within the specified time, Party B is more willing to arrange 10 people to do it for 10 days rather than 105 people to do it for one day. Since 100 people are to be managed, the latter needs five additional managers.

Cluster is a complex project, which involves many basic distributed algorithms, such as election algorithm, consistent algorithm, wave algorithm, snapshot, fault detection and so on.

Thank you for your reading. the above is the content of "what are the knowledge points of server cluster technology". After the study of this article, I believe you have a deeper understanding of the knowledge points of server cluster technology. the specific use of the situation also needs to be verified by practice. Here is, the editor will push for you more related knowledge points of the article, welcome to follow!

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.