A fundamental topic in CloudWave is that of execution analytics which deals with the gathering, processing, and analysis of the vast amounts of monitoring data available in the Cloud. This data may originate from the underlying Cloud infrastructure, the service application running in the Cloud, and even the end devices that access and use this service. The data is also of many different types - from resources consumed by the service, all the way to patterns of user behavior. Execution analytics is about extracting a single holistic picture of the world from all this data.
What's new? So far, most analytics solutions have been focusing either on data from the Cloud infrastructure, or from the application. The purpose of the analytics also varies - Cloud infrastructure data is used for Cloud optimization, while application data is used for application optimization. Realizing this gap, several commercial players had recently started to offer solutions based on the combination of application and infrastructure monitoring. In CloudWave, we take this approach a step further by developing a unified open approach - one that considers data from all possible sources (the infrastructure, the application, and even the end-user devices), and extracts a more complete and informed picture.
Execution Analytics in CloudWave will be specially designed to handle the huge amounts of available data at every single point in time. It will do so by advanced methods, starting with preliminary filtering at the collection points, and all the way to applying in real-time different levels of abstraction to the data. The results of the analytics are designed such that they can then be used both by the engineers of the application, as well as by the CloudWave system itself for further development and performance optimization.
