Massive Data Following Database Research

In the information era, there are more and more information. But row-oriented database found it more and more difficult to query and analyze these information. It needs other ways to solve this problem.Column-oriented database is different from traditional row-oriented database. It stored it's data by columns. What will be taken from the small change?In this paper, in comparing with row-oriented storage, we research the principles of column-oriented storage, and introduce its key technology. Basing on these researches, we found that column-oriented database have two significant advantages:one is good compressing; the other is good at querying and analyzing vast data. We study on these two advantages. By summarizing the existing compress algorithms which are suitable to column-oriented database, and according to the character of security data, we made a type-adaptive compression algorithm that is special for column-oriented database with vast data. It's proved to have good compress effect. And in the research of query of vast data, by building a simulated security trading analysis system, we study the practical performance of column-oriented database in the real environment. And take MonetDB and MySQL for compare. Then verify the advantage of column-oriented database in querying and analyzing fields by experimental data.