Performance prediction for concurrent database workloads

Abstract

Current trends in data management systems, such as cloud and multi-tenant databases, are leading to data processing environments that concurrently execute heterogeneous query workloads. At the same time, these systems need to satisfy diverse performance expectations. In these newly-emerging settings, avoiding potential Quality-of-Service (QoS) violations heavily relies on performance predictability, i.e., the ability to estimate the impact of concurrent query execution on the performance of individual queries in a continuously evolving workload.This paper presents a modeling approach to estimate the impact of concurrency on query performance for analytical workloads. Our solution relies on the analysis of query behavior in isolation, pairwise query interactions and sampling techniques to predict resource contention under various query mixes and concurrency levels. We introduce a simple yet powerful metric that accurately captures the joint effects of disk and memory contention on query performance in a single value. We also discuss predicting the execution behavior of a time-varying query workload through queryinteraction timelines, i.e., a fine-grained estimation of the time segments during which discrete mixes will be executed concurrently. Our experimental evaluation on top of PostgreSQL/TPC-H demonstrates that our models can provide query latency predictions within approximately 20% of the actual values in the average case.

Distribution-based query scheduling

“…Without any knowledge about r, one cannot infer the urgency of each query and as a result, one cannot apply any cost-or urgency-aware scheduling. Recently, there have been many efforts in the database community (e.g., [12,33,10]) to reduce the errors in the prediction of r. Moreover, recent studies (e.g., [11]) have shown that although it can be extremely difficult to predict the exact execution time of a particular query instance, it is relatively easy, e.g., through a brief period of online monitoring, to learn about the distribution of the execution time for queries with the same template or from the same family.…”

“…Solutions proposed so far cover both OLTP [12] and OLAP [22,36,35] queries, as well as both stand-alone [12,36] and concurrent [10,35] queries. Because most of these solutions are based on machine-learning approaches, and because real data contain noise and inherent uncertainties, it is reasonable to assume there exists unavoidable imprecision in the predicted query execution time.…”

“…Despite a great deal of decade-old and recent effort [12,10,22,36,35], accurate prediction of query running times for dynamic, concurrent workloads remains largely elusive.…”

Towards predicting query execution time for concurrent and dynamic database workloads

“…The current trend of offering database as a service (DaaS) makes this capacity even more attractive, since a DaaS provider needs to honor service level agreements (SLAs) to avoid loss of revenue and reputation. Recently, there has been substantial work on query execution time prediction [2,3,6,8,13,28]. Much of this work focuses on predicting the execution time for a single standalone query [3,8,13,28], while only a fraction of this work considers the more challenging problem of predicting the execution time for multiple concurrently-running queries [2,6].…”

“…systems usually allow multiple queries to execute concurrently. The existing work on concurrent query prediction [2,6], however, assumes that the workload is static, namely, the queries participating in the workload are known beforehand. While some workloads certainly conform to this assumption (e.g., the report-generation workloads described in [1]), others do not.…”