As
it stores records, it compresses them so that less disk storage is required. How
much it compresses them depends on many factors, but a general rule of thumb is
that a cube will normally take up about 30??“40 percent of the size of the relational
warehouse. Therefore, a 1TB relational data warehouse would translate to a cube of
approximately 300??“400GB.
48 B u s i n e s s I n t e l l i g e n c e w i t h M i c r o s o f t O f f i c e P e r f o r m a n c e P o i n t S e r v e r 2 0 0 7
The reason a cube is stored in a binary format, at least with Analysis Services
2005, is so the data can be retrieved extremely quickly. The binary format is
designed for extremely fast retrieval of data. Not only is the format designed to
be read quickly and return data as fast as possible, the compression means that
reads need to access less disk, so that actual reading off the drive is faster than for
uncompressed data.
Part of what the cube can do as well is preaggregate the data. Cubes can store
higher level summarizations of data, meaning that requests for summarized data can
be returned very quickly. A cube that stores sales data at a daily level may contain
aggregated data that stores sales at the month, quarter, and year level, for example.
Being able to retrieve these aggregations means that the engine does not have to
spend time calculating the values on the fly.
This means that cubes have two major advantages over storing the same information
in a relational database: first, they are built for extremely fast retrieval of data and
second, the inclusion of preaggregated data means that queries for higher-level values
is extremely fast.
Pages:
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92