The Cloud Performance Benchmark from ThousandEyes compares global network performance and connectivity differences between the five major public cloud providers — Amazon Web Services (AWS), Google Cloud Platform (GCP), Microsoft Azure, Alibaba Cloud and IBM Cloud — proving that, when it comes to performance, not all clouds are created equal.
"When businesses need to decide which cloud provider best meets their needs, one metric that's notably missing from their assessments has been performance data, mainly because it's never been available or has, at best, been myopic," said Archana Kesavan, research author and Director of Product Marketing at ThousandEyes. "The second annual Cloud Performance Benchmark gives businesses that comparative data. Understanding cloud performance is essential for planning and for ongoing measurement so you can be assured that you're providing customers and employees with the best possible performance."
Results were derived from an analysis of over 320 million data points collected from 98 global metro locations over the course of 30 days, using both end-to-end performance and in-depth path measurement techniques.
Notable findings include:
RELIANCE ON PUBLIC CLOUD
Some cloud providers rely heavily on the public Internet to transport traffic instead of their backbones, which can impact performance predictability. While Google Cloud and Azure rely heavily on their private backbone networks to transport their customer traffic, protecting it from performance variations associated with delivering over the public Internet, AWS and Alibaba Cloud rely heavily on the public Internet for the majority of transport, resulting in greater operational risk that can impact performance predictability. IBM takes a hybrid approach that varies regionally.
Regional performance differences
Latin America and Asia have the highest performance variations across all clouds, whereas in North America, cloud performance is generally comparable. Regional performance differences can make a significant impact in terms of performance gains or losses.
AWS Global Accelerator
AWS Global Accelerator doesn't always out-perform the Internet. AWS Global Accelerator was introduced in November 2018 to let customers use the AWS private backbone network for a fee — rather than use the public Internet, which is AWS' default behavior.
While there are many examples of performance gains in various regions around the world, the Global Accelerator is not a one-size-fits-all solution, as there are several examples where the Internet actually performs faster and more reliably than Global Accelerator, or, the results are negligible.
Broadband ISP choice makes a difference in cloud performance. Businesses looking to get every performance edge possible should consider which broadband ISP provider they select, depending on which cloud they most heavily rely upon. There are performance gains and losses depending on which broadband provider businesses use to connect to each cloud.
The Great Firewall of China
All cloud providers pay a performance toll when crossing the Great Firewall of China. Despite Alibaba's origins in China, it experiences packet loss when crossing through China's Great Firewall just as all the other cloud providers do, showing it does not get any preferential treatment.
Enterprises serving customers in China, but hesitant to pick a hosting region in China due to stringent regulations on traffic and data privacy, might consider Hong Kong as a viable option — Alibaba Cloud traffic experienced the least packet loss from Hong Kong to China, followed by Azure and IBM.
Benchmark Findings By Cloud Provider
The inaugural Cloud Performance Benchmark published in November of 2018 benchmarked the three biggest cloud providers (by market share) AWS, Azure and GCP. Alibaba Cloud and IBM Cloud are new this year, meaning they have no prior comparison points. There are notable differences between each cloud in terms of overall performance and connectivity architectures.
Key findings, by cloud provider include:
■ AWS generally demonstrates low latency and made some minor improvements over the last year. Its performance predictability metrics, however, noticeably improved, with its most dramatic change in Asia, which showed a 42 percent reduction in variability over last year. However, when compared to Azure and GCP, it still has lower performance predictability due to its extensive reliance on the Internet rather than leveraging its own backbone for delivery.
■ Azure continues its strong network performance based on aggressive use of its own backbone to carry user traffic to cloud hosting regions. Standout changes from 2018 to 2019 include on average a 50 percent improvement in performance predictability in Sydney, whereas in India, there was a 31 percent decrease in performance predictability. Despite a slight decrease year over year, Azure continues to lead in performance predictability in Asia when compared to the other cloud providers.
■ Google Cloud continues to favor use of its own backbone for user to cloud hosting region traffic delivery and has strong performance in most regions, but it still has some significant global gaps that haven't been resolved since last year's report. Traffic from Europe and Africa takes 2.5-3 times longer to get to India, going around the rest of the world instead of taking a direct route. Google Cloud also decreased visibility into their internal network, making it harder for its users to understand its network paths and performance.
■ Alibaba Cloud offers comparable performance to other cloud providers. It closely resembles AWS in terms of connectivity patterns, region locations, and even region naming constructs. Like AWS, Alibaba Cloud leans heavily on the Internet rather than their own private network backbone for the majority of user traffic transport. Uniquely, inter-region traffic between Alibaba Cloud regions is not contained within their own cloud infrastructure as it is for the other four, instead it exits Alibaba Cloud, traverses the Internet and then makes its way back into Alibaba Cloud.
■ IBM Cloud performance is comparable to the major players. IBM takes a hybrid approach to traffic delivery, fluctuating between using its own private backbone and the public Internet, depending on which regions user traffic is accessing.
"Historically, most IT professionals select cloud providers based on factors such as price or proximity to users, but often overlook the underlying network architecture of the cloud providers, which can have a significant impact on performance," said Zeus Kerravala, principal analyst, ZK Research. "The 2019 ThousandEyes Cloud Performance Benchmark shows just how widely performance can vary from the top cloud service providers and reinforces the point that businesses should remain diligent about collecting network intelligence and measure their own performance to ensure they are getting what they expect from their cloud providers."
In order to properly sort through all monitoring noise and identify true problems, their causes, and to prioritize them for response by the IT team, they have created and built a revolutionary new system using a meta-cognitive model ...
As we shift further into a digital-first world, where having a reliable online experience becomes more essential, Site Reliability Engineers remain in-demand among organizations of all sizes ... This diverse set of skills and values can be difficult to interview for. In this blog, we'll get you started with some example questions and processes to find your ideal SRE ...
US government agencies are bringing more of their employees back into the office and implementing hybrid work schedules, but federal workers are worried that their agencies' IT architectures aren't built to handle the "new normal." They fear that the reactive, manual methods used by the current systems in dealing with user, IT architecture and application problems will degrade the user experience and negatively affect productivity. In fact, according to a recent survey, many federal employees are concerned that they won't work as effectively back in the office as they did at home ...
Users today expect a seamless, uninterrupted experience when interacting with their web and mobile apps. Their expectations have continued to grow in tandem with their appetite for new features and consistent updates. Mobile apps have responded by increasing their release cadence by up to 40%, releasing a new full version of their app every 4-5 days, as determined in this year's SmartBear State of Software Quality | Application Stability Index report ...
In this second part of the blog series, we look at how adopting AIOps capabilities can drive business value for an organization ...