Hi,
I'm developing a NT Service using .NET 2.0, its a multithreaded TCP
server, this service will handle connections from many enterprise
applications.
The challenge is to avoid a single-point-of-failure. My first idea was
create a service instance capable of just redirect incoming calls to
service instances which implement the concrete service.
This solution decrease service unavailability risk, but AFAIK real
cluster services like App Center uses another approach, it creates a
virtual server that can redirect incoming calls to an available server,
this approach really solve the single-point-of-failure problem.
I'd like to use the same solution, how it can be done ?


Thanks for any help !