I work in a company that employs 'Quants' to develop statistical predictive models. One issue they often have is the time it takes to run optimisations of their data. I was thinking that this can be alleviated by providing them with a framework (They dev in c#) which they can extend and it will spread their computing across any registered machines.
So their are 2 parts to this
1 - A way of providing a framework which they can easily extend and will look after pushing their code to any registered machines.
2 - Having the node be unobtrusive to other applications running (A node will just be another users machine in this case)
I've been looking at implementations of the Actor pattern to achieve the first part and saw one that looks great (http://actorfx.codeplex.com/
). Unfortunately this seems to only run on Azure which doesn't fit with my idea.
So my question is whether there are similar solutions already available in .net, and if not what recommendations would you have for me to start with?