Request Dispatching Over Distributed SDN Control Plane: A Multiagent Approach.

Software-defined networking (SDN) allows flexible and centralized control in cloud data centers. An elastic set of distributed SDN controllers is often required to provide sufficient yet cost-effective processing capacity. However, this introduces a new challenge: Request Dispatching among the controllers by SDN switches. It is essential to design a dispatching policy for each switch to guide the request distribution. Existing policies are designed under certain assumptions, including a single centralized agent, global network knowledge, and a fixed number of controllers, which often cannot be satisfied in practice. This article proposes MADRina, Multiagent Deep Reinforcement Learning for request dispatching, to design policies with high dispatching adaptability and performance. First, we design a multiagent system to address the limitation of using a centralized agent with global network knowledge. Second, we propose a Deep Neural Network-based adaptive policy to enable request dispatching over an elastic set of controllers. Third, we develop a new algorithm to train the adaptive policies in a multiagent context. We prototype MADRina and build a simulation tool to evaluate its performance using real-world network data and topology. The results show that MADRina can significantly reduce response time by up to 30% compared to existing approaches.

Load-sensitive dynamic workflow re-orchestration and optimisation for faster patient healthcare.

Hospital waiting times are considerably long, with no signs of reducing any-time soon. A number of factors including population growth, the ageing population and a lack of new infrastructure are expected to further exacerbate waiting times in the near future. In this work, we show how healthcare services can be modelled as queueing nodes, together with healthcare service workflows, such that these workflows can be optimised during execution in order to reduce patient waiting times. Services such as X-ray, computer tomography, and magnetic resonance imaging often form queues, thus, by taking into account the waiting times of each service, the workflow can be re-orchestrated and optimised. Experimental results indicate average waiting time reductions are achievable by optimising workflows using dynamic re-orchestration.

Preference based load balancing as an outpatient appointment scheduling aid.

Load balancing is a performance improvement aid in various applications of distributed systems. In this paper we propose a preference based load balancing strategy as a scheduling aid in an outpatient clinic of an online medical consultation system. The performance objectives are to maximizing throughout and minimizing waiting time. Patients will provide a standard set of preferences prior to scheduling an appointment. The preferences are rated on to a scale and each service request will have a respective preference score. The available doctors will also be classified into classes based on their clinical expertise and the nature of the past diagnosis and the types of patients consulted. The preference scores will then be mapped on to each class and the appointment will be scheduled. The proposed scheme was modeled as a queuing system in Matlab. Matlab SimEvents library modules were used for constructing the model. Performance was analysed based on the average waiting time and utilization. The results revealed that the preference based load balancing scheme markedly reduce the waiting time and significantly improve the utilization under different load conditions.

A cardiod based technique to identify cardiovascular diseases using mobile phones and body sensors.

To prevent the threat of Cardiovascular Disease (CVD) related deaths, the usage of mobile phone based computational platforms, body sensors and wireless communications is proliferating. Since mobile phones have limited computational resources, existing PC based complex CVD detection algorithms are often unsuitable for wireless telecardiology applications. Moreover, if the existing Electrocardiography (ECG) based CVD detection algorithms are adopted for mobile telecardiology applications, then there will be processing delays due to the computational complexities of the existing algorithms. However, for a CVD affected patient, seconds worth of delay could be fatal, since cardiovascular cell damage is a totally irrecoverable process. This paper proposes a fast and efficient mechanism of CVD detection from ECG signal. Unlike the existing ECG based CVD diagnosis systems that detect CVD anomalies from hundreds of sample points, the proposed mechanism identifies cardiac abnormality from only 5 sample points. Therefore, according to our experiments the proposed mechanism is up to 3 times faster than the existing techniques. Due to less computational burden, the proposed mechanism is ideal for wireless telecardiology applications running on mobile phones.