Backend-as-a-Service

Performance Shoot-out


Who is faster?
Baqend
Load time: ...

Global Benchmark Results


Benchmarked on AWS c4.xlarge instances in the different locations. The result of each vendor is the average over 10 consequitive loads with warmed up servers but cleared browser sessions to simulate first time visits.
The benchmark app runs in Baqend's free tier. Try for yourself!

Implementation Remarks


The web application used for the comparison is a proof of concept that simulates typical web workloads. It loads 30 small news articles, each including an image. The project can be found at GitHub. For the comparison, the same application is deployed in the free tier of all the competitors. There are, however, some limitations for the individual vendors that need to be considered:

  • Parse delivers hosted images by default with caching headers. On the one hand, this speeds up the page load for Parse. But on the other hand, a news editor could not change images anymore without customers seeing stale data because caches are not invalidated. We wanted to change this behaviour by configuring custom HTTP headers but could not find such settings.
  • Apiomat does not offer hosting in their free tier. That is why the website and the files are served from AWS S3 in the comparison. Only the news articles themselves come from Apiomat.
  • Kinvey does not support actual hosting of websites. You can, however, upload files and get an url back for it. Files are hosted on Google cloud storage.

Baqend achieves its unique performance by leveraging the HTTP caching infrastructure and combines it with new algorithms for cache invalidation. For more details on how Baqend enables caching of arbitrary resources and database objects visit our blog. To see how this benchmark translates into the real world take a look at our e-commerce case study.

Try Baqend Cloud Today!

Build your first instantly loading website

Get Started for Free