The dynamic nature of resource allocation and runtime conditions on Cloud can result in high variability in a job's runtime across multiple iterations, leading to a poor experience. Identifying the sources of such variation and being able to predict and adjust for them is crucial to cloud service providers to design reliable data processing pipelines, provision and allocate resources, adjust pricing services, meet SLOs and debug performance hazards. In this paper, we analyze the runtime variation of millions of production Scope jobs on Cosmos, an exabyte-scale internal analytics platform at Microsoft. We propose an innovative 2-step approach to predict job runtime distribution by characterizing typical distribution shapes combined with a classification model with an average accuracy of >96%, using an innovative interpretable machine-learning algorithm out-performing traditional regression models and better capturing long tails. We examine factors such as job plan characteristics and inputs, resource allocation, physical cluster heterogeneity and utilization, and scheduling policies. To the best of our knowledge, this is the first study on predicting categories of runtime distributions for enterprise analytics workloads at scale. Furthermore, we examine how our methods can be used to analyze what-if scenarios, focusing on the impact of resource allocation, scheduling, and physical cluster provisioning decisions on a job's runtime consistency and predictability.

R is one of the most popular languages in the data science, statistical and machine learning (ML) community. However, when it comes to scalable data analysis and ML using R, many data scientists are blocked or hindered by (a) its limitations of available functions to handle large datasets efficiently, and (b) knowledge about the appropriate computing environments to scale R scripts from desktop exploratory analysis to elastic and distributed cloud services. In this tutorial we will discuss solutions that demonstrate the use of distributed compute environments and end to end solutions for R. We will present the topics through presentations and worked-out examples with sample code. In addition, we will provide a public code repository that attendees will be able to access and adapt to their own practice. We believe this tutorial will be of strong interest to a large and growing community of data scientists and developers using R for data analysis and modeling.