The option of cloud or on-premise are available for hosting a Jiwa deployment.

Leveraging platforms such as Microsoft Azure or Amazon Web Services Jiwa can be hosted to automatically scale to any number of users by using the Infrastructure-As-A-Service (IaaS) aspects of cloud platforms.

Use Microsoft SQL Azure for the database - Start from the S3 Performance tier (100 DTU’s) and scale up as needed.

If using Microsoft Azure, you have a choice of an RDS topology or Azure Virtual Desktops (AVD).

Azure Virtual Desktops are an alternative to RDS, but keep in mind if you also want services such as the REST API or Plugin Scheduler to be configured, then these should be on a dedicated virtual machine.

Please ensure your Azure SQL Server is in the same Region as where you plan to deploy your RDS or AVD.

This has the same requirements as the on premise RDS server - see Deployment Planning | RDS Server SpecificationsPreview

See also: Provisioning an Azure environment

If your cloud of choice is not Microsoft Azure, then your requirements are the same as the RDS topology for on premise - see Deployment Planning | On PremisePreview

There are two topologies possible - with a Remote Desktop Services (RDS) Server and without.

Both topologies require a dedicated SQL Server and each user requires their own PC (Workstation).

Management of more than a handful of users can become difficult without an RDS server - software updates need to be deployed to each individual users workstation, whereas with an RDS environment only one update on the RDS Server is required.

Jiwa is just a Windows desktop app, so your average PC will be sufficient - whether the user is going to connect via an RDS server or not.

As a general rule, Jiwa the application requires 512MB of RAM and one CPU core or VCore. Under provisioning the CPU cores is usually acceptable, as the Jiwa CPU usage is bursty - it does depend on the usage patterns of the users, but plan for 0.5 cores per user as a starting point, in addition to 2 cores for the operating system.