A report that compares global network performance of the three major public cloud providers – Amazon Web Services (AWS), Google Cloud Platform (GCP) and Microsoft Azure has revealed striking differences in their capabilities across different geographies.
Network monitoring company ThousandEyes analysed over 160 million data points for the report, assessing 55 cloud regions using TCP-based probes to collect “hop-by-hop network path data” along with network metrics like loss, latency and jitter, the company said.
Among the key takehomes: AWS underperforms significantly in Asia, while GCP is almost three times slower than its two rivals to connect Europe and India. Multi-cloud performance between the three providers meanwhile is consistent and reliable, however, the report found.
AWS vs Azure vs GCP: Core Differences
One of the most striking findings was AWS’s reliance on the public internet to handle network traffic. Amazon’s network design forces traffic from the end user through the public Internet, only to enter the AWS backbone closest to the target region, ThousandEyes said.
“This behavior is in stark contrast to how Azure and GCP design their respective networks. In the latter, traffic from the end-user, irrespective of geographical location, is absorbed into their internal backbone network closest to the user, relying less
on the Internet to move traffic between the two locations.”
ThousandEyes said: “Google and Microsoft have the historical advantage of building and maintaining a vast backbone network. AWS, the current market leader in public cloud ofierings, focused initially on rapid delivery of services to the market, rather than building out a massive backbone network. Given their current position, increasing profitability and recent investments in undersea cables, it is likely that their connectivity architecture will change over time.”
Geographical Latency: A Cloud Contrast
Locations of cloud data centers naturally also directly impact network latency.
When connecting Europe to India, for example GCP exhibited three times the network latency compared to AWS and Azure.
In Asia, GCP and Azure exhibited more network performance stability than AWS, which demonstrated 35 percent less network performance stability than GCP and 56 percent less than Azure.
When connecting Europe to Singapore, AWS and GCP were 1.5 times slower than Azure.
ThousandEyes noted: “Performance variations in certain geographies highlight the reality that public cloud vendors do not yet have consistent performance globally. For an enterprise with a global customer base, the insights from these findings should factor into cloud architecture decisions.”
The full 2018 Public Cloud Performance Benchmark Report can be found here.