Research On Mechanisms Of Optimizing Resource Allocation And Its Security In Smart Identifier Network

Limitation of the so-called "triple bindings",namely resource/location binding,user/network binding and control/data binding, the current Internet has many inhernet shortcomings, such as poor security, low mobility and high consumption, which all seriously hinder its further development. Smart Identifier NETwork (SINET) is a novel future Internet architecture,which removes the restrictions from the "triple bindings"completely. SINET contains three layers and two domains horizontally. As the part of the SINET theory, targeting at implementing the inherent wisdom in SINET, this thesis aims at studying its smart feature, mainly focus on mechanisms of optimizing resource allocation and its security. The main contributions of the thesis are as follows:Firstly, we propose an adaptive approach for elephant flow detection, which could efficiently identify elephant flows with low-latency and low-overhead. Particularly, in order to meet the the demands of the traffic characteristics in SINET, dynamical traffic learning algorithm is adopted to configure the threshold value real-timely and dynamically. Furthermore, we propose hierarchical statistics pulling to save bandwidth comsumption and processing time with two supplement functions called elephant store and range splitting, when the controller detects the elephant flows. We use OMNET++emulator and mathematical analysis to verify our methods. Both of them confirm the benefits of our approaches.Secondly, we propose an enhanced mechanism of elephant flow scheduling in SINET-based network. The main idea of this mechanism is to use the resource adaptation resolving server to dynamically perceive the network resource state, and calculate the link utilization in real time. Then the approach optimizes the dispatching of the detected large flow according to the current link utilization situation. The mechanism can efficiently reduce the overhead and improve the scalability of control plane by using Parametric Minimum Cross Entropy (PMCE) algorithm. We describe the proposed approach in detail, and evaluate it in OMnet++ to verity its feasibility and effectiveness. Numerical results show that the benefits of the scheme are better than previous methods and the extra delay caused by PMCE algorithm is controllable.Thirdly, we we proposed an effective detection method, which has low-overhead and high-accuracy. We first classify the potential switches that are compromised using Bayesian Network, which is a supervised learning algorithm. Then, we deploy the anomaly detection on the vulnerable switches to detect the Packet-In messages flooding attack based on Fuzzy C-Means. Extensive simulations and testbed-based experiments show that the presented solution can defeat the Packet-In messages flooding attack with low-overhead and high-accuracy.