In order to be able to scale horizontally and vertically with no limits all the server platform infrastructure had been splitted in several boxes one of them dedicated to a particular task and also independent of the rest so it can be scalled independently of the rest ones.
The names of each box follows this scheme: WU-<type_of_client>
The WU-Global-Server server is the responsible to intermediate between the central Notification Server infrastructure (3rd party) and each WU-Local-Server deployed in each OB.
This server exposes a public API through Internet which only trusted partners can consume (see Security chapter).
The WU-Local-Server server is a proxy between the global WU servers and the client equipment (device). This service will receive petitions through a standard HTTP port and will send UDP datagrams or TCP packets (for pinging purposes) inside the OB private network to the private IP of the client equipment. This server must be placed inside the OB private network or in a zone that must see that private IPs.
This server is responsible to ping to the correct client inside each OB (using UDP datagrams). It must be placed inside the OB or in a zone that can see the devices inside that private network.
This server will receive the ping orders through a standard HTTP port which will be connected to InterNodo network to receive the data from the VDC inside operator network.
The following scheme shows who the notification is sent with a wakeup server: