Dwarf尺寸的进一步缩减

Further Compression of Dwarf Size

Dwarf不仅降低了数据立方的存储开销,而且具有结构简单、易于实现、查询和维护等优点,是一种比较理想的数据立方组织方法。为了进一步缩减Dwarf的存储尺寸,本文通过研究Dwarf结构,分别提出了浓缩Dwarf和冰山Dwarf:前者从Dwarf结构中删除了对于查询来说冗余的内容,而后者从Dwarf结构中去掉了对于用户来说琐碎的内容。实验和分析表明,浓缩Dwarf有效地减小了Dwarf的存储尺寸,而冰山Dwarf适合于忽略细节的应用场合,极大地降低了Dwarf的存储开销。

Dwarf is an appropriate way for data cube store because it not only reduces the storage size, but also has a simple structure and is easy to be queried and maintained. For further compression of Dwarf, we proposes Condensed Dwarf and Iceberg Dwarf respectively, the former deletes from Dwarf structure redundant store, while the latter deletes from Dwarf structure trivial store. Our experiments and analysis show that Condensed Dwarf reduces the storage size of Dwarf effectively, while Iceberg Dwarf works well in detail-overlooked situation, and it can reduce the storage size of Dwarf significantly in such cases.