Conventional approaches to data warehousing have significant
problems with a number of types of query. For example, when attempting
to answer complex analytical queries it typically takes either a very
long time (certainly measured in hours and often in days) or requires an
inordinate amount of processing power. This is especially true when the
query cannot be predicted in advance. Secondly, traditional relational
databases simply cannot cope with truly real-time environments. And,
thirdly, standard approaches are not well-engineered to cope with
large-scale queries whose business rules change on a regular basis.
Today, a number of companies are addressing these concerns and offering
products that can be used to supplement existing solutions (though some of them may also replace traditional approaches). Typically, these
vendors report performance improvements of hundreds of times, frequently on considerably reduced hardware platforms and this report examines these claims, which might be thought to strain credulity (actually they are real).
Specifically, this report compares six of the vendors in this space, including Aleri, Alterian, Apama, Aruna, Sand Technology and Sybase. The
report also investigates the technologies (including column-based relational databases, tokenisation, and vector processing and databases) that underpin their various solutions.