What are the simulated system's properties while the simulating workflows on cloud ennvironment (input parameters and variables), e.g., how the simulated datacentre is to be set up and the networking capabilities of the centralised storage the individual VMs, discussion of the monetary costs of data transfers (in data-centre, between centralised storage and VMs), discussion about how the simulated datacenter descriptions was loaded during the simulations.
Looking for detailed description.
I would enumerate some:
resources characterization, at least ram and cpu capabilities
number of resources
interarrival time between jobs (generating an exponential distribution from that)
tasks per job
resource requirements for jobs/tasks
distribution between batch and service jobs
scheduler approach (monolithic, shared-state, two level, omega, mesos, yarn)
scheduler think time
a lot more...
- Badges
- Science method