Mohammad Fairus Khalid

A survey on SLA-based brokering for inter-cloud computing

Service Level Agreement (SLA), a contract between IT service provider and customer, clearly defines hardware and software parameters to ensure Quality of Service (QoS). Cloud Service Brokering (CSB) provides either API or User Interface (UI) between service subscriber and service provider which enable interoperability between several separate clouds to improve service intermediation, aggregation and arbitrage. Today, SLA-based brokering technology is emerging and has been used to ensure QoS for Inter-Cloud environment. In this paper, we surveyed and evaluated some of the latest SLA-based service brokering projects. Additionally, we discussed a high level architectural view of cloud service broker derived from these projects.

Architecture of scalable backup service for private cloud

In this paper we propose a scalable backup service for virtualization infrastructure such as private cloud called hBaaS. hBaaS is designed to utilize existing infrastructure without the needs of additional equipment to provision it as-a-Service model. We propose an integrated solution for virtualization platform that scales as the total VMs increases. The solution addresses the issue of concentrated backup schedule on a specific time range which potentially cause resource bottleneck. We discuss two methods to reduce the backup workload footprint; by distributing the backup workload operation into multiple proxy servers and push the backup operations into background which increases the duration of time to backup but lessen the impact of backup operation to the infrastructure. Indirectly, hBaaS will be able to cater more backup requests from users.

Determining Overhead, Variance & Isolation Metrics in Virtualization for IaaS Cloud

Infrastructure as a Service cloud provides an open avenue for consumer to easily own computing and storage infrastructure. User can invent new services and have it deploy in a fraction of a time. As services being consumed, user expect consistent performance of their virtual machines. To host an IaaS, virtualization performance and behaviour need to be understood first. In this paper we present virtualization benchmarking methods, results and analysis in various resource utilization scenarios. The scenarios are Performance Overhead; the cost of virtualization compared to physical, Performance Variance; the VM instance performance when more VM is instantiate and Performance Fairness or Isolation; the study of fairness on each of the guest instances under stress environment. These studies provides us the fundamental understanding of virtualization technology.

Policy management for Docker ecosystem

Edge Computing is a form of geographically and dispersed distributed system. It shifts computing resources from a centralized paradigm i.e. Cloud or data center closer to the target objects e.g. data, user or sensors. By de-emphasizing centralization, it reduces communication bottlenecks and potential failure. Due to the nature of Edge Computing architecture challenges, resource management is complex. Policy management can automate and alleviate the management process that indirectly reduces the intervention of system administrators. A policy can be applied to multiple target edge devices to ensure its behavior uniformity. In this paper, we present our studies of the policy management system for managing edge resources. We present Mi-Focus Policy, which consist of Policy Hub for managing policies and Changkuk, a local Enforcement Engine which enforces policies in order to obtain a self-desired state required by the administrator.