James Xue

Dynamic Resource Allocation in Enterprise Systems

It is common that Internet service hosting centres use several logical pools to assign server resources to different applications, and that they try to achieve the highest total revenue by making efficient use of these resources. In this paper, multi-tiered enterprise systems are modelled as multi-class closed queueing networks, with each network station corresponding to each application tier. In such queueing networks, bottlenecks can limit overall system performance, and thus should be avoided. We propose a bottleneck-aware server switching policy, which responds to system bottlenecks and switches servers to alleviate these problems as necessary. The switching engine compares the benefits and penalties of a potential switch, and makes a decision as to whether it is likely to be worthwhile switching. We also propose a simple admission control scheme, in addition to the switching policy, to deal with system overloading and optimise the total revenue of multiple applications in the hosting centre. Performance evaluation has been done via simulation and results are compared with those from a proportional switching policy and also a system that implements no switching policy. The experimental results show that the combination of the bottleneck-aware switching policy and the admission control scheme consistently outperforms the other two policies in terms of revenue contribution.

Partition-based Profit Optimisation for Multi-class Requests in Clusters of Servers

This paper investigates profit optimisation by the partitioning of server pools. Different types of web requests are considered in this paper: best-effort requests and multi-class QoS-demanding requests. Each request is associated with a certain amount of revenue. The system earns associated revenue when completing a best effort request, while the revenue earned for serving a QoS-demanding request depends on the degree of satisfaction. The complete server pool is divided into two clusters of servers, each dedicated to serving one type of request. The total profits accrued when serving these requests are mathematically modelled in this paper. The optimisation equations for the pool partitioning have also been constructed so that the maximum profit can be achieved. An optimal server switching policy is also developed. The server switching policy is optimal in the sense that the maximum profit can be maintained by switching servers from one partition to the other. Costs involved in server switching are factored into the switching model. Supportive experimentation based on simulation have been conducted and the results verify the effectiveness of the pool partitioning policy and the server switching policy developed in this paper.

A System for Dynamic Server Allocation in Application Server Clusters

Application server clusters are often used to service high-throughput web applications. In order to host more than a single application, an organisation will usually procure a separate cluster for each application. Over time the utilisation of the clusters will vary, leading to variation in the response times experienced by users of the applications. Techniques that statically assign servers to each application prevent the system from adapting to changes in the workload, and are thus susceptible to providing unacceptable levels of service. This paper investigates a system for allocating server resources to applications dynamically, thus allowing applications to automatically adapt to variable workloads. Such a scheme requires meticulous system monitoring, a method for switching application servers between \text it {server pools} and a means of calculating when a server switch should be made (balancing switching cost against perceived benefits). Experimentation is performed using such a switching system on a Web application testbed hosting two applications across eight application servers. The testbed is used to compare several theoretically derived switching policies under a variety of workloads. Recommendations are made as to the suitability of different policies under different workload conditions.

Distributed Broadcast Scheduling in Mobile Ad Hoc Networks with Unknown Topologies

Broadcasting is a fundamental communication task in mobile ad hoc networks, and minimizing broadcasting time (or latency) is crucial to the performance ofmany applications. Extensive studies have been conducted on the minimization of broadcasting time in the context of radio networks, which are usually modeled as general graphs. In this paper, we consider how to achieve this goal with distributed algorithms based on a more realistic (and restricted) network model. We propose a randomized algorithm that completes broadcasting in O(D log(n/D)+log2 n) time, where n is the number of nodes in the network and D the eccentricity (maximum distancefrom the source node to any other node). Compared with a previous optimal algorithm that achieves the same result for general networks, our algorithm obviates the need to know the network eccentricity D beforehand We also propose a deterministic broadcasting algorithm that works in O(n) time, which is in contrast with the best known result of O(n log2 D) for general networks.

A scheduling algorithm for revenue maximisation for cluster-based Internet services

This paper proposes a new priority scheduling algorithm to maximise site revenue of session-based multi-tier Internet services in a multicluster environment. This research is part of a larger study in support of large-scale online trading systems and, as a result, this case study is chosen as a demonstrator for the techniques presented in this paper. The trading system is partitioned into a number of operations (trade, query etc.), which by their very nature are divided into orders of importance in terms of transactional response. The algorithm in this paper is based on Mean Value Analysis (MVA), which is used for the calculation of performance metrics concerning the queuing networks and workload allocation decision support in the multicluster. In addition to this, the priority assignment is based on combination of three attributes of any given request: (i) the sender class; (ii) the operation and, (Hi) the status of the users portfolio (i.e.number of items in the users portfolio). A discrete event simulator has been developed to evaluate the performance of the priority scheduling scheme with different combinations of request attributes in various experimental scenarios. Our study aims to develop a dynamic scheduling policy, which takes into account real-time system parameters and optimises the site revenue. Although our priority scheduling algorithm is designed for an online trading system, it can be applied to most e-Commerce systems in which differentiated services are required.

Predicting the Effect on Performance of Container-Managed Persistence in a Distributed Enterprise Application

Container-managed persistence is an essential technology as it dramatically simplifies the implementation of enterprise data access. However it can also impose a significant overhead on the performance of the application at runtime. This paper presents a layered queuing performance model for predicting the effect of adding or removing container-managed persistence to a distributed enterprise application, in terms of response time and throughput performance metrics. Predictions can then be made for new server architectures - that is, server architectures for which only a small number of measurements have been made (e.g. to determine request processing speed). An experimental analysis of the model is conducted on a popular enterprise computing architecture based on IBM Websphere, using Enterprise Java Bean-based container-managed persistence as the middleware functionality. The results provide strong experimental evidence for the effectiveness of the model in terms of the accuracy of predictions, the speed with which predictions can be made and the low overhead at which the model can be rapidly parameterised.

QoS-Aware Service Selection

With the widespread use of the Internet, the number of web services that can provide similar functionality has increased rapidly in recent years. Web service selection has to be based on some non-functional attributes of the services, such as the quality of service (QoS). In this chapter, we use a server switching service that is commonly used in Internet hosting environments to explain how an agent can use a performance model to evaluate services and select the most suitable services among a number of functionally similar services returned by the service discovery. The various criteria that can be used to assess QoS are introduced in this chapter, including mean response time, throughput, system utilisation and others closely related to business such as revenue and operating costs. Service selection in the chosen case study depends on the quality and suitability of various switching policies, in other words, different switching policies can be selected depending on the QoS of the services and the run-time system state. Since the system performance can be evaluated using an analytic model, therefore, the QoS of services is assessed based on the output of the performance model.

Mining association rules for admission control and service differentiation in e-commerce applications

Workload demands in e‐commerce applications are very dynamic in nature, therefore it is essential for internet service providers to manage server resources effectively to maximize total revenue in server overloading situations. In this paper, a data mining technique is applied to a typical e‐commerce application model for identification of composite association rules that capture user navigation patterns. Two algorithms are then developed based on the derived rules for admission control, service differentiation, and priority scheduling. Our approach takes the following aspects into consideration: (a) only final purchase requests result in company revenue; (b) any other request can potentially lead to final purchase, depending upon the likelihood of the navigation sequence that starts from current request and leads to final purchase; (c) service differentiation and priority assignment are based on aggregated confidence and average support of the composite association rules. As identification of composite association rules and computation of confidence and support of the rules can be pre‐computed offline, the proposed approach incurs minimum performance overheads. The evaluation results suggest that the proposed approach is effective in terms of request management for revenue maximization.

Simplified deployment of virtual machines using an intelligent design engine

With the advent of cloud computing and the virtualisation of computer systems, challenges exist in managing resources and Virtual Machines (VMs) effectively. Cloud computing platform services offer scalable and resilient platforms, but are still not without complexity to the end user community; this paper proposes an alternative solution for the deployment of an intelligent private or public cloud compute platform, built around a set of predefined rule based parameters with the purpose of providing a simplified process for provisioning VMs. The proposed solution incorporates an architecture and design, utilising existing utilities, tools and protocols; it provides example output from the experiment conducted to prove the overall process and concept are feasible for implementation in terms of presenting a simplified deployment mechanism. By using an Intelligent Decision Engine (IDE), it enables strict adherence to a set of standard build parameters based around expert system administration knowledge, which uniquely supports a one-click VM deployment procedure. Subsequent control and management of the system environment and VMs then continue to be managed through use of the IDE and its expert knowledge base.

Routing algorithm optimization for software defined network WAN

Software Defined Network (SDN) provides a new fine-grained interface enables the routing algorithm to have an a global view of the network throughputs, connectivity and flows at the data-path. This paper aims to provide a novel approach for dynamic routing algorithm for Software Defined Network in Wide Area Network (SDN-WAN); based on using a modified shortest-widest path algorithm with a fine-grained statistical method from the OpenFlow interface, called Shortest-Feasible OpenFlow Path (SFOP). This algorithm is designed to identify the optimal route from source to destination, providing efficient utilization of the SDN-WAN resources. It achieves this aim by considering both the flow requirements and the current state of the network. SFOP computes the optimal path which provides the feasible bandwidth with the lowest hop count (delay). That will present better stability in SDN communication, QoS, and usage of available resources. Moreover, this algorithm will be the base for an SDN controller because it extracts the widest available bandwidth from source to destination for a single path. It enables the controller to decide whether it is enough to use this simple algorithm only, or if a more complicated algorithm that provides larger bandwidth such as multiple-path algorithms is needed. Finally, a testbed has been implemented using MATLAB Simulator, Pox controller, and Mininet emulator will be discussed. The latency comparison of SFOP algorithm with three other algorithms latencies shows that this algorithm finds better latency for an optimal path. Evidence will be shown that demonstrates that SFOP has good stability in dynamic changes of SDN-WAN.