The cloud database check: Leading providers in analytics and telecommunication 

In this article you will read,

• what strengths and weaknesses the PostgreSQL databases from Azure, AWS and T Cloud Public have,
• what advantages T Cloud Public offers beyond price and performance,
• and where you can download the independent benchmark study.

The use of databases from the cloud, especially open source variants such as PostgreSQL, has become well-established. But users get different performance for their money, even if they use seemingly identical configurations. The benchANT benchmark shows the differences in latency, throughput and price for the PostgreSQL DBaaS offerings Azure, AWS and T Cloud Public.

Databases from the cloud are convincing because they offer tangible speed and cost benefits. Their use is significantly simplified. However, despite identical specifications for hardware and software on paper, there are significant differences. The implementation of the DBaaS layer has a significant influence on the performance of a database instance – and therefore also the price/performance ratio.

One of the most popular databases in the IT world is PostgreSQL. According to DB-Engine, the open source database is in fourth place on the popularity scale. It is steadily catching up with the "big three" (Oracle, MySQL, Microsoft SQL). PostgreSQL is part of the standard portfolio of every cloud provider.

In March, the independent benchmarker benchANT analyzed the PostgreSQL offerings from AWS, Azure and T Cloud Public on behalf of T Cloud Public. The result is a benchmark study that shines light on the performance and resulting price/performance ratio.

Two business scenarios were used to test the performance of the PostgreSQL databases: data analysis and support for the business of a telco provider.

Mobile telecommunications providers maintain so-called HLR databases (Home Location Register). Master data such as the telephone number (no changes), the subscribed services (infrequent changes), but also the current location of the cell phone (very frequent changes) are stored in these databases. For each call, the HLR databases of the caller and the receiver are contacted to establish the connection. Speed and high scalability are important here – a demanding deployment scenario with high data throughput. The TATP benchmark (Telecommunication Application Transaction Processing) simulates such an HLR database.

Analytics cases describe a completely different application scenario. TPC-H (Transaction Processing Performance Council) is a benchmark that describes how well databases support business decisions. TPC-H triggers (complex) ad-hoc queries that examine large volumes of data and generate analyses or answers to business questions. The metric that is recorded here is "queries per hour". The requests to the database are significantly more extensive and time-consuming than simple OLTP (Online Transaction Processing) tasks. The focus here is on mastering large volumes of data (big data).

Overall, the results show that the PostgreSQL RDS service of T Cloud Public is on par with the DBaaS solutions of the US hyperscalers in terms of performance and price/performance ratio in the selected benchmarking scenarios.

T Cloud Public is the only provider that includes support by e-mail and phone in the price. Together with the GDPR advantage (and the associated independence from the US Cloud Act), it offers a well-rounded and powerful PostgreSQL DBaaS product for European companies.

You want the data? Click here to download the report. benchANT also did a benchmark for MySQL – and tested this database in use for e-commerce. You can find the results of this benchmark here.