A Parallel Network Emulation Method for Evaluating the Correctness and Performance of Applications

Developing new solutions for challenging communication environments requires extensive testing in various different scenarios. While discrete-event network simulators scale well, their main use is in evaluating specific algorithms and not necessarily assessing the real-world performance of an actual application. In contrast, full network emulation is a costly but more realistic way to perform an evaluation. These types of evaluation provide more valuable insights for both developers and users. Nevertheless, oftentimes the wheel is reinvented to provide a newly written simulation environment for a specific networking software or handwritten evaluation and reporting scripts for common metrics. Unfortunately, the downside of customized solutions is potential oversights regarding important metrics, and later finding out various problems when deploying the software in the field. In this paper, we present coreemu-lab, a novel framework that automates the process of orchestrating different monitoring services in an emulated, lightweight networking environment. This includes simulated mobility, automated data collection, and analysis of common metrics. It is specifically designed for the evaluation of resilient, decentralized communication software in challenging, mobile scenarios. Our evaluation shows the flexibility and ease of use of our proposed multi-platform, open-source solution.

Using ad hoc networks as alternative means of communication in disaster situations is a salutary solution. However, analysing application performance is mandatory for evaluating such a possibility. In this paper we present the two aspects of our approach to application performance assessment on wireless ad hoc networks. The first aspect refers to real-world tests in which we quantify objectively the relationship between network conditions and application performance. The second aspect is represented by the wireless network (WLAN) emulator that we design to run on StarBED, the large-scale network experiment environment of the Hokuriku Research Centre in Ishikawa, Japan. By combining these two aspects we perform experiments with real applications, while having full control of the network conditions in which the application is tested (when using emulation).

Network softwarization has revitalized the interest of the network community towards emulation as an effective mechanism for network experimentation. Relevant platforms automate the deployment of virtual network topologies on a host, providing users the ability to manually run experimental scenarios. Whilst this may suit prototyping, modern development and deployment practices such as CI/CD depend on fully automated testing processes, built around high-level testing APIs and abstracting the challenges involved with synchronizing complex node interaction scenarios. In this paper, we present Network Emulation System (NES): a cloud-native, and highly-parallelizable Network Emulation as a Service (NEaaS) platform designed from the ground up to facilitate codeless experiment specification and to automate network testing workflows in cloud CI/CD environments. We demonstrate that NES offers a 8x speedup improvement in topology instantiation times in comparison to existing emulation platforms, and its life-cycle model can automate testing processes for complex service configurations using existing CI/CD platforms such as GitHub Actions.

A cooperative shared haptic virtual environment, where the users can kinesthetically interact and simultaneously feel each other, is beneficial for many VR simulations. A little is known about the influences of the network delay on the quality of haptic and the performance in such environments. This paper has addressed this issue by exploring the haptic sensation and the performance in a tele-handshake cooperative shared haptic system, where network emulator was exploited to simulate the real network cloud. An experiment for different delay settings has been conducted and an objective and subjective evaluation has been described. An objective evaluation of the force feedback and the performance showed that there was no noticeable effect of the delay on the force feedback, but it had a negative impact on the performance. Meanwhile, the subjective evaluation suggested that the quality of haptic sensation deteriorated as the delay increased. Nevertheless, the examined tele-handshake system was able to deliver a high quality of haptic sensation, good performance, and stability for large delays.

This paper describes the DIStributed Network Emulation Sys tem (DISNES) modeling architecture for a gigabit per second Lo cal Area Network (LAN). The emulation of networks is an im portant capability for rapidly bringing new network technology to market because it allows for the simultaneous research and devel opment of network hardware and software. A modular, object oriented strategy is employed to model network components fea turing different functionalities. The distributed emulation system has demonstrated its usefulness by allowing the rapid development of a network management interface and a model for network man agement information. In particular, DISNES incorporates an ex perimental portion of the Internet Management Information Base (IMIB), operated through the Simple Network Management Pro tocol (SNMP), which defines objects for managing Permanent Vir tual Circuit (PVC) parameters on Asynchronous Transfer Mode (ATM) switches. The flexibility of DISNES makes it especially suitable for the development of distributed networked protocols and applications.

We have proposed a constructive learning method, called multi-sieving learning, for implementing automatic decomposition of learning tasks and a parallel and modular multi-sieving network architecture in our previous work (1995). In this paper we present a new parallel and modular multi-sieving neural network architecture to which multiple control networks are introduced. In this architecture the learning task for a control network is decomposed into a finite set of manageable subtasks, and each subtask is learned by an individual control sub-network. An important advantage of this architecture is that the learning tasks for control networks can be learned efficiently, and therefore automatic decomposition of complex learning tasks can be achieved easily.

In the last few years many emulation systems have been developed to evaluate the effectiveness of new protocols and applications in realistic and controllable network scenarios. NEPTUNE is a flexible architecture for cluster-based network emulation systems designed and implemented at University of N

Cross-layer optimization for 3-D video transmission over cooperative relay systems

Virtual reality (VR) is a rapidly developing technology that has a wide spectrum of industrial and commercial applications. Networked (distributed or shared) virtual environments (VE) are of growing interest to modern manufacturing industry; a dominating use of networked virtual manufacturing environments (VMEs) is on-line visualisation and collaborative control of 3D information. This has to be supported by real-time data transfer. To meet a broad range of common requirements for Internet-based VE communications, particularly for virtual manufacturing and collaborative design and control, this paper presents a networked virtual environment system that is designed to support networked virtual design and manufacturing. The system is implemented with manufacturing message specification (MMS) standards so as to integrate a range of manufacturing services into networked VEs over the Internet.

Existing testing systems are not suited for testing of tactical hardware in complex battlefield configurations due to the increased complexity of sensors that include network interfaces and sensor fusion algorithms. We are developing a system for augmented testing of wireless sensor networks that supports virtual system configurations interacting with the real-time hardware in the loop. The testbed provides a realistic simulated representation of a tactical network that allows faithful testing of networked systems focusing on hardware in the loop sensors and sensor fusion systems. Systems can be tested using this method in a controlled repeatable environment not feasible in field testing. The system design combines dedicated high performance computing (HPC) resources with a scalable, high fidelity network emulation and a computer generated forces (CGF) model to virtually represent the tactical network, force movement, interactions, and communication loads to systems under test. The system is implemented using QualNet network emulation and OneSAF CGF running on a parallel Linux blade system and integrates sensors to the simulated virtual environment through wireless gateways and Ethernet connections. This paper presents the design framework, current implementation of the system, and real-time performance measurements.

The influences of network characteristics on the task performance in a desktop collaborative virtual environment (CVE) system were studied. In the experiment, two participants worked together to accomplish a collaborative task and a network emulator was utilized to simulate real network characteristics. The task used in the experiment was designed to be highly interactive and had rigid requirements on network characteristics (e.g. latency and packet loss). The objective performance and subjective assessment were recorded and analyzed quantitatively as a function of network latency and packet loss rate. The experiment results show that network latency has no distinct effect on task performance when network latency is less than 200 ms and lightly affect task performance when network latency is between 200 and 600 ms; packet loss has no effect on task performance when there are 40 state update messages received per second and lightly affect task performance when there are between 15 and 40 state update messages received per second.

For many large scientific applications, computing on a cluster is a viable, economical alternative to a dedicated parallel machine. Application performance on a cluster is largely determined by the speed of the underlying communication network. The authors use a parallel network approach to improve the communication network performance. More specifically, they use multiple networks based on Ethernet to improve the performance. The authors modified the parallel virtual machine (PVM) environment to take advantage of multiple Ethernets by using round robin scheduling strategy. The modified PVM, referred as PPVM, showed an improvement of 35% over PVM under no background load, and an improvement of 100% under background load on two channels. >

Summary form only given. An algorithm for longitudinal motion stereo based on a discrete parallel neural network is discussed. Longitudinal motion stereo is a method to infer depth information from a forward or backward moving camera. Existing algorithms have some problems associated with the location of the focus of expansion (FOE), camera orientation, and surface orientation. The present algorithm allows the camera to move along its optical axis freely, needs no information on the FOE, and makes no requirements on the surface orientation. The algorithm uses a Gabor correlation operator to extract image features and employs a discrete parallel neural network to compute the disparity field based on the Gabor features. A depth map is then derived from the disparity field by simple algebraic computations. >

Virtual reality (VR) enables the study of visually-guided behavior in situations that are otherwise not amenable to laboratory research. An example is a fire evacuation scenario, in which building occupants have to find their way to safety through smoke and flames, identify safe egress routes, and avoid threats. First, however, it must be determined whether evacuation behavior in VR is comparable to that in a real environment. We are thus initially studying participant responses to a fire alarm in matched virtual and real environments. The present experiment investigates how the decision to evacuate is influenced by visual cues about a neighbor’s behavior. In the Control group, a lone participant is exposed to a fire alarm while performing a bogus task. In the Passive group, a participant performs the same task with a confederate who ignores the fire alarm. In the Active group, the confederate immediately leaves the room when the fire alarm sounds. The gender of the participant is crossed with the gender of the confederate. 75 participants experience the scenario in a real environment (25 per group), and 75 others in a matched virtual environment presented in a wireless head mounted display (HMD), with avatars of the confederates. In the real environment, two-thirds of participants in the Control group evacuated the room. In the Active group, nearly all did so, whereas in the Passive group all but one did. More female than male participants evacuated in the Control group, but there were no other gender effects. Data collection in the virtual environment is in progress, and will be completed by spring. The findings thus far demonstrate a very strong influence of neighbor behavior on a participant’s evacuation decision. The results will provide answers to the much-debated question of whether evacuation behavior is similar in virtual and real environments. Meeting abstract presented at VSS 2015

This paper presents an ontology-based approach to the problem of jobs scheduling in case where jobs are processed by applications running in virtual environments and number of applications and their performance varies over time. Using ontology-based framework brings benefits when system has a varying number of components and their performing properties are also non-constant. The work is focused on ontology model needed to organize information exchange for intelligent agents embedded into virtual machines and gathering information about applications performance. In cases when jobs of one type can be processed by several applications having different performance, the existence of optimal threshold queuing policy has been proven earlier. It can reduce the average job processing time. In order to calculate thresholds we need relevant information about active applications and their current performance, the rate of jobs stream, the number of jobs in the queues, etc. The presented approach solves the problem of effective gathering of relevant information about the system state based on intelligent agents interaction where each intelligent agent uses ontology to publish only information about changes that are relevant to decision making. This reduces the system’s overhead for monitoring of ongoing parameters.