Reader survey Once upon a time, not so very long ago, relational databases were pretty much the only game in town – for mainstream organizations, at least. As a result, they underpin huge numbers of business applications of all kinds, and when we surveyed readers of The Register recently, it was clear many organisations expect their relational database usage to increase.
And yet the database world is changing. Relational database management systems (RDBMSes) still dominate, though application needs have been evolving and broadening, and in recent years new approaches, such as NoSQL and Graph databases, have developed at a rapid rate. While such systems are much less widely implemented, our research indicates their use is expected to grow as well.
The problem, of course, is that this means many organizations find themselves running a mix of database technologies. Worse, the server and storage infrastructure used to support both long-established RDBMSes and newer technologies may not always be the most up to date. Indeed, the research results show that a significant proportion of those taking part in the survey believe many of their RDBMS implementations are sub-optimal and would benefit from the software being upgraded or replaced.
Creaking at the seams
In particular, the survey results show that considerable numbers of RDBMS applications run on old server or storage hardware that is no longer considered fit for purpose. To be fair, this is not unusual for systems built to last over long periods of time, though deeper analysis did throw up some interesting correlations.
Most notably, where databases are running on storage platforms that are considered to be too old or not up to the job, there was a clear correlation with negative factors such as poor performance and availability issues. This result was predictable because more than half of the respondents also agree that the underlying storage layer impacts how well database application needs are met.
So given that the challenges of inadequate infrastructure are acknowledged, what can be done to improve things? The survey did find that upgrading some RDBMSes to more modern database platforms would be attractive for some workloads, though many respondents also thought it would be tricky to achieve. In contrast, survey respondents consider that moving existing relational database workloads to new storage platforms would be less disruptive, yet still able to deliver substantial benefits.
What’s the plan?
Considering the importance that relational databases and the applications they support have in business, it is no surprise to find many respondents indicating that action is planned. The obvious step of upgrading to a newer release of the database is something that a majority will either definitely or probably take, but this can almost be regarded as "business as usual." Interestingly though, considerable numbers will at least look at either re-engineering some of their systems to run on an alternative database architecture or switching to a different RDBMS.
The other general area where we expect activity concerns the storage platforms on which database systems run. As we have mentioned, the importance of storage to database applications is reasonably well understood, and large numbers of respondents expect to be expanding the storage platforms they use. Almost half said that there was at least a possibility of moving databases to entirely new storage platforms. Sadly, many also noted that the usual challenges concerning lack of budget or not having access to the required resources and skills could defer some projects.
If this has grabbed your attention and you would like to find out more of the detail, download the report Modernizing database technology stacks. ®