Testing at Scale

  • Comments posted to this topic are about the item Testing at Scale

  • Wow, a tool like that is long overdue. By benchmarking your dev environment with production you could also "trend" how your development releases should do in the production environment. I'm sure that most development shops dev environments are nothing like the production hardware.

    Question though, from SQL viewpoint how would this be different than capturing a trace and then playing it back? Then the next question would be how to systematically increase (scale up) the work load of the captured trace (with realistic database load). That is probably the hard part of the solution.

    Mike Byrd

  • Steve:

    I am surprised that a tool like that is not already available. I used to do Capacity Planning when I worked for a vendor back in the late 80's and early 90's and we had tools like that to help us justify hardware configurations for our customers. We could show growth of transactions based upon changes to CPU, disk, memory.

  • That would be really handy for replacing end-of-life hardware as well. Gather query info on the old system, replay it on the new box and see what shakes loose. Testing the hit on the new system would also give a bit of insight on what tweaks (partitioning, tempdb, file placement, etc.) would impact the new box as well.

    Where do I sign up to be a beta tester? I'm in!

    <lol>

  • The hard part about profile is scaling it up. If you double or triple up the trace, you need other clients. I'd think that a tool like this would be easier to just send a load to the CPU/Disk/Memory easier from a single machine.

  • Maybe this will show my lack of knowledge about VM, but shouldn't you be able to spin up a VM with the production Config, and test against that? I guess that assumes you have a machine that exceeds the specs of the production machine, to VM against.

Viewing 6 posts - 1 through 5 (of 5 total)

You must be logged in to reply to this topic. Login to reply