Aggregates caching in columnar in-memory databases

Abstract. The mixed database workloads found in enterprise applica-tions are comprised of short-running transactional as well as analyti-cal queries with resource-intensive data aggregations. In this context, caching the query results of long-running queries is desirable as it in-creases the overall performance. However, traditional caching approaches are inefficient in a way that changes in the base data result in invalidation or recalculation of cached results. Columnar in-memory databases with a main-delta architecture are well-suited for a novel caching mechanism for aggregate queries that is the main contribution of this paper. With the separation into read-optimized main storage and write-optimized delta storage, we do not invalidate cached query results when new data is inserted to the delta storage. Instead, we use the cached query result and combine it with the newly added records in the delta storage. We evaluate this caching mechanism with mixed database workloads and show how it compares to existing work in this area.