Abstract
From a historical perspective, relational database management systems (RDBMSs) have integrated many specialized systems and data models back into the RDBMS over time. New workloads motivated specialized systems for performance, but over time, general-purpose RDBMSs absorbed this functionality to avoid boundary crossing. We already witnessed this process for object-relational functionality, XML and JSON data types, OLAP/HTAP systems, and RDF/graph processing, while for natural language processing (NLP), time series, and machine learning (ML), the outcomes remain unclear. Interestingly, graph processing, NLP, and time series are largely ML workloads too. For this reason, integrating data management and ML is of high practical relevance and has been addressed by (1) integrating ML into RDBMSs, and (2) specialized ML systems. The paper "Declarative Recursive Computation on an RDBMS" [3] by Jankov et al. makes a very valuable contribution by reconciling these two areas and showing the potential of recursive computations on an RDBMS, as the backend-not necessarily frontend-for large-scale machine learning.
- Z. Cai, Z. Vagena, L. L. Perez, S. Arumugam, P. J. Haas, and C. M. Jermaine. Simulation of Database-Valued Markov Chains Using SimSQL. In SIGMOD, 2013. Google ScholarDigital Library
- Z. J. Gao, S. Luo, L. L. Perez, and C. Jermaine. The BUDS Language for Distributed Bayesian Machine Learning. In SIGMOD, 2017. Google ScholarDigital Library
- D. Jankov, S. Luo, B. Yuan, Z. Cai, J. Zou, C. Jermaine, and Z. J. Gao. Declarative Recursive Computation on an RDBMS. PVLDB, 12(7), 2019. Google ScholarDigital Library
- S. Luo, Z. J. Gao, M. N. Gubanov, L. L. Perez, and C. M. Jermaine. Scalable Linear Algebra on a Relational Database System. In ICDE, 2017. Google ScholarDigital Library
Recommendations
Technical Perspective: Conjunctive Queries with Comparisons
Query processing, the art of efficiently executing a relational query on a given database, is a foundational and core area in data management research. Established at the dawn of relational database systems in the 1970's, relational query processing ...
Technical Perspective of Concurrent Prefix Recovery: Performing CPR on a Database
Where do novel database system research results come from? In the 1970's, most systems research papers proposed mechanisms to support abstractions that were being explored for the first time, such as data translation, indexing, query optimization, high ...
Technical Perspective: Probabilistic Data with Continuous Distributions
The paper entitled "Probabilistic Data with Continuous Distributions" overviews recent work on the foundations of infinite probabilistic databases [3, 2]. Prior work on probabilistic databases (PDBs) focused almost exclusively on the finite case: A ...
Comments