At first let's discuss what features differ DCOM server in service from the ordinary one. In brief these features are the same as features of service application in comparison with an ordinary executable. At first it is a remote controlling and monitoring capability. There is a set of Windows API functions that are used for this purpose. The service can be started, stopped and paused and all these operations can be done over network even if no interactive user is logged on. The service can start automatically when the system starts operating. And, in addition, service can work in the Local System account. So if the DCOM server is implemented as a service then it should have the same properties. Consider for example the pausing capability. It allows to implement a server that continues to work with existing connections after pausing but declines any new connection requests. After disconnection of all clients such server can be stopped safely protecting clients from data loss.

Another feature is a possibility to implement a multi-service application. All services of this application can be controlled independently and moreover they can run in contexts of different users. At the same time all this services will use the same address space. The access and launch rights can also be configured independently.

Below is a table that summarize the common features and differences between ordinary DCOM exe-server and DCOM server in a service.

Feature	Ordinary DCOM server	DCOM server in a service
Automatic start due to client request
Starting in given user context
Starting in local system context
Starting as interactive user
Access/launch rights configuration
Remote starting	(due to client request only)
Remote stopping
Pausing
Several independent parts in one application
Independent profile configuration of each part
Independent controlling of each part

The only necessary thing to use the specific features is to implement the DCOM server in service. Let's try to do that...