Hoi Chan

Ranking and mapping of applications to cloud computing services by SVD

Cloud computing promises to provide high performance, on-demand services in a flexible and affordable manner, it offers the benefits of fast and easy deployment, scalability and service oriented architecture. It promises substantial cost reduction together with increased flexibility than the traditional IT operation. Cloud service providers typically come with various levels of services and performance characteristics. In addition, there are different types of user applications with specific requirements such as availability, security and computational power. Currently, there are no standard ranking and classification services for the users to select the appropriate providers to fit their application requirements. Determining the best cloud computing service for a specific application is a challenge and often determines the success of the underlying business of the service consumers. In this paper, we propose a set of cloud computing specific performance and quality of service (QoS) attributes, an information collection mechanism and the analytic algorithm based on Singular Value Decomposition Technique (SVD) to determine the best service provider for a user application with a specific set of requirements. This technique provides an automatic best-fit procedure which does not require a formal knowledge model.

Dynamic Resource Allocation via Distributed Decisions in Cloud Environment

The adaptation of virtualization technologies and the Cloud Compute model by Web service providers is accelerating. These technologies commonly known as Cloud Compute Model are built upon an efficient and reliable dynamic resource allocation system. Maintaining sufficient resources to meet peak workloads while minimizing cost determines to a large extend the profitability of a Cloud service provider. Traditional centralized approach of resource provisioning with global optimization and statistical strategies can be complex, difficult to scale, computational intensive and often non-traceable which adds to the cost and efficiency of Cloud operation, especially in industrial environments. As we have learned in real life, the most efficient economic system is the one that provides individuals with incentives for their own decisions. It is also true for computing systems. In this paper, we present an architecture for dynamic resource provisioning via distributed decisions. We will illustrate our approach with a Cloud based scenario, in which each physical resource makes its own utilization decision based on its own current system and workload characteristics, and a light-weight provisioning optimizer with a replaceable routing algorithm for resource provisioning and scaling. This approach enables resource provisioning system to be more scalable, reliable, traceable, and simple to manage. In an industrial setting, the importance of these characteristics often exceeds the goal of squeezing the absolute last CPU cycles of the underlying physical resources.

Autonomic Ranking and Selection of Web Services by Using Single Value Decomposition Technique

With the proliferation of Web services as a business solution to enterprise application integration, ranking and selecting the best Web services among the providers become an important factor in the success of the business solution. Quality of service (QoS) determines the quality and usability of a service including its price, performance, reliability, integrity, accessibility, availability, interoperability, and security. Given a set of QoS attributes from a variety of sources, it is a challenge to sort through all of them and be able to get the best services that meet QoS requirement. In this paper, we describe a novel method by which Web services can be ranked and selected automatically based on a number of observed QoS parameters and feedback responses learned from prior knowledge. This new approach treats the observed Web services QoS attributes and target Web services relationship, represented by a matrix, as a statistical problem. Using singular value decomposition (SVD) technique, and an user assisted weighting system, implicit higher order correlations among Web services and their associated QoS attributes are extracted and used to estimate the selection of recommended Web services.

An approach to monitor application states for self-managing (autonomic) systems

Autonomic Computing has gained widespread attention over the last few years for its vision of developing applications with autonomic or self-managing behaviors[1]. One of the most important aspects of building autonomic systems is the ability to monitor applications and generate corrective actions should exceptions occur. The problem lies in those applications where source code is not available and therefore it is virtually impossible to modify the application code to include monitoring functions, or the application code is too tangled with other components which make modification difficult. This hinders the inclusion of autonomic features in many of the legacy applications. In this report, we will describe an approach to build generic monitoring systems for legacy applications.

An approach to high availability for cloud servers with snapshot mechanism

Virtualization technologies enable the execution of multiple virtual machine instances (VMs) with different operating systems (OSs) on the same physical host. Each VM instance functions independently as an isolated system with its own physical resources, OS and applications. Due to significant cost saving and efficiency, the virtualization model has been increasingly adapted by enterprises and service providers as their main computing and service delivery infrastructure, running critical internal business and external customer facing applications. To minimize down time due to unexpected VM crashes, a high availability or backup system is usually built into the infrastructure. There are many high availability technology options available such as replication, mirroring and fail over clustering. Most of these solutions are usually designed based on the traditional computing model, they are costly to implement, complicated and tedious to maintain, especially in a virtualized environment, and they often require additional expensive hardware and software components. In this paper, we introduce a simple, flexible, scalable, extensible, efficient and cost effective system which utilizes the current VM infrastructure and common utilities to provide a high availability solution in the virtualization environment. Our smart adaptive snapshot replication technique provides a smooth and reliable mechanism for cost-performance, wherein the amount of resources allocated for high availability solution can be adjusted based on available resources, utilization and customer requirements.

Virtual machines with sharable operating system

Virtualization technologies commonly known as Cloud model enable the execution of multiple virtual machine instances (VMs) with different operating systems (OSs) on the same physical host. Each VM instance functions independently as an isolated system with its own physical resources, OS copy and applications. There is only a limited number of currently available and widely used OSs used by most of the running VM instances; it is wasteful to store all the VM images with virtually the same common OS code. It is also inefficient in terms of performance and system resources utilization to virtually clone the entire image each time a new VM instance is provisioned. In addition, performing OS updates and patches are complicated, tedious and error prone since not only the stored images need to be updated, all the running VM instances must be properly refreshed. More importantly, faster provisioning of VM instances in respond to workload changes is critical to the successful operation of Cloud service providers. In this paper, we show our exploration work to address these performance issues by using a common, sharable operating system approach which provides run-time on-demand operating system components to individual VM instances in Cloud environment. This new approach allows optimized VM image storage, faster VM provisioning and efficient OS updates with minimum interruption.

A policy based system to incorporate self-managing behaviors in applications

With the rapid increase in complexity of software systems and applications, it becomes necessary to develop tools to simplify the incorporation of self-managing features into applications. The use of object and component technologies, together with a policy system which externalizes business logic from an application, plays an important role in enabling systems with greater manageability and variability. In this presentation, we will show an approach to build policy based applications with much greater flexibility, expressiveness and reusability . The key concept in this approach is the separation of business logic expressed as policy or rules into its various sub-components and these sub-components are dynamically configurable.

A robot as mobile sensor and agent in data center energy management

In this poster/software demonstration we illustrate the integration of an autonomous mobile robot into a slightly customized version of a commercially available asset and data center energy management application known as Maximo Asset Management for Energy Optimization (MEO), version 7.1.1, through a number of practical scenarios. The scenarios showcase increasing degrees of autonomy and sophistication in the areas of data center mapping, monitoring and thermally-aware diagnostics.

Aspect-Oriented Web Services for Distributed Resource Monitoring in Utility Computing

Monitoring resource utilization is an essential task in utility computing (UC). Typically, a UC manager with adaptors is used to orchestrate and collect the monitoring results. However, large IT systems, such as those in data centers are required to monitor dynamic resources running in distributed platforms, thus demanding a complex UC manager structure. This paper introduces a novel approach that utilizes a platform specific aspect-oriented programming (AOP) tool to dynamically weave a monitoring Web service into a running resource to enable communication with a UC manager exposed with standard monitoring Web service interfaces. This AO-Web service approach provides a simple yet powerful and effective means for the dynamic monitoring of distributed resources running in heterogeneous platforms.

Sensor composer: composing intelligent virtual sensors for collaborative sensing

Sensor networks have been gaining popularity with the proliferation of internet and wireless networks. Millions of sensors (physical, Webs and social Webs) are available in sensor networks via the internet around the globe collecting enormous amount of data. Generally, sensors provide information about various aspects of the physical world. They can provide information about weather, traffic conditions, efficiency of data center operation, enable autonomic management of smart buildings, and are being used in numerous mobile applications. Due to the diversity and the number of available sensors, it is practically impossible for the general users to manually define the set of sensors that best meets their application needs. The ability to select automatically and intelligently the best set of sensors to form a high level reusable virtual sensor will create a new class of applications and expand the horizon of numerous current applications. To meet these challenges, we developed the Sensor Composer, a foundation tool in which we treat each sensor as a sharable and reusable entity and assemble these sensors to form higher level abstract sensor based on the user requirements.