Search

Edge computing is envisioned as a promising enabler to leverage computation capacities at the edge and address the issues faced in cloud computing. In this paper, we propose an advanced decision model to solve the computation offloading problem in edge computing. It utilizes the intrinsic hierarchical topology of the Internet and performs online scheduling in a decentralized manner to eliminate anticipated modeling. It also admits the fact that congestion can take place unexpectedly, for the possible reasons such as stale information. Our decision model is enhanced in two approaches to endow both sender and receiver (for one offloading action) with the ability of “dropping” the request, so as to avoid congestion. Random Early Detection(RED) algorithm is incorporated. The results of simulation demonstrate that our decision model can handle workloads well and these two enhancements are effective.

With the rapid development of cloud computing, more and more application providers are deploying their applications in the cloud in order to be free from the burden of system administration. Meanwhile, the world is full of information explosion, which makes some applications become hot in a short period of time unexpectedly. Thus these applications in cloud may encounter sudden traffic increment or decrement. Usually, developers use the alarm service or the auto scaling service provided by the cloud provider to tackle sudden traffic change. But these are reactive methods which have time latency and usually they only consider horizontal resizing of scaling group. In this paper, we propose PRMRAP: a proactive framework based on the prediction of resource amount to cope with sudden traffic change. Compared with post-action methods and existing proactive methods, we have lower time latency and we consider not only horizontal resizing but also vertical resizing of scaling group which makes our model quicker and much more cost-effective.

Because of the popularity of cloud computing, Cloud Service Providers (CSPs) can rent virtual machines (VMs) from Cloud Providers (CPs) conveniently. In our previous work, we proposed an autonomic and elastic resource scheduling framework, named AERS, which made full use of both proactive and reactive controllers in the field of dynamic resource provision and was integrated with an availability-aware and communication overhead optimized placement strategy. In this paper, we propose an enhanced version of AERS, named EAERS. It eliminates modeling the specific cloud application and instead determines the relationship between workloads and the number of virtual machines (VMs) through self-learning so that the whole scheduling can be carried out in a more transparent manner. Dynamic consolidation is also designed, implemented and integrated into EAERS. Experiments on OpenStack show that EAERS performs as expected.

The elasticity of cloud computing is able to help Cloud Application Providers (CAPs) adjust the number of rented virtual machines (VMs) for cloud applications according to actual varying demands while enforcing SLAs. In this paper, we design a framework called AERS (Autonomic and Elastic Resource Scheduling Framework). AERS makes full use of the profits of both proactive controllers and reactive controllers in the field of dynamic resource provision. It can not only adjust the number of available VMs for cloud applications with fluctuation of workloads but also react fast when cloud applications break SLAs. In addition, We build a model for the availability of cloud applications. Based on this model, we propose an availability-aware and communication overhead optimized placement strategy integrated in AERS to help CAPs choose proper availability zones for launching VMs. Experiments on Openstack show that AERS is able to provide VMs elastically while enforcing SLAs.