APImetrics Releases Largest Ever Cloud API Performance Report
The efficiency of cloud services is getting better, however, depending too heavily on just one data center location poses a significant threat to the overall quality and effectiveness.
APImetrics is pleased to announce the publication of its most extensive Cloud API Performance Report to date.
The Comprehensive Yearly Industry Report provided by the group, community, and information backing APImetrics.
APImetrics, the leading expert in API quality and efficiency, has made public the 2022 Cloud Performance Report, as stated by the company.
"APIs play a crucial role in our daily lives, and APImetrics is dedicated to assessing the quality of these APIs and the underlying clouds that power them," stated David O'Neill, APImetrics' CEO. "Many individuals fail to recognize the significance of the cloud in ensuring optimal API performance and functionality, resulting in wasted time, money, and customer dissatisfaction."
The blog post relies on the distinctive and constantly growing API dataset produced by the APImetrics platform to create an impartial, all-encompassing foundation for evaluating API quality. It encompasses information derived from over one billion API interactions with over 8,400 distinct API entry points spread across 70 geographically varied cloud data centers provided by AWS, Azure, Google, and IBM.
The article centers around information gathered from top API services, encompassing well-known corporate infrastructure providers, financial services firms, social media platforms, and search engines.
Relying solely on one provider in one location for a substantial portion of the world's cloud services poses a formidable threat to cybersecurity.
Marketing Technology News: Simpler Loyalty: Clever Methods to Enhance Customer Loyalty in the Grocery Industry via Technological Advancements.
"APImetrics has been conducting standardized evaluations on some of the most essential APIs worldwide for an extended period. Throughout this time, we have accumulated an extensive and unparalleled collection of data that provides valuable insights into not only the performance of these APIs but also the cloud platforms they depend on," expressed O'Neill.
One of the main discoveries is that AWS US-East (N. Virginia) remains the quickest cloud region for Time to Connect, maintaining its position for the fourth consecutive year. With an average time of 1.23 ms, it exemplifies the preference for hosting numerous solutions in this particular location.
O'Neill expressed concerns about the immense cybersecurity vulnerability associated with relying heavily on one provider and one location for hosting a substantial portion of the world's cloud services. He emphasized the need for enterprises to prioritize this issue. Despite Amazon's strong emphasis on performance, there have been several instances over the past few years where critical services like Auth0 and others experienced complete shutdown due to a single point of failure, impacting all services and APIs simultaneously.
"The improvement of edge internet performance has undergone significant transformations since our initial analysis," stated Paul Cray, the Chief of Machine Learning and Artificial Intelligence. "This poses an additional threat to inexperienced operations teams. Simply relying on the connection time for receiving a response from the internet can be highly deceptive. It is crucial for businesses to consider the overall transaction times to avoid making detrimental hosting decisions."
The initial findings from APImetrics in 2017 indicated a significant gap in median TCP Connect Time between South America and Europe, with South America being ten times slower. However, as of today, South America has managed to reduce this gap and is now 40% faster than Europe and more than three times faster than North America. Despite this improvement, South America is still approximately 50% slower in completing the entire transaction compared to North America.
According to O'Neill, there is a commonly held belief that the cloud is uniform and lacking diversity. However, this is not the case. The location and infrastructure on which an API is hosted can greatly affect both the operational efficiency and the experience of customers using the APIs. It is often observed that dissatisfied customers are actually unhappy due to decisions they have made in their own system design. This poses a challenge for API providers as they are unable to directly address issues that originate externally to the API itself, making it difficult to ensure customer satisfaction.
Cray stated that one significant advancement since 2020 has been the enhancement of Connect Time in various areas, particularly in traditionally considered 'peripheral' regions like Oceania and especially South America. In South America, the TCP Connect Time has reduced significantly from 113 ms in 2020 to just 3 ms in 2022. The reason behind this improvement is attributed to the implementation of enhanced infrastructure and increased hosting of local services. It is anticipated that this positive trend will persist. It will be intriguing to observe if South Asia can reach a similar level in 2023.
The article offers suggestions for achieving the best possible API performance. It highlights that if there is a delay of 100 milliseconds or longer (which is more than 10% of the average call time), it could be partly attributed to the cloud service chosen by the company. It emphasizes that all clouds are not equal and they evolve over time. Therefore, API users should not suffer from significant latency delays due to a decision made many years ago.
"It's simple to observe a primary figure like 100ms and deem it as fairly decent, but dismiss the consequences," O'Neill expressed. "When there is a procedure that involves ten pivotal stages and each one exceeds the expected time by 100ms, it accumulates to one second per transaction. If this occurs in 1,000 transactions within an hour, it equates to an extra six hours of computational resources required per day to manage this. Cloud resources have monetary value. Now contemplate this scenario on a larger scale, with 10,000 or even 10 million instances!"
APImetrics is dedicated to providing broader access to data, enabling individuals to comprehend their cloud vulnerability and the potential risks associated with their system structures. The API Directory, set to launch this month on the APImetrics website, marks the initial phase of making this information accessible to the public.
"We cannot deny the fact that the business environment has undergone significant changes," stated O'Neill. "APImetrics aims to provide individuals with genuine insights on how to enhance their cost structure and ensure that they are receiving a worthwhile return on their investment."
Marketing Technology News: An Interview with Denise Vu Broady, Chief Marketing Officer (CMO) at Collibra
