By Bob Scheier
Informix Dynamic Server
compresses, consolidates
data to boost data
warehousing
t
b
Mention Informix Dynamic Server (IDS) and most DBAs think of a fast, reliable, and low-maintenance platform for online transaction processing (OLTP). But those transactions hold valuable insights into business trends, leading many
organizations to also use IDS as a data warehouse against which they can run
business intelligence (BI) queries.
To meet those needs, Informix has unveiled a series of enhancements that
ring its long-standing speed and ease-of-use benefits to data warehousing and
BI. These include tools to help customers model, schedule, and execute the data
transformation and data flows required to create data warehouses. Most recently,
IBM announced new data compression and consolidation features that reduce the
cost (and boost the performance) of both BI and OLTP applications on IDS.
The key new technology is the IDS Storage Optimization Feature. It reduces
he size of not only OLTP data stores but also data warehouses that, if they grow
too large, can make business analysis overly complicated and expensive.
a
e
Introducing Storage Optimization
Based on technology used in IBM DB2, Storage Optimization compresses and consolidates the data within IDS. Early customer trials sho w that Storage Optimization
reduces the amount of space required to store data either in memory or on disk by
an average of 50 percent. This can cut the time required to process queries by as
much as 20 percent because more data can be kept in memory, reducing the number of I/O operations to slower mechanical disk drives. Cutting the database size
can also, of course, delay or even eliminate the need to upgrade disk storage.
The first of the three Storage Optimization components is compression, which
xamines each row in the database for recurring patterns of data. It stores the
individual recurring patterns in a dictionary, replacing those patterns with shorter
strings of symbols. Unlike other compression techniques that scan only a portion
of each row for repeating patterns, Storage Optimization scans the entire row,
regardless of how many columns it intersects. By scanning a larger area than other
compression techniques, it can find and compress more repeating patterns, and
thus achieve very high compression ratios (see Figure 1).
Repack (or Coalesce), the second component, consolidates the free space creted within each partition, while the final capability, Shrink, removes the unused
portion of the partition and returns it for reuse by IDS. These larger, contiguous
spaces are much easier for IDS to reuse than smaller, isolated free spaces. It is
this compression and consolidation that speeds query performance, while holding
down the amount of physical disk space required for data warehousing.
