Die Leistungsfähigkeit von virtuellen Maschinen in der Cloud ist von mehreren Faktoren abhängig. Vor allem CPU- und RAM-Ausstattung, die verwendete Virtualisierung, aber auch die Speicheroptionen determinieren die Bearbeitungsgeschwindigkeit.
Für diesen Benchmark hat das Beratungsunternehmen Cloud Spectator die Rechen- und Speicher-Performance der IONOS Cloud gemessen und sie in Relation zu entsprechenden Angeboten von Amazon Web Services (AWS), Google Cloud Engine (GCE) und Microsoft Azure gesetzt. Sie erfahren, welche Cloud die höchste Performance bietet, und wo Sie das beste Preis-Leistungsverhältnis erhalten.
Originalauszug aus dem Dokument:
All VMs went through the exact same setup process, which involved updating all packages then rebooting, followed by entire disk partitioning. Each VM was left alone for one hour after mounting each storage device to allow the operating system to initialize the filesystem (via ext4lazyinit). Each VM was tested “as is” with no kernel or operating system tweaks or optimizations applied.
Test Design Considerations
Infrastructure performance testing was conducted on specific VM and Storage types for each provider.
Provider VM configurations may yield different results based on underlying infrastructure, time of day, number of tenants running other workloads on the same hardware, virtualization technology and settings (e.g. shared resources), and other technology factors.
Furthermore, issues such as user contention or physical hardware malfunctions can also cause suboptimal performance. The VMs and Storage volumes selected for this engagement were generally available specified offerings from the various providers.
Infrastructure Testing: Compute
Compute results were obtained using GeekBench5, SysBench CPU, and SysBench RAM benchmarks. GeekBench5 was performed with default / standard testing options. SysBench CPU and RAM benchmarks were run with total thread counts equaling the vCPU count of the VM.
Infrastructure Testing: Storage
Storage results were obtained using FIO (Flexible I/O tester) using 4KB block size and a total job / thread count that matches each VMs vCPU count. Multiple 300-second iterations were conducted to compensate for the high variability often seen when stressing storage volumes. Results were gathered and represented in IOPs (input/output operations per second). The results displayed in this report are derived from the average IOP value recorded during each test.