Load balancing/clustering and DSC/QMS/Reload Engine

[northerner]

I'm working at a site where the Qlikview infrastructure has been built over time.  There are two servers that host a number of QVWs (different files on each machine).  There is no clustering or load balancing on these servers and so the situation is as follows:

Server 1: QVS, QWS, DSC, QMS and Reload Engine

Server 2: QVS, QWS, DSC, QMS and Reload Engine

I want to create a resilient environment and I'm planning to do this as follows:
a) Load balance the two QVS instances from QMS.
b) Add a third-party load balancer to route requests to either of the two QWS instances.

I'm not as certain about what needs to be done with the other components. 
1) Will I need a DSC service running on each machine?  (I believe that QWS needs this?)
2) I've read that only one instance of QMS and the Reload Engine should be running at any one time.  In this case do I need a third party clustering mechanism to ensure that a QMS/Reload Engine is started on server 2 if there is a failure of the QMS/Reload Engine on server 1?

I'm waiting to obtain a server LEF that allows clustering, so at the moment I can't experiment with the set-up.  I'd appreciate any input that anyone may have.

[northerner]

In case this is useful for anyone else, I've found some information that answers my questions:

1) Section 26.2 of the Qlikview Server Reference Manual has instructions for how to cluster the DSC service, which should make this service resilient.

2) Section 26.1 deals with how to cluster the QDS (i.e. ReloadEngine), so I won't need a 3rd party clustering mechanism for this.  However ,it looks like I will something like Microsoft Cluster Server to provide resilience for the QMS.