Most companies today are faced with silos of data that might be located in legacy on premise systems, modern software as a service providers and/or born in the cloud applications. Given the location of the data, how can we mash up and model the data to define metrics that are import to the business users? Azure Analysis Services is a platform as a service offering that eliminates the need for managing more hardware and software. The trusted tabular semantic data model can be used to compress large amounts of data into a small memory model. We can use row level security and active directory to eliminate the worry of users viewing the wrong information. In this presentation, we will be reviewing how to create and deploy a model using Visual Studio 2017. The various maintenance tasks such as backup, restore, firewalls and security will be reviewed. When a model is refreshed or processed, data can come from the azure cloud or on premise. Knowing which data can be stored in memory or must be directly queried is key to understanding performance. Advanced topics such as on premise data gateways, redundant servers and scaling out will be covered.