Top 5 Considerations When Evaluating NoSQL Databases: Conclusion

The Nexus Architecture

MongoDB’s design philosophy is focused on combining the critical capabilities of relational databases with the innovations of NoSQL technologies. Our vision is to leverage the work that Oracle and others have done over the last 40 years to make relational databases what they are today. Rather than discard decades of proven database maturity, MongoDB is picking up where they left off by combining key relational database capabilities with the work that Internet pioneers have done to address the requirements of modern applications.

Figure 1: MongoDB Nexus Architecture, blending the best of relational and NoSQL technologies

Relational databases have reliably served applications for many years, and offer features that remain critical today as developers build the next generation of applications:

  • Expressive query language & secondary Indexes. Users should be able to access and manipulate their data in sophisticated ways to support both operational and analytical applications. Indexes play a critical role in providing efOcient access to data, supported natively by the database rather than maintained in application code.
  • Strong consistency. Applications should be able to immediately read what has been written to the database. It is much more complex to build applications around an eventually consistent model, imposing significant work on the developer, even for the most sophisticated engineering teams.
  • Enterprise Management and Integrations. Databases are just one piece of application infrastructure and need to fit seamlessly into the enterprise IT stack. Organizations need a database that can be secured, monitored, automated, and integrated with their existing technology infrastructure, processes, and staff, including operations teams, DBAs, and data engineers.

However, modern applications impose requirements not addressed by relational databases, and this has driven the development of NoSQL databases which offer:

  • Flexible Data Model. NoSQL databases emerged to address the requirements for the data we see dominating modern applications. Whether document, graph, key-value, or wide-column, all of them offer a flexible data model, making it easy to store and combine data of any structure and allow dynamic modification of the schema without downtime or performance impact.
  • Scalability and Performance. NoSQL databases were all built with a focus on scalability, so they all include some form of sharding or partitioning. This allows the database to be scaled out across commodity hardware deployed on-premises or in the cloud, enabling almost unlimited growth with higher throughput and lower
    latency than relational databases.
  • Always-On Global Deployments. NoSQL databases are designed for continuously available systems that provide a consistent, high-quality experience for users all over the world. They are designed to run across many nodes, including replication to automatically synchronize data across servers, racks, and geographically-dispersed data centers.

While offering these innovations, NoSQL systems have sacriOced the critical capabilities that people have come to expect and rely upon from relational databases. MongoDB offers a different approach. With its Nexus Architecture, MongoDB is the only database that harnesses the innovations of NoSQL while maintaining the foundation of relational databases.


As the technology landscape evolves, organizations increasingly Ond the need to evaluate new databases to support changing application and business requirements. The media hype around non-relational databases and the commensurate lack of clarity in the market makes it important for organizations to understand the differences between the available solutions. As discussed in this paper, key criteria to consider when evaluating these technologies are the data model, query model, consistency model and APIs, as well as commercial support and community strength. Many organizations Ond that document databases such as MongoDB are best suited to meet these criteria, though we encourage technology decision makers to evaluate these considerations for themselves.