May 14, 2014 by Matt Kalan | Comments
With such a large shift towards NoSQL technologies like MongoDB, there is a lot of discussion about the changing role of the Database Administrator (DBA). Many go so far as to say DBAs are no longer needed, an idea driven by new database capabilities that focus on agility with a dynamic schema instead of a fixed schema and features that make operational management easier in areas like high availability and horizontal scaling.
Based on my experience working with hundreds of customers to implement MongoDB in their organizations, there is still room for the DBA, even if their everyday tasks might take less time with MongoDB.
First, let's talk about the parts of their role that stay the same. There still needs to be someone to set up back-up/recovery processes, handle capacity planning, run maintenance tasks (e.g. upgrades), diagnose issues, do configuration management, and set up replication and sharding. In enterprises, often separate operational teams handle security, monitoring, and diagnosing common issues but that could be part of the DBA's role in some firms as well.
Notice I left out schema management; in MongoDB, the implementation of the schema is NOT predefined but is rather determined at the application level and the structure of the object is stored in the application team’s favorite programming language. This is incredibly valuable when business units request new features or data be added to the application. The application developer can simply add it in, and no change needs to be made in the database, enabling rapid agility. In this case, the DBA is no longer the middleman, and can focus on keeping the database up and running.
But in some cases, application changes are more complex, and development teams need a database expert to consult about schema design and its impact on performance, maintainability, and other factors. In this case, the role of the DBA transforms into the DB Adviser. In this role, the DB Advisor would work closely with the application development team that implements the schema. The DBA provides the process and due diligence to manage the pace of change rather than enforce the limitations of the relational schema.
Implementing the schema in the application might be disconcerting to some DBAs and others out there who rightfully worry about application developers making uninformed decisions and bringing down an application. Letting go of schema management might be a difficult step, but DBAs rely on application developers to make decisions in their application code on a regular basis. Creating a poorly constructed algorithm will bring the system to a crawl and querying fields in unexpected ways would cause performance problems as well.
Is choosing schema design in the database so much more responsibility than any other development decision?
The DBA should still guide developers, but without the enormous overhead of having to always update a relationship database schema to simply add a field! Your schema should be as dynamic as your business: agile with the optimal amount of control.
The DBA has always been the person accountable to ensure the reliability and performance of the database alongside development teams. However, in today's market, the DBA will continue performing most of the activities they do in the former relational-only world, but will advise in other areas to allow their business groups to innovate and iterate faster than their competition.
In this shift, the DBAs can take on new roles to help their businesses achieve more, faster. See how you can help your organizations get faster, better and leaner with MongoDB.