My company’s needs are quite simple: we have a multi-threaded .Net computer program that reads many Gb binary files, processes massive calculations, and stores the results in a SQL Server database. We would like to do this in the cloud to complete this repetitive task as soon as possible.
So, we are right in the cloud / grid / cluster operation.
I thought that there would be a ton of resources on this and a lot of alternatives available. I was just stunned to see how wrong I was. While installing / starting EC2 instances was easy, finding a relatively simple and easy way to parallelize and aggregate the processing power of these EC2 instances was not easy. Amazon customer service is holding on, and I just couldn't get a specific answer from them.
I found utilify that sounds promising. It is designed by alchemi people. However, the documentation link was broken, and I didn’t have an answer to my letters when I contacted support, so it was not very encouraging.
We chose Amazon over Azure, since AMI is a simple virtual virtual machine (no need to “bind” an application or another), and EBS is a more convenient storage, since it is a “real” file system. Azure, on the other hand, seems HPC ready for windows, while AWS only offers Linux-powered AMIs.
Any help and suggestions are more than welcome.
EDIT:
A .Net application is multi-threaded and consists of hundreds of parallel workers performing exactly the same task asynchronously.
source
share