Automatically classifying database workloads

The Psychic–Skeptic Prediction framework for effective monitoring of DBMS workloads

“…The prediction problem is the focus of this paper. Our background work on classifying workloads is described elsewhere [14,15,17] and is summarized below.…”

Consistent on-line classification of dbs workload events

“…Related research approaches do not meet the requirements of DBS workload classification. In [3] the classification of database workload into OLTP and DSS is described. The approach is intended only to distinguish between these two workload types and cannot be used for DBS workload classification in general.…”

“…1. As for large workloads it is likely that the quality loss is exceeded even when the class limit is reached, we avoid the effort for a binary search in this case (1)(2)(3)(4).…”

Poster Session Problem definition for effective workload management

“…Recent work [6,5] on workload classification presents techniques of on-line identification of workload types like the ones used in On-line Transaction Processing (OLTP) and Decision Support Systems (DSS). While the type of workload submitted is important for an efficient database tuning we consider the problem of tuning the database system in accordance with different classes of users that have different performance expectations from the database system.…”

“…While the type of workload submitted is important for an efficient database tuning we consider the problem of tuning the database system in accordance with different classes of users that have different performance expectations from the database system. The approach presented in [6,5] is orthogonal with ours in the sense that for each monitored class of users that we define, an on-line classifier can be used to identify the OLTP and DSS workloads. In this case two performance goals can be specified for each class of users.…”