A Relay Server deployment consists of the following:

• Mobile devices running client applications and services that need to communicate with back-end servers running in a corporate LAN.
• Optional load balancer to direct requests from the mobile devices to a group of Relay Servers.
• One or more Relay Servers running in the corporate DMZ.
• Back-end servers running in a corporate LAN that are responsible for servicing client requests.
• One Relay Server Outbound Enabler (RSOE) per back-end server. The Outbound Enabler manages all communication between a back-end server and the Relay Server farm.

The following diagram shows the Relay Server architecture.

The Relay Server consists of a set of web extensions, a background process for maintaining state information, and a web server.

Because the Relay Server is a web extension running in a web server, all communication is performed using HTTP or HTTPS. Using HTTP easily integrates with existing corporate firewall configurations and policies. The Relay Server requires that the connection from the corporate LAN to the Relay Server be initiated from inside the corporate LAN. This provides a more secure deployment environment because it does not require inbound connections from the DMZ into the corporate LAN.

The Relay Server contains two web extensions: a client extension and a server extension. The client extension handles client requests made from applications running on mobile devices. The server extension handles requests made by the Outbound Enabler on behalf of a back-end server.