The good news is that there are a number of processing
techniques. One is a full process, in which all existing data is removed from the cube
and it is completely reloaded. Another only adds new records to the cube. There are
others as well, and they exist to cover different needs in different situations. There is
also a technology in Analysis Services 2005 called Proactive Caching, which allows
a cube to update an in-memory cache instead of having to physically reprocess the
cube. This feature is designed to allow cubes to be more real-time than those that are
processed in the traditional, scheduled fashion.
There are also several different storage formats, such as leaving all data in the
relational engine. The specifics of these are beyond the scope of the book, but the
overall process is basically the same as described here. Understand that in almost
all cases, storing data in the binary Analysis Services format, called MOLAP
for Multidimensional OLAP, is the fastest way to retrieve data for reporting and
analysis.
Delivering Data to Consumers
After the cube is built, the final step is to deliver data to the consumers. This is
where, oddly enough, many companies fall down. They build a warehouse and then
wonder why no one uses it. Most companies make one, or both, of the following two
mistakes:
They provide inadequate or suboptimal tools for working with the data in a
cube.
They overcomplicate the cube so much that few people can use it.
Pages:
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106