Shulou(Shulou.com)06/03 Report--

MapReduce can roughly see the reason from its name. The two verbs Map and Reduce, "Map" is to decompose a task into multiple tasks, and "Reduce" is to aggregate the results of multitasking after decomposition to get the final analysis results. This is not a new idea, in fact, its essence is a kind of "divide and conquer" idea, which divides a huge task into many small task units, and finally summarizes the results of each small task unit, and obtains the final result. In a distributed system, a machine cluster can be regarded as a hardware resource pool, splitting parallel tasks and then handing them over to each idle machine resource, which can greatly improve computing efficiency and this kind of resource independence. It undoubtedly provides the best design guarantee for the expansion of computing clusters. After the task is broken down, the results of the processing need to be summarized again, which is what Reduce does.

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.