A Security-Driven Approach for Energy-Aware Cloud Resource Pricing and Allocation
Auctions are often recommended as effective cloud resource pricing and allocation mechanism. If adequately set, auctions provide incentives for cloud users’ truthful bidding and support cloud provider’s revenue maximization. In such a cloud system, resources are offered via an auction mechanism as...
Guardado en:
| Autores principales: | , |
|---|---|
| Formato: | article |
| Lenguaje: | EN |
| Publicado: |
Stefan cel Mare University of Suceava
2021
|
| Materias: | |
| Acceso en línea: | https://doaj.org/article/e5d2118454834388bc087d5194370ab6 |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| Sumario: | Auctions are often recommended as effective cloud resource pricing and allocation mechanism. If adequately set,
auctions provide incentives for cloud users’ truthful bidding and support cloud provider’s revenue maximization.
In such a cloud system, resources are offered via an auction mechanism as Virtual Machines (VMs). Due to the
virtualization of the cloud system, VMs’ security becomes a critical factor. However, security requirements
are often in contrast with performance requirements since the operation of security mechanism inevitably consumes
a certain amount of Central Processing Time (CPU) and memory. Thus, delays and energy consumption increase.
In this paper, we propose a novel simulation model based on a truthful auction mechanism to address revenues,
security, and energy consumption in a cloud system. The VMs security modeling is introduced to assess the security
level of VMs. A Vickrey-Clarke-Groves (VCG) driven algorithm is established for winner determination.
The proposed simulation model is used to observe cloud provider’s revenues, lost revenues, cloud users'
task rejection rate and energy consumption depending on the offered security level. This model supports
decision making in terms of investments in security and selection of security scenario that maximizes
revenues and minimizes task rejection rate and energy consumption. |
|---|