The Amazon public cloud platform’s reliance on the public internet to handle network traffic could put users at a heightened risk of performance issues, claims a new report by network monitoring company ThousandEyes.
In its analysis of how the Amazon, Google and Microsoft public cloud platforms process traffic, featured in its inaugural Public cloud performance benchmark report, the firm details how it claims all three firms handle traffic passing through their respective network architectures.
According to its findings, there is a stark contrast between how much exposure to the public internet traffic passing along the network paths within Google and Microsoft’s cloud setup gets compared to Amazon’s.
“Amazon Web Services’ [AWS’s] network design forces traffic from the user through the public internet, only to enter the AWS backbone closest to the target region. This is in stark contrast to how Azure [Microsoft] and Google Cloud Platform design their respective networks,” the report states.
“In the latter, traffic from the user, irrespective of geographical location, is absorbed into their internal backbone network closer to the user, relying less on the internet to move traffic between the two locations.”
And this distinction is important for enterprise IT buyers to know about, the report continues, because the more reliant a cloud provider’s network is on the public internet to function, the higher risk of experiencing technical difficulties.
“For enterprises building their services on the public cloud, cloud connectivity architectures can directly impact the end-user experience,” the report continues. “AWS deployments feature increased reliance on the internet and are thereby subject to lower performance predictability, plus higher operational challenges and risks.”
Computer Weekly contacted AWS for a response to the research’s findings, but had not received a response at the time of publication.
As for why Amazon’s approach to network design differs so widely to Microsoft and Google’s could be down to the differences in how they came to be cloud providers in the first places, the report suggests.
“Google and Microsoft have the historical advantage of building and maintaining a vast backbone network. AWS focused initially on rapid delivery of services to the market, rather than building out a massive backbone network,” the report adds.
“Given their current position, increasing profitability and recent investments in undersea cables, it is likely that their connectivity architecture will change over time.”
The report is, ThousandEyes claims, the first to benchmark the global network performance of Amazon, Google and Microsoft’s public cloud platforms, and is intended to provide enterprise IT buyers with an unbiased view of how they compare.
“Previously available studies on the ‘big three’ focus on services offered, pricing tiers, and global datacentre presence. However, performance studies of public cloud providers have been missing in action,” the document states.
“It is imperative for enterprise IT leaders to understand that cloud architectures are complex and not rely on network performance and connectivity assumptions or instincts while designing them.”
Mix and matching services
And for the enterprises looking to adopt a multi-cloud strategy, that will see them mix and match services from Amazon, Google and Microsoft, the report also suggests they should not run into many network-related difficulties when attempting to do so.
“Multi-cloud performance reflects a symbiotic relationship between AWS, Azure and GCP. Traffic between cloud providers almost never exits the three provider backbone networks, manifesting as negligible loss and jitter in end-to-end communication,” the report states.
“Despite being competitors, the three providers peer directly with each other, eliminating the dependence on third-party ISPs for multi-cloud communication. All three cloud providers have vast networks and are well connected across multiple popular colocation facilities.”