by: Paul M. Cray
In the API age, every millisecond counts. Understanding cloud performance is vital for businesses because the cloud powers mission-critical web services - applications on which they, their clients, partners, suppliers and users depend.
The 2024 edition of the APIContext Cloud Service Provider API Report uses our unmatched repository of historical test call data to shed light on the evolving landscape of cloud services in 2023, with actionable insights into the best cloud choices to optimize API quality.
Here are a few key findings.
The Speed Race: AWS Dominates
Dominance for DNS Time
AWS was (once again) totally dominant on DNS Time with an average for the year of 2 ms; no other managed better than 13 ms, and Azure, at 20 ms, was ten times slower than AWS.
Four of the top five Time to Connect locations are AWS data centers (across Asia, Europe, North America, and Oceania).
AWS US-East Performance Shift
AWS US-East (N. Virginia) has long been a performance benchmark, reflected in its popularity as the host location for many popular services. Yet, after four years of leading the pack in Time to Connect, it experienced a significant decline. In 2022, the average Time to Connect was 1.23 ms, but in 2023, it increased to 2.50 ms. The new champion for Time to Connect is AWS Asia-Pacific Northeast 3 (Osaka, Japan), with an average time of 2.28 ms.
IBM Leads for Total Time
IBM beats AWS for Total Time averaging 443 ms. AWS was second with 450 ms, and Azure landed at the bottom with 529 ms, so it’s not necessarily just DNS Time and Time to Connect that matters for best API performance. In 2023, Azure was consistently >75 ms slower globally than AWS.
Regional deterioration in latency
All regions were slower in terms of both DNS Time and Connect Time in 2023 compared to 2022. In the case of Connect Time, South America was slower by 3500%! North America has only improved by 1 ms for DNS Time since 2018, and, at 12 ms, is no longer the fastest region for that latency component. Europe and South America tie for first at 8 ms, and South Asia (11 ms) and East Asia (9 ms) are also faster.
Service Availability Challenges
In 2022, 18% of observed API services achieved 99.99% service availability. However, in 2023, this figure plummeted to just 7%.
PagerDuty stands out as a beacon of reliability, showcasing exceptional performance with Four 9s availability, managing to hit this mark for two consecutive years, something no other service achieved.
The Clouds are Not Flat
These are just a few of the findings from the APIContext 2024 Cloud Service Provider API Report. The report has more usable insights including recommendations for API owners, and detailed performance analysis. We share this report every year so you can make informed choices about which clouds, regions and locations are best for you. I hope it’s helpful and look forward to your feedback. Download the full report now.
And, if you want to know more about how APIContext can help you understand better the APIs you care about, get in touch with us.