Search

Similar to terrestrial networks where edge computing facilities have already been introduced to decrease the user request response delay and to reduce the backhaul bandwidth consumption, low earth orbit (LEO) satellite constellation networks can also be benefited by adopting edge computing technologies. The computation tasks generated by ground users can be offloaded to their accessing LEO satellites to enhance network QoS and user QoE. This paper focuses on modeling and simulating computation offloading at LEO constellation network edge. A one-dimensional networking model for edge computing enabled LEO constellation networks is derived, and on that basis, a Monte Carlo simulator is developed from scratch to evaluate system performance. As a case study, three different computation offloading schemes are elaborated and implemented on the simulator. Comparative evaluation experiments have been conducted and the results indicate that, in resource restricted scenarios, allowing computation offloading to the neighbors of the access satellites can considerably reduce the request blocking probability with only slightly increasing the average request response delay.

Named Data Network (NDN) technology can efficiently mitigate the negative influence due to instable links and dynamic topology on MANET performance. Considering MANET nodes’ mobility, this paper enhances NDN’s shortest path routing algorithm by storing the interest packets generated by the consumers in the pending interest tables (PITs) of the shortest path’s neighboring nodes, aiming at building the backup paths for transmitting back the requested data packets, so as to mitigate the shortest path disruption problem caused by MANET topology dynamicity. This paper designs and implements a simulation platform based on NS3/ndnSIM, and the experimental results indicate that, the enhanced shortest path plus backup (SPPB) algorithm can effectively decrease the average request response delay and improve successful request response ratio.

Network emulation is an essential method to test network architecture, protocol and application software during a network’s entire life-cycle. Compared with simulation and test-bed methods, network emulation possesses the advantages of accuracy and cost-efficiency. However, legacy network emulators are typically restricted in scalability, agility, and extensibility, which builds barriers to prevent them from being widely used. In this paper, we introduce the currently prevalent cloud computing and related technologies including resource virtualization, NFV (network functional virtualization), SDN (software-defined networking), traffic control and flow steering to the network emulation domain. We design and implement an innovative cloud-based network emulation platform, aiming at providing users Network Emulation as a Service (NEaaS), which can be conveniently deployed on both public and private clouds. We carried out performance evaluation and discussion on this platform. It turns out, the platform can significantly outperform most legacy network emulators regarding to the scalability, agility, and extensibility, with much lower emulation costs.

In the last decade, various resource virtualization technologies have been widely applied in ICT industry, particularly the cloud computing domain. These virtualization technologies can squeeze out hardware potential and consequently can save expenditure. Virtualization technologies are used in the network emulation domain to emulate network nodes, which could be quite heterogeneous in terms of hardware architecture. Currently, many network emulators utilize x86 based virtual machines (VMs) to emulate target network nodes of heterogeneous architectures, i.e. ARM, SPARC, PPC, etc., which may introduce incompatibility to the original system and application software of the target nodes, and will consequently jeopardize the emulation fidelity. This paper focuses on alleviating the emulation incompatibility caused by node heterogeneity. Firstly, this emulation incompatibility problem is investigated and analyzed. Then, a hybrid virtualization approach to emulate heterogeneous nodes is elaborated and implemented in a cloud-based network emulation system. A case study of applying the proposed approach to emulate a space-ground integrated network (SGIN) is conducted. Functional verification and performance evaluation experiments lead to the results, which show the hybrid approach can effectively dispose of the incompatibility problem with an affordable performance degradation.