Information Technology Reference
In-Depth Information
cooling mode is applied, and the value represents the fan speed as a percentage of
its maximum; (b) 101-200 : the air conditioning unit is assumed to be operating,
and (c) 201-300 : neither air conditioning nor free cooling are in operation.
Evolutionary operators : We apply a three-point crossover (using cross points
p 1 , p 2 ,and p 3 ); p 1 is selected randomly in (1, K ), p 2 is K and p 3 is K + p 1 .This
approach assures that portions representing the same time interval for both
cooling and server power move together from parents to offspring.
Mutation is applied to each gene with probability p M .Foracoolingpower
gene (position 1 to K ) its value v is replaced with mod ( v + rand ()
MAX HVAC,
MAX HVAC)). For the other genes, they are redefined with a random value
between 0 and the maximum server power, i.e. representing all servers on.
×
4.3 Energy and QoS-aware Scheduling
In order to model a realistic task planning in the simulated datacenter, we apply
three heuristic energy-aware QoS schedulers. They are based on the ones defined
from our previous work [6], but adapted to deal with the specific features of the
problem addressed in this article.
The schedulers apply different backfilling-oriented techniques to work with
computing resources that are available in certain periods of time (we call these
periods slack-times or simply holes ) and unavailable in other moments (e.g. due
to sleeping and/or switching off servers). The heuristics differ in the way they
fill holes/slack-times that are left after a given policy for sleeping/shutting down
holes is applied to reduce the energy consumption:
1. Best Fit Hole (BFH): Tasks are first sorted according to their arrival times,
and then assigned to computing resources to fill their existing holes/slack-
times. If a task fits into more than one hole, the one that “best fits” the task
(i.e., the hole that minimizes the difference between the hole duration and
task execution time) is selected. Holes within each machine are processed
according to their finishing times. A specific logic is included to deal with
deferrable tasks. When no hole is available to execute a task, BFH assigns it
to the machine that provides the minimum finishing time for that task. The
rationale behind this strategy is to use available holes and spare unoccupied
large holes and empty machines for upcoming tasks with potential larger
execution times.
2. Best Deadline (BD): This scheduler applies a greedy approach to select the
slack to execute each incoming task, improving the QoS of the resulting
schedule. As in BFH, a specific logic is included to deal with deferrable
tasks, which can be scheduled in any available hole within the simulation
period. When no hole is available to execute a task, BD also assigns it to
the machine that provides the minimum finishing time for that task.
3. Earliest Finishing Time Hole (EFTH): In this strategy, holes/slack-times are
selected to minimize the tasks' finishing times. That is, instead of finding the
hole that best fits a given task (as in BFH), a hole that can finish it earlier
is selected regardless of its length. As a result, EFTH should lead to fewer
 
Search WWH ::




Custom Search