In part 1 of this post, we discussed the advantages a NoSQL database like MongoDB has over a traditional RDBMS. Of course, selecting the right database is not just about comparing features, it’s also a matter of what business need is being filled. Here are some of the use cases where a NoSQL database like MongoDB is the best choice.
Web applications. MongoDB is well suited as a primary database for web applications because it provides a feature-rich document model, allows for easy implementation of queries, is highly scalable, and can process and store real-time analytical data. This data, along with other visualization tools, can produce information that can help organizations improve their business value.
Agile development. Many organizations prefer an agile development model to build applications, and because MongoDB does not have a fixed schema, it works well with this model.
Analytics and logging. MongoDB’s relevance to analytics is derived from its speed, its targeted atomic updates, and its capped collection ability. Atomic updates allow applications to efficiently push data into an array. Capped collection used for logging, provides fixed allocation and gets aged out automatically. Storing log information in the database rather than a file system provides easier organization and superior query power. Instead of relying on another programming tool to read the log file, this is accomplished with MongoDB’s query language.
Variable schema. If an application needs to consume JSON API, having a database that can easily consume JSON data format directly is an advantage. If the data structure cannot be determined before consuming it, MongoDB’s schema-less approach significantly simplifies the data-modeling task.
The latest version of MongoDB has improved on some of the limitations present in earlier releases, including reducing row-level concurrency issues and adding secondary indexes to improve performance. It also provides a more mature level of administrative tools, similar to those in RDBMSs.
Nevertheless, MongoDB does have limitations to keep in mind, such as virtual memory mapping for shared environments. Because of this, memory-mapped files should be used with journaling or with replication enabled (to avoid any issues caused by unclean shutdowns). Also, MongoDB runs significantly better on 64-bit machines than on 32-bit machines.
MongoDB is designed for modern applications, providing dynamic queries, indexes, aggregation framework, replication, sharding, and native drivers for most modern programming languages. Its document data model compliments the object structure in applications, making it an ideal choice for organizations requiring dynamic schemas, scalability, and shorter development cycles.
Because some of MongoDB’s features are very similar to those of an RDBMS, it’s ideal for companies that want to migrate some of their small and business critical applications away from RDBMS. However, a successful transition depends on having a partner with in-depth experience working with nextgen databases, proven data management strategies, and a roadmap for migrating from RDBMS to NoSQL.
Date de la publication : 2016-04-18