Embedding and reconfiguration algorithms for service aggregation in network virtualization

Abstract

Network virtualization and transmission quality optimization are promising techniques for future Internet. Previous researches on virtual network embedding focused on efficient utilization of network resources. In this paper, we present a virtual network embedding scheme that aims at improving the effect of transmission quality optimization. In order to enable more service benefit from optimization, the embedding algorithm presented in this paper follows the service aggregation embedding principle. We also develop a reconfiguration algorithm based on service aggregation and load balance. Simulation experiments demonstrate that the proposed algorithms can achieve good performance on service aggregation as well as efficient resource utilization. average of μ = 1000 time units. The date set contained 2000 VN requests. All the VN requests were randomly classified to 4 ToS, which were Application Control, Media-Oriented, Data, and Best Effort [1].The parameter settings of simulations were α = 0.15, ε 1 = 0.09, ε 2 = 0.1, k = 3, i = 10, and the reconfiguration period was 200 time units. We compared our algorithms with G-SP [13] and VNA [12], and moreover, compared the performance of VRSA with the performance without reconfiguration. SERVICE AGGREGATION VNE ALGORITHM

Self‐adaptive online virtual network migration in network virtualization environments

“…VNRs are generated following a Poisson distribution that determines the order and arrival time for each VNR. A summary of online VNE parameters of our simulation setup is shown in Figure , which has been elaborated taking into account setups from the literature for assessment purposes …”

“…The network topologies (substrate and virtual) are produced with the GT-ITM tool. 18,[35][36][37] The setting shown in Figure 4 allowed us to experiment in scenarios where migration triggerings are likely to be unprofitable in the long term (by the time of the last VNR arrival), hence helping us to assess the capacity of the RL agent to self-adaptively learn to avoid unprofitable VN migrations. The values of CPU and BW capacities as well as link delays have been randomly generated using a minimum and maximum range following a uniform distribution.…”

“…A summary of online VNE parameters of our simulation setup is shown in Figure 4, which has been elaborated taking into account setups from the literature for assessment purposes. 18,[35][36][37] The setting shown in Figure 4 allowed us to experiment in scenarios where migration triggerings are likely to be unprofitable in the long term (by the time of the last VNR arrival), hence helping us to assess the capacity of the RL agent to self-adaptively learn to avoid unprofitable VN migrations. Finally, our simulations were performed on SNs of sizes 100, 150, 200, and 250 and densities of 0.06, 0.12, 0.25, and 0.6, to demonstrate the robustness of our self-adaptive approach on different SN topologies.…”

Enabling Adaptive Routing Service Customization via the integration of SDN and NFV

“…Especially for the novel and complex applications, such as video teleconference, live TV and online gaming, the routing service customization is needed urgently (Lima and Carvalho, 2011). However, the traditional routing service configuration methods are still mainly in manual or command-default mode (Jiang et al, 2016), which makes them difficult to adaptively adjust the QoS according to the changing network status and the application demands in an effective way. Therefore, the Internet is in need of a kind of adaptive routing service customization mechanism with functional as well as non-functional requirements considered (Klein et al, 2014).…”

Service‐oriented 5G network architecture: an end‐to‐end software defining approach

“…Consequently, SDN not only enables the centralized control plane to make the global decision and to schedule the behaviors of data‐plane devices according to the service requirements but also easily provides abundant open interfaces for the service providers to flexibly program. There are several existing studies focused on the QoS of future network or software defined network . Differently, this article addresses aforementioned challenges from the global network's perspective.…”