Most databases expected to be cloud-hosted by 2022

News  /  Published 08 Dec 2020  / 

According to a report by Gartner, 75% of all databases will be moved to the cloud by 2022, with only 5% considering repatriation on on-premises.

The analysts who co-authored the report said that they expect the cloud database management systems (DBMS) revenue will make for 50% of the total market revenue in 2023, leading also to fewer vendors. A growth in multi-cloud is also expected, increasing the complexity of deployments.

However, there is a seeming contradiction in the report. Indeed, it stated that DBMS vendors such as Redis Labs and Databricks are listed alongside cloud platform providers like AWS. Yet this is not comparable as the main options for running Redis in the cloud are AWS, Microsoft Azure, and Google Cloud, while it is Azure, AWS, and Alibaba for Databricks.

The report still makes good observations of multiple vendors. For instance, it informed us that AWS DBMS services are preferred for their performance and reliability. Indeed, AWS seems to have a better track record for availability and reliability than the other hyperscale providers. Yet, the analysts pointed out that AWS is rather behind some other hyperscale providers and most independent service providers.

On the contrary, Google was quite praised for its multi-cloud efforts as it is the first hyperscale CSP to deliver on the multi-cloud promise. However, there are still some concerns over unspecified issues over GCP’s service and support.

IBM and Microsoft are also at the top of cloud vendors. The report noted that Microsoft’s multi-cloud support with SQL Server is tainted by the fact that the versions vary, and that pricing and licensing is structured to favor Azure.

Concerning IBM, the report praised Oracle although running Oracle risks a double license penalty when run on other clouds, and the Oracle Database is not certified to run on all CSP infrastructure. Moreover, it was revealed that Oracle Database on-premises is quite expensive and difficult to manage.

 

Source: DevOpsOnline.co.uk