An IQ database is fundamentally different from a conventional relational database because it focuses on readers, not writers. In a conventional database, it is most important to allow many users to update the database instantly and accurately, without interfering with each other. By contrast, in an IQ database, fast query response for many users is most important.
This reader-oriented approach drives the design of IQ architecture and features. The IQ way means:
Data is stored in columns, not rows.
Placing indexes on all columns gives you a performance advantage.
A large page size gives you a performance advantage.
A large temporary cache gives you a performance advantage for most operations.
Access to data is controlled at the table level, using table-level locks rather than row locks, and a technique called table-level versioning.
Most query results focus on data at the table level.
Most insertions and deletions write data for an entire table, not for a single row.
To get the best performance from your IQ database, keep these fundamental differences in mind. For details on setting up IQ to get the best performance, see Sybase IQ System Administration Guide and Sybase IQ Performance and Tuning Guide.
