We are planning to move towards scrum based approach for software development. However the thing that is still unclear is how can the testing process be adopted in the Scrum approach. It seems to be easier to break features into tasks based on dev resources available and start and end of a Sprint. Would QA typically start after the first dev Sprint 1 is completed? If so, the challenge would the testing volume may not same always but they start and end date for QA would have to line up with Sprint 2 of dev cycle and in this case the number of QA resources would never be same in each Sprint. Hope I am making sense. The other alternative we see is to have QA process be running in parallel with Sprints and as soon as a feature is completed, they take it and test it. But they will never be having any Sprints.