QoS lake: Challenges, design and technologies

QoS evaluation based on their historical data not only helps in getting more accurate QoS, but also helps in making future QoS prediction, recommendation and knowledge discovery. [1] designed a generic QaaS (Quality as a service) model in the same line as PaaS and SaaS, where users can provide QoS attributes as inputs and the model returns services satisfying the user's QoS. It uses historical data to evaluate accurate QoS. Storing and evaluating QoS based on historical data and managing QoS for all services on the internet is challenging. This paper proposed a QoS lake in the same line of Data Lake for implementing QaaS model using big data technologies like Hadoop, Spark, and Yarn etc. The QoS Lake is a very large repository that stores all logs generated from services and its evaluated QoS data in its original context for all services on internet. The log data are processed to evaluate QoS either in batch or real time. QoS Lake is integrated with cutting-edge analytics, automation, orchestration and machine intelligence tools and languages which are used for future prediction, recommendation and knowledge discovery. QoS Lake has four loosely coupled layers namely; Ingestion layer, data layer, Analysis layer and Visualization layer. The challenges and advantages of the data lake are also discussed. The paper also presented the technologies available today to realize each layer and functionalities of the QoS Lake.