NoSE: Schema Design for NoSQL Applications

Database design is critical for high performance in relational databases and a myriad of tools exist to aid application designers in selecting an appropriate schema. While the problem of schema optimization is also highly relevant for NoSQL databases, existing tools for relational databases are inadequate in that setting. Application designers wishing to use a NoSQL database instead rely on rules of thumb to select an appropriate schema. We present a system for recommending database schemas for NoSQL applications. Our cost-based approach uses a novel binary integer programming formulation to guide the mapping from the application's conceptual data model to a database schema. We implemented a prototype of this approach for the Cassandra extensible record store. Our prototype, the NoSQL Schema Evaluator (NoSE) is able to capture rules of thumb used by expert designers without explicitly encoding the rules. Automating the design process allows NoSE to produce efficient schemas and to examine more alternatives than would be possible with a manual rule-based approach.

1 Introduction

NoSQL systems have become a popular choice as database backends for applications because of the high performance, scalability, and availability that they provide. In this paper, we focus on systems that Cattell termed extensible record stores in his taxonomy of NoSQL systems [1]. In these extensible record stores, applications can create tables of records, with each record identified by a key. However, the application need not define the set of columns in the records in advance. Instead, each record can have an arbitrary collection of columns, each with an associated value. Because of this flexibility, applications can encode their data in both the keys and column values. We refer to tables in such systems as column families . Examples of extensible record stores that support this column family model include Cassandra  [2], HBase [3], and Bigtable  [4].