Title | ||
---|---|---|
Mapping of time-consuming multitask applications on a cloud system by multiobjective Differential Evolution |
Abstract | ||
---|---|---|
The paper deal with the mapping problem of multitask parallel applications on cloud systems.Specific reference is made to task interaction graph applications.A new form of service contract is hypothesized.A tool based on a multiobjective Differential Evolution algorithm is proposed to deal with the problem.The results demonstrate the effectiveness of the proposed evolutionary mapper. Cloud computing is on-demand provisioning of virtual resources aggregated together so that by specific contracts users can lease access to their combined power.Here we hypothesize a new form of service contract by means of which users do not explicitly require resources, but simply supply information about their time-consuming multitask applications and specify their needs through some quality of service (QoS) parameters. The individuation of the virtual machines (VMs) onto which map and execute them is left to the cloud manager. Unfortunately the task/node mapping, already known as NP-hard for conventional parallel systems, becomes more challenging when application tasks must be run on VMs hosted on heterogeneous and shared cloud nodes, and when it must comply with QoS requests too. To support this new cloud service, a novel mapper tool, based on a multiobjective Differential Evolution algorithm, is proposed. Such a tool defines the mapping of the tasks on the VMs with the aim to exploit as much as possible the available cloud resources without penalizing the execution time of the submitted applications and, at the same time, to respect users' QoS requests.To reveal the robustness of this evolutionary tool, an experimental analysis on artificial time-consuming parallel applications, modeled as task interaction graphs, has been effected. |
Year | DOI | Venue |
---|---|---|
2015 | 10.1016/j.parco.2015.04.001 | Parallel Computing |
Keywords | Field | DocType |
Cloud computing,Mapping,Differential Evolution,Multiobjective problems | Time-Consuming,Virtual machine,Computer science,Parallel computing,Quality of service,Differential evolution,Robustness (computer science),Exploit,Provisioning,Cloud computing,Distributed computing | Journal |
Volume | Issue | ISSN |
48 | C | 0167-8191 |
Citations | PageRank | References |
1 | 0.35 | 69 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Ivanoe De Falco | 1 | 242 | 34.58 |
Umberto Scafuri | 2 | 116 | 16.33 |
Ernesto Tarantino | 3 | 361 | 42.45 |