In the Cloud environment, the dynamic task scheduling approach schedules the incoming tasks to the resources based on their current load to execute the task efficiently. An effective dynamic task scheduling approach aims to reduce the completion time, energy consumption and improves the utilization rate of the Cloud resources. In addition, it maintains the load in the Cloud resources effectually for minimizing the migration rate and performance degradation and improve the Quality of Service (QoS). Nowadays, the demand of the Cloud resources increases abruptly; however Cloud data center contains reliable resources, the Cloud providers struggle to manage the requirements of the users. To solve this constraint, the dynamic resource allocation technique is used in Cloud computing. It allocates the available resources to the users based on their requirements. An effective resource allocation minimizes resource wastage, cost, and SLA violations.
CloudSim, the simulation framework can be extended to experiment the innovations in both system and behavior modeling of cloud system components and environment.
Basically large scale cloud environment can be modeled and simulated by configuring the following components with the parameters
VmScheduler such as VmSchedulerSpaceShared, and VmSchedulerTimeShared can be extended to simulate user defined policies for allocating VM to hosts.
CloudletScheduler such as CloudletSchedulerSpaceShared, and CloudletSchedulerTimeShared can be extended to simulate user defined policies for allocating cloudlets to VM.
Solutions concerning with Network topology can be simulated using the class NetworkTopology and its methods such as buildNetworkTopology, mapNode.
Resource utilization is computed on the basis of CPU MIPS, storage and bandwidth using getAvailable methods. Based on the utilization, VM migration and task migration can be accomplished by extending the DataCenter class in which processCloudletSubmit() method is overridden with “VM_MIGRATE” tag and “CLOUDLET_MOVE” tag respectively.
Resource allocation is carried out in number of ways such as auction, game theory. It can be carried out by creating Auctioneer class and extending DataCenterBroker and DataCenter class.
Attacks in cloud computing environment such as Denial Of Service (DoS) attack can be modeled in cloudsim by submitting large number of cloudlets by the malicious user.
Optimization algorithms such as ACO algorithm, Genetic algorithm, and PSO algorithm can be applied for scheduling the task in VMs in large scale experiments.
Simulating storage system is required for deduplication techniques. HarddriveStorage class can be extended to simulate the solutions regarding Deduplication techniques.