A Statistical Model for Early Detection of DDoS Attacks on Random Targets in SDN

Wireless Personal Communications - Software Defined Networks (SDNs) have accelerated and simplified the management, configuration and error detection in today’s networking systems. However,...     

A Stochastic Approximation Approach for Foresighted Task Scheduling in Cloud Computing

Wireless Personal Communications - With the increasing and elastic demand for cloud resources, finding an optimal task scheduling mechanism become a challenge for cloud service providers. Due to...     

A token-based incentive mechanism for video streaming applications in peer-to-peer networks

Free-riding is one of the main challenges of Peer-to-Peer (P2P) streaming systems which results in reduction in video streaming quality. Therefore, providing an incentive mechanism for stimulating cooperation is one of the essential requirements to maintain video Quality of Experience (QoE) in such systems. Among the existing mechanisms, payment-based schemes are most suitable for streaming applications due to their low overhead. However, to date, no dynamic payment mechanism has been proposed which can take the stochastic dynamics of the video streaming ecosystem (e.g., the request arrival, demand submission, bandwidth availability, etc.) into account. In this paper, we propose a dynamic token-based payment mechanism in which each peer earns tokens by admitting other peers’ requests and spends tokens for submitting its demands to the others. This system allows the peers to dynamically adjust their income level in adaptation to changes in the system state. We propose a Constrained Markov Decision Process (CMDP) formulation in which the goal of each peer is to obtain a request admission policy which minimizes the expected cumulative cost of consumed bandwidth, while satisfying a long-term constraint on the Mean Opinion Score (MOS) of the users as the measure of QoE. The proposed admission policy is adaptive to the request arrival process, bandwidth state and the token bucket length of each peer. To make up for the lack of design-time knowledge of the system’s statistics, each individual peer is equipped with a model-free algorithm to learn its optimal admission policy over the course of real-time interaction with the system. Simulation results are presented to compare the performance of the proposed algorithm against baseline schemes such as: random, token-threshold, bandwidth-threshold and myopic algorithms.     

A preliminary comparison of epidural lidocaine and xylazine during total intravenous anaesthesia in Iranian fat-tailed sheep

The effects of total intravenous anaesthesia using diazepam-ketamine (D-K) mixture in combination with epidural lidocaine or xylazine were studied in 17 healthy, Iranian fat-tailed sheep undergoing hindlimb orthopaedic surgery. All sheep were given diazepam (0.4 mg/kg) and ketamine (4 mg/kg) as induction agents. Following endotracheal intubation and administration of oxygen, the animal received lidocaine (2%, 0.2 ml/kg = 4 mg/kg) or xylazine (0.08 mg/kg, diluted in 0.9% NaCl to a volume of 0.2 ml/kg) epidurally. Anaesthesia was maintained for 174.2 +/- 7.8 minutes by intermittent injection of D-K (2.5 mg/ml and 25 mg/ml, respectively). This drug combination provided satisfactory anaesthesia for more than 2.5 hours. The quality of recovery was good. Our results demonstrate that the combination of total intravenous anaesthesia (D-K) and epidural analgesia (lidocaine or xylazine) provides a suitable technique for hindlimb orthopaedic surgery in sheep. Epidural administration of lidocaine or xylazine provided effective analgesia and significantly decreased the dose of D-K required to maintain anaesthesia. Further studies would be required to determine details of cardiopulmonary effects of D-K infusion.