A Cost Effective Scalable Framework for Dynamic Threshold Based Autoscaling in Cloud

Venish Raja C
2020 International Journal of Advanced Trends in Computer Science and Engineering  
Cloud computing essentially has infinite resources from any single computing applications perspective, and supports those unlimited resources by on-demand scaling. An effective resource allocation strategy for client satisfaction and the maximization of profit for cloud providers is expected within the cloud paradigm.Previous resource allocation strategies are much focused on computation intensive task,distribution of task and not on the type and size of the task.Less focus is on data-intensive
more » ... tasks in which resource management approaches are not as effective in minimizing costs and lead to over-resource provisioning.In this approach,resource allocation is studied at the events raised in the application,number of task,size of the task and the number of threshold users.A new architecture has been proposed based on task defined and service oriented resource allocation.The proposed architectural framework uses the dynamic threshold based auto scaling mechanism in allocating the resources. It also aims to find the maximum threshold values for the task related metrics with minimum resources. In addition the proposed work mainly focuses on the reduction of cost and provide better efficiency with minimum resources used. The resource allocation is based on the events raised by the cloud users while using the cloud services. The experimental results indicates that the proposed auto scaling approach is better in terms of cost and obtain maximum throughput with minimum resources utilized.
doi:10.30534/ijatcse/2020/01942020 fatcat:ios7xnf35jf4bpppn3zjlfiwcm