Information Technology Reference
In-Depth Information
units called package-based scheduling and task level scheduling which allocates tasks
to the virtual machine of data center dynamically. Service-level scheduling should
satisfy the QoS constrain of each task and the dependency between tasks. The task-
level scheduling research genetic algorithm, ant colony algorithm and particle swarm
algorithm under QoS of each task and the total cost reduced. In brief, this strategy
selects suitable cloud provider and topics the service resource; allocates and optimizes
tasks to virtual computing resource using intellectual algorithms in task-level.
Concurrent Level Based Workflow Scheduling Algorithm
Due to Deadline Bottom Level (DBL) hasn't considered the concurrence during the
real executing process that cause much more shatter time. In order to solve such prob-
lem, Guangzhen Lu [13] et al. proposed a novel heuristic workflow scheduling algo-
rithm CLWS, which it distributes task levels by their concurrence, and adopts the
efficiency algorithm MDP to optimize the sequential tasks with time dependency. It
not only can decrease the time pieces, but also can optimize the total executing cost.
The experiments demonstrate that CLWS has better performance than DBL and
Deadline Min-Cost.
Adaptive Workflow Scheduling
Mustafizur Rahman [16] et al. developed a hybrid heuristic that can effectively inte-
grate most of the benefits of both heuristic and metaheuristic-based approaches to
optimize execution cost and time as well as meet the user's requirements through an
adaptive fashion. They proposed Adaptive Hybrid Heuristic scheduling algorithm,
which is designed to first generate a task-to-service mapping with minimum execution
cost using GA with user's budget and deadline as well as satisfying the service and
data placement constraints specified by the user.
Table 1.
The typical of clouding workflow algrithms
Scheduling
Algorithm
Scheduling
Method
Scheduling
Parameters
Findings
Environment
Tools
Cloud
workflow
scheduling
algorithm
oriented to
dynamic price
changes
It decrease 5%
cost compared
to state space
algorithm with
considering the
changing price
Batch
/dependency
mode
Execution
cost, Dead-
line time
Cloud Envi-
ronment
Java
The algorithm
decrease the
cost compared
the built-in
algorithm and
has better
performance
Instance-
intensive cost-
constrained
workflows
scheduling
algorithm
Batch
/dependency
mode
Resource
Utilization,
Total cost
Cloud Envi-
ronment
Cloud-
Sim
