所有标记为AWS的博客条目//www.billcordell.com/blog/tags/134-aws 2022年6月17日星期五11:50:10 +0000 en - //www.billcordell.com/blog/1950-how-pricey-are-the-hyperscalers-internet-of-things-offerings //www.billcordell.com/blog/1950-how-pricey-are-the-hyperscalers-internet-of-things-offerings 首席研究员:数字经济部门研究总监Owen Rogers和物联网451 Research研究总监Christian Renaud最近发表了《Technology &188宝金博网址是多少Business Insight: The Economics of IoT报告，该报告比较了AWS、谷歌和微软这三家领先的超大规模企业的云物联网产品的定价模式和成本。451 Research Voice of The Enterprise IoT调查受访者将这三家大型公共云提供商视为领先的物联网平台供应商。188宝金博网址是多少要发现这些结果并非易事。AWS物联网核心、谷歌物联网核心和微软Azure物联网中心的成本结构如此不同，以至于人工比较成本几乎是不可能的。即使在定价模型看似简单的地方，细微的差异也会产生重大的成本影响。例如，影响成本的一个隐藏元素是，每个提供者对构成消息的内容都有自己的定义。一些物联网平台对只有几个字节的“保持活跃”消息不收费，而其他平台则要收费。事实上，一些云提供商将这样的小消息集中到最近的千字节。这将导致64字节的“ping”消息被收取17倍的容量。为了解释这些差异并有效地完成分析，分析师仔细确定了9个影响超大规模云物联网平台成本的定价参数。这些参数仍然导致了数以百万计的排列，其中一家提供商比另一家更具成本效益。 The only approach was to price all possible combinations and understand the differences using machine-learning techniques.The analysts constructed a Python simulation that performed 10 million comparisons of US pricing. Each simulated scenario had a randomly selected configuration of the nine price-impacting parameters. Resulting data was then fed into an analytics platform that produced a Chi-square automatic interaction detection (CHAID) decision tree. The tree revealed which combinations of the parameters drove which provider (AWS, Google or Microsoft) to be the cheapest with a predictive strength of 96%. Overall, the 451 Research Digital Economics Unit's methodology deduced that Microsoft Azure is the cheapest at scale, but AWS is the least expensive IoT provider for deployments of less than 20,000 devices that each send an average of three messages or fewer at under 6KB a minute. Google was not found to be exclusively the cheapest in any scenario.To learn more about what factors drive the cloud-related costs of Internet of Things deployments, and in which scenarios each cloud provider has a cost advantage, register below for our webinar on June 13 at 1:00 pm ET. admin@www.billcordell.com(卡桑德拉·罗) 研究与数据 2018年6月1日星期五18:30:43 +0000