Self-service BI works well if users only explore data and use it on their local PC. When they want to share their discoveries, they need a portal for it. This is where the Power BI service enters the scene. Users publish their Power BI models on the cloud and they can schedule data refresh if the data is available on the cloud. For imported data, there's a limit on the size of the model that can be published on the cloud. Currently, this limit is set to 1 gigabyte of compressed data. Power BI models try to compress the data by column to enhance query performance.

For on-premise data sources, data management gateways must be used. These gateways are like the self-hosted integration runtimes we used in previous chapters. They allow secure access to on-premise data from the cloud. We'll talk about data management later in this chapter.

There are two different versions of Power BI available on the cloud: personal and professional. The personal version allows publication of dashboards and reports to the cloud, but they cannot be shared. This can be seen as an extension of personal BI. We can show what we did but cannot share it. The professional version allows us to share our Power BI dashboard and reports with other users in our organization.

More information about the features of Power BI on the cloud can be found at this URL: https://powerbi.microsoft.com/en-us/pricing/.