توضیحات
The roots of column-oriented database systems can be seen beginning from 1970s, but it was not until 2000s that some researches and applications started to be done. In the last recent years some column store databases namely MonetDB
[3], [4] and C-Store [5] has been introduced by their authors, with the claim that their performance gains are quite noticeable against traditional approaches. These traditional approaches are for row-oriented database systems that have physical designs such that almost all tables in the database have one-to-one mappings to the tables in the logical schema. One of the important points that it will be looked in this paper is to figure out if the gains of column-stores are due to their own
internal physical design or if the same can be achieved using some column-oriented design in a row-stores database.MonetDB [3], [4] is a column-oriented open source database management system for high-performance
applications in data mining, business intelligence, scientific databases, XML query and text and multimedia retrieval. It focuses on exploiting column-oriented data processing as a way to improve the computational efficiency of database engines. This system will be talked about more in Section II and VI
To understand better the internal structure of column-stores,it can be observed that the attempt of simulating column-stores
inside row-stores [2]. Star Schema Benchmark (SSBM) [6],
[7] is a recently proposed data warehousing benchmark that has been implemented with column-oriented internal design as possible, in addition to some traditional designs. Columnoriented approach techniques which are used in SSBM
(vertical partitioning, index-only plans and materialization) will be explained in Section III. This system is taken as an
example and examined to address our questions. After observing performance of the open source C-store
database [5], which is a column-store, on SSBM; it will be shown that although the internal structure of a column store is emulated inside a row store, the query processing performance is quite poor. This observation raises the question of which kind of optimization techniques can be used to improve the performance of column-stores over row-stores on warehouse workloads. Four optimization techniques are defined which are specific to column-oriented database management systems(DBMS)as late materialization block iteration [8], compression [9] and invisible joins (a new technique); and
they will be discussed in Section IV.
year : 2012
by :S.G.Yaman
Information File :English langauge/6 page/size:332k
Download Link :click
سال :2012
کاری از :S.G.Yaman
اطلاعات فایل :زبان انگلیسی /6صفحه /حجم :332k
لینک دانلود :روی همین لینک کلیک کنید
نقد و بررسیها
هنوز بررسیای ثبت نشده است.