Chitose King
Links: https://www.zhihu.com/question/27974418/answer/39845635
Source: Know
Copyright belongs to the author, please contact the author for authorization.
Google has begun to play big data, found that the times can't keep up with their rhythm, worried about the technology successor, so published three papers (Search GFs bigtable mapreduce). There are a few work unsaturated, all the people who have nothing to do, want to engage in an open source web search (Lucene nutch). The three papers were shocked and began to practice in a second-rate internet company (Yahoo). That's what Google wants. Daoteng a few, out of an elephant (Hadoop), this is just a code. Big data, not just storing huge amounts of data, but emphasizing the value of good data, is analysis and computation. Like a huge atomic bomb research and development team, Einstein only one, to crush Einstein into a madman's appearance is only a drop in the bucket, but can be put into the capacity of a poor, the universities, research institutions of mass production, also have a certain ability of the slag (i), come together, human sea tactics proved to be feasible, Because the CPU is not a lot of diodes (2 goods) composed of. Each slag should be able to memorize some information and process some information. This is the distributed storage and computing (HDFs mapreduce), the upper layer by the Einstein and the like to unify the control. Well, start running, and Roosevelt asked Einstein if the dregs were reliable. Einstein replied that the system was supposed to be unreliable, they every day DotA, bubble sister, but the system is good enough fault tolerance, one can not change another, a too slow on two run together, who quickly with WHO, the internal credit mechanism and blacklist it. Roosevelt said, I see the line.
How does "Hadoop" describe the big data ecosystem?
