Results of industry-wide survey on Postgres high availability

Technology decision-makers have moved many critical business processes to the cloud—yet cloud service outages are far from rare. Annually, there are dozens of significant outages with an inevitable handful being extremely serious, lasting many hours and disrupting businesses of every size. Outages bring risks of data loss, including vital transaction records, not to mention missed business opportunities, poor customer service, and reputational damage.

A new survey conducted by Foundry (publisher of CIO, and other leading IT publications) and pgEdge shows that IT leaders are acutely aware of the risks of cloud region failure given that 82% cite it as a concern. While in the last year, 21% directly experienced cloud failure that exceeded their downtime requirements.

Seventy-five percent of survey respondents list their functional role as IT / Networking / Security, with 13% at the C-level (CIO / top IT executive / Chief Data Scientist). Three-quarters (75%) report that their companies are currently using PostgreSQL in development / testing, highlighting the practical need for these solutions to manage business processes.

The company size of respondents varied, with 62% in the 1,000 - 9,999 employee range and 12% employing 10,000 or more people. This shows that there is a common need for high availability strategies and failover tools among small, midsize, and large enterprises alike. Of the top industries represented, Financial services was  37%, software and computing was 20% and manufacturing was 12%.

DOWNLOAD FULL SURVEY REPORT

As seen above, the highest percentage of respondents are opting for cloud-native tools. However, while this might seem sensible, doing so fails to take into account the likelihood of a cloud region failure (noting the 21% that have experienced such a failure in the past year, as per Figure 1 above).

Still, the variety of strategies used for managing failover that are reflected in the survey show that IT decision-makers are open to different approaches to providing high availability. Respondents anticipate a number of benefits from a high-availability or distributed PostgreSQL solution led by increased uptime or zero downtime operations (53%).  However, cost savings compared to proprietary databases (36%) is also sought by respondents, as is the ability to deploy databases across multiple or hybrid cloud environments.

To that end, multi-master replication (MMR) is generating strong interest (47%) among organizations that deploy applications across multiple cloud regions.

As the name suggests, a multi-master (also referred to as active active) implementation establishes several master databases spread across different geographic regions. Each database can handle read and write traffic simultaneously, ensuring eventual data consistency through bi-directional replication (BDR) and conflict resolution.

By contrast, in a traditional monolithic architecture, only one database can perform read-write operations while distributed nodes are read-only.

It's important to note, in a multi-master configuration, any of the several master databases can perform a failover if one of the other masters go offline. Consequently, all nodes collectively enable higher uptime than is possible in a monolithic database implementation.