Open source DBMS software is on the rise, Gartner analysts Merv Adrian and Donald Feinberg said in a February 2018 report. They predict that by 2022, more than 70% of new applications developed by corporate users will run on an open source database management system, and that half of the existing relational database installations built on commercial DBMS technologies will be converted to open source platforms or in the process of being converted.
That's four years later than an earlier Gartner projection, which foresaw those usage levels being reached in 2018. But the report points to high growth rates for open source database revenues in recent years, as well as a "profound shift" in development and software packaging among vendors to put more emphasis on the use of open source technology.
For users, the biggest appeal of an open source DBMS isn't the ability to access the software's source code and develop new features themselves. The real lure is lower cost, according to the analysts, who wrote that open source offerings are usually far less expensive than their commercial counterparts. The report includes an example that shows an open source relational DBMS costing as little as 1% of the price of Oracle Database over three years.
But there are a lot of open source technologies for prospective users to evaluate -- and other issues beyond cost to consider, including technical support, application needs and skills requirements. In this Q&A, Adrian discusses the status of open source database software and offers advice on working with it.
How widely are open source databases being used at this point?
Merv Adrian: Gartner believes that 95% of mainstream IT organizations use open source software overall in their mission-critical IT portfolios, and that open source DBMS technology will account for more than 10% of total spending on database software by 2019 due to increased enterprise adoption.
Obviously, it's not being used everywhere -- and open source vendors love to say they've replaced existing vendors as soon as they have one installed instance. But the reality is that many open source products have a foothold, and they're all doing their best to expand to broader usage.
Are open source relational databases a viable alternative now to Oracle, SQL Server and the other top commercial database technologies for enterprise applications in general?
Adrian: Although open source RDBMS [relational database management system] options have typically been aimed at providing good enough functionality for a reduced total cost of ownership, they have matured and gained ground. Sometimes, though not always, open source databases can be used for significant, mission-critical applications.
In most cases, though, it's important to ensure that the selected open source product, even after it has proved its capability with appropriate testing at scale and under relevant usage loads, is offered by a significant, viable vendor with demonstrable, enterprise-class support capabilities and features.
Is there any difference between the skills needed to work with an open source database compared to Oracle and the like?
Adrian: It depends on the type of DBMS being used. But, in general, the database administration tools provided by open source vendors are often only a subset of what the commercial players offer. More hands-on operation is the norm, and the required skills will be at a premium.
Is the growth of the cloud helping to boost the adoption of open source DBMS software?
Adrian: As elsewhere in the IT landscape, the impact of the cloud has been profound. Gartner estimates that Amazon Web Services [AWS] is now the leading vendor of open source databases. In addition, the cloud provides a low-cost, low-impact sandbox for assessing open source DBMS technologies. Cloud platforms like AWS and Microsoft Azure offer utilities that can be used to do a preliminary impact analysis on migrating databases to open source alternatives and to select pilot-project candidates.
In the past, it would have been necessary to buy and provision hardware and install the relevant software, then design the target environment and move data to it before testing applications against an open source system. These utilities have created a more frictionless environment that is driving increased trials.
Do open source database vendors have trouble keeping up with commercial competitors on technology improvements because of the community development process?
Adrian: At the margin with new functionality, they can be even more aggressive than the traditional vendors, though they may skip some of the features that the latter have hardened over years.
On the other hand, the big commercial DBMS vendors are increasingly incorporating open source components into their proprietary offerings, creating a competitive challenge for the pure-play vendors. The result is that open source database technology is being weaponized as a foundation for the newer vendors and a force multiplier by the large ones.
What do you think about the relative strengths of the three most prominent relational open source DBMS technologies: MySQL, MariaDB and PostgreSQL?
Adrian: There are vendors offering good support and continuing R&D on all three of those offerings, which represent the two most widely deployed RDBMS platforms in the world, since MariaDB is a fork of MySQL. The differences between the three technologies are worth a discussion of their own, but they all can be used to good effect on a large proportion of the business needs of most organizations -- if care is exercised on application fit.