Network processing plays an important role in the development of Internet as more and more complicated applications are deployed throughout the network. With the advent of new platforms such as network processors (NPs...Network processing plays an important role in the development of Internet as more and more complicated applications are deployed throughout the network. With the advent of new platforms such as network processors (NPs) that incorporate novel architectures to speedup packet processing, there is an increasing need for an efficient method to facilitate the study of their performance. In this paper, we present a tool called SimNP, which provides a flexible platform for the simulation of a network processing system in order to provide information for workload characterization, architecture development, and application implementation. The simulator models several architectural features that are commonly employed by NPs, including multiple processing engines (PEs), integrated network interface and memory controller, and hardware accelerators. ARM instruction set is emulated and a simple memory model is provided so that applications implemented in high level programming language such as C can be easily compiled into an executable binary using a common compiler like gcc. Moreover, new features or new modules can also be easily added into this simulator. Experiments have shown that our simulator provides abundant information for the study of network processing systems.展开更多
Software-defined networking(SDN)algorithms are gaining increas-ing interest and are making networks flexible and agile.The basic idea of SDN is to move the control planes to more than one server’s named controllers a...Software-defined networking(SDN)algorithms are gaining increas-ing interest and are making networks flexible and agile.The basic idea of SDN is to move the control planes to more than one server’s named controllers and limit the data planes to numerous sending network components,enabling flexible and dynamic network management.A distinctive characteristic of SDN is that it can logically centralize the control plane by utilizing many physical controllers.The deployment of the controller—that is,the controller placement problem(CPP)—becomes a vital model challenge.Through the advancements of blockchain technology,data integrity between nodes can be enhanced with no requirement for a trusted third party.Using the lat-est developments in blockchain technology,this article designs a novel sea turtle foraging optimization algorithm for the controller placement problem(STFOA-CPP)with blockchain-based intrusion detection in an SDN environ-ment.The major intention of the STFOA-CPP technique is the maximization of lifetime,network connectivity,and load balancing with the minimization of latency.In addition,the STFOA-CPP technique is based on the sea turtles’food-searching characteristics of tracking the odour path of dimethyl sulphide(DMS)released from food sources.Moreover,the presented STFOA-CPP technique can adapt with the controller’s count mandated and the shift to controller mapping to variable network traffic.Finally,the blockchain can inspect the data integrity,determine significantly malicious input,and improve the robust nature of developing a trust relationship between sev-eral nodes in the SDN.To demonstrate the improved performance of the STFOA-CPP algorithm,a wide-ranging experimental analysis was carried out.The extensive comparison study highlighted the improved outcomes of the STFOA-CPP technique over other recent approaches.展开更多
Network virtualization(NV) is considered as an enabling tool to remove the gradual ossification of current Internet. In the network virtualization environment, a set of heterogeneous virtual networks(VNs), isolated fr...Network virtualization(NV) is considered as an enabling tool to remove the gradual ossification of current Internet. In the network virtualization environment, a set of heterogeneous virtual networks(VNs), isolated from each other, share the underlying resources of one or multiple substrate networks(SNs) according to the resource allocation strategy. This kind of resource allocation strategy is commonly known as so called Virtual Network Embedding(VNE) algorithm in network virtualization. Owing to the common sense that VNE problem is NP-hard in nature, most of VNE algorithms proposed in the literature are heuristic. This paper surveys and analyzes a number of representative heuristic solutions in the literature. Apart from the analysis of representative heuristic solutions, a taxonomy of the heuristic solutions is also presented in the form of table. Future research directions of VNE, especially for the heuristics, are emphasized and highlighted at the end of this survey.展开更多
Network virtualization(NV) is widely considered as a key component of the future network and promises to allow multiple virtual networks(VNs) with different protocols to coexist on a shared substrate network(SN). One ...Network virtualization(NV) is widely considered as a key component of the future network and promises to allow multiple virtual networks(VNs) with different protocols to coexist on a shared substrate network(SN). One main challenge in NV is virtual network embedding(VNE). VNE is a NPhard problem. Previous VNE algorithms in the literature are mostly heuristic, while the remaining algorithms are exact. Heuristic algorithms aim to find a feasible embedding of each VN, not optimal or sub-optimal, in polynomial time. Though presenting the optimal or sub-optimal embedding per VN, exact algorithms are too time-consuming in smallscaled networks, not to mention moderately sized networks. To make a trade-off between the heuristic and the exact, this paper presents an effective algorithm, labeled as VNE-RSOT(Restrictive Selection and Optimization Theory), to solve the VNE problem. The VNERSOT can embed virtual nodes and links per VN simultaneously. The restrictive selection contributes to selecting candidate substrate nodes and paths and largely cuts down on the number of integer variables, used in the following optimization theory approach. The VNE-RSOT fights to minimize substrate resource consumption and accommodates more VNs. To highlight the efficiency of VNERSOT, a simulation against typical and stateof-art heuristic algorithms and a pure exact algorithm is made. Numerical results reveal that virtual network request(VNR) acceptance ratio of VNE-RSOT is, at least, 10% higher than the best-behaved heuristic. Other metrics, such as the execution time, are also plotted to emphasize and highlight the efficiency of VNE-RSOT.展开更多
This paper investigates the maximal achievable multi-rate throughput problem of a multicast ses-sion at the presence of network coding. Deviating from previous works which focus on single-rate network coding, our work...This paper investigates the maximal achievable multi-rate throughput problem of a multicast ses-sion at the presence of network coding. Deviating from previous works which focus on single-rate network coding, our work takes the heterogeneity of sinks into account and provides multiple data layers to address the problem. Firstly formulated is the maximal achievable throughput problem with the assumption that the data layers are independent and layer rates are static. It is proved that the problem in this case is, unfortunately, Non-deterministic Polynomial-time (NP)-hard. In addition, our formulation is extended to the problems with dependent layers and dynamic layers. Furthermore, the approximation algorithm which satisfies certain fair-ness is proposed.展开更多
We address the problem of optimizing a distributed monitoring system and the goal of the optimization is to reduce the cost of deployment of the monitoring infrastructure by identifying a minimum aggregating set subje...We address the problem of optimizing a distributed monitoring system and the goal of the optimization is to reduce the cost of deployment of the monitoring infrastructure by identifying a minimum aggregating set subject to delay constraint on the aggregating path. We show that this problem is NP-hard and propose approximation algorithm proving the approximation ratio with lnm+1, where is the number of monitoring nodes. At last we extend our modal with more constraint of bounded delay variation. Key words network - distributed monitoring - delay constraint - NP-hard CLC number TP 393 Foundation item: Supported by the National Natural Science Foundation of China (60373023)Biography: LIU Xiang-hui(1973-), male, Ph. D. candidate, research direction: algorithm complexity analysis, QoS in Internet.展开更多
This paper develops a deep learning tool based on neural processes(NPs)called the Peri-Net-Pro,to predict the crack patterns in a moving disk and classifies them according to the classification modes with quantified u...This paper develops a deep learning tool based on neural processes(NPs)called the Peri-Net-Pro,to predict the crack patterns in a moving disk and classifies them according to the classification modes with quantified uncertainties.In particular,image classification and regression studies are conducted by means of convolutional neural networks(CNNs)and NPs.First,the amount and quality of the data are enhanced by using peridynamics to theoretically compensate for the problems of the finite element method(FEM)in generating crack pattern images.Second,case studies are conducted with the prototype microelastic brittle(PMB),linear peridynamic solid(LPS),and viscoelastic solid(VES)models obtained by using the peridynamic theory.The case studies are performed to classify the images by using CNNs and determine the suitability of the PMB,LBS,and VES models.Finally,a regression analysis is performed on the crack pattern images with NPs to predict the crack patterns.The regression analysis results confirm that the variance decreases when the number of epochs increases by using the NPs.The training results gradually improve,and the variance ranges decrease to less than 0.035.The main finding of this study is that the NPs enable accurate predictions,even with missing or insufficient training data.The results demonstrate that if the context points are set to the 10th,100th,300th,and 784th,the training information is deliberately omitted for the context points of the 10th,100th,and 300th,and the predictions are different when the context points are significantly lower.However,the comparison of the results of the 100th and 784th context points shows that the predicted results are similar because of the Gaussian processes in the NPs.Therefore,if the NPs are employed for training,the missing information of the training data can be supplemented to predict the results.展开更多
文摘Network processing plays an important role in the development of Internet as more and more complicated applications are deployed throughout the network. With the advent of new platforms such as network processors (NPs) that incorporate novel architectures to speedup packet processing, there is an increasing need for an efficient method to facilitate the study of their performance. In this paper, we present a tool called SimNP, which provides a flexible platform for the simulation of a network processing system in order to provide information for workload characterization, architecture development, and application implementation. The simulator models several architectural features that are commonly employed by NPs, including multiple processing engines (PEs), integrated network interface and memory controller, and hardware accelerators. ARM instruction set is emulated and a simple memory model is provided so that applications implemented in high level programming language such as C can be easily compiled into an executable binary using a common compiler like gcc. Moreover, new features or new modules can also be easily added into this simulator. Experiments have shown that our simulator provides abundant information for the study of network processing systems.
文摘Software-defined networking(SDN)algorithms are gaining increas-ing interest and are making networks flexible and agile.The basic idea of SDN is to move the control planes to more than one server’s named controllers and limit the data planes to numerous sending network components,enabling flexible and dynamic network management.A distinctive characteristic of SDN is that it can logically centralize the control plane by utilizing many physical controllers.The deployment of the controller—that is,the controller placement problem(CPP)—becomes a vital model challenge.Through the advancements of blockchain technology,data integrity between nodes can be enhanced with no requirement for a trusted third party.Using the lat-est developments in blockchain technology,this article designs a novel sea turtle foraging optimization algorithm for the controller placement problem(STFOA-CPP)with blockchain-based intrusion detection in an SDN environ-ment.The major intention of the STFOA-CPP technique is the maximization of lifetime,network connectivity,and load balancing with the minimization of latency.In addition,the STFOA-CPP technique is based on the sea turtles’food-searching characteristics of tracking the odour path of dimethyl sulphide(DMS)released from food sources.Moreover,the presented STFOA-CPP technique can adapt with the controller’s count mandated and the shift to controller mapping to variable network traffic.Finally,the blockchain can inspect the data integrity,determine significantly malicious input,and improve the robust nature of developing a trust relationship between sev-eral nodes in the SDN.To demonstrate the improved performance of the STFOA-CPP algorithm,a wide-ranging experimental analysis was carried out.The extensive comparison study highlighted the improved outcomes of the STFOA-CPP technique over other recent approaches.
基金supported by the National Natural Science Foundation of China under Grants 61372124 and 61401225the National Science Foundation of Jiangsu Province under Grant BK20140894the Postgraduate Research&Practice Innovation Program of Jiangsu Province under Grant KYCX17_0784
文摘Network virtualization(NV) is considered as an enabling tool to remove the gradual ossification of current Internet. In the network virtualization environment, a set of heterogeneous virtual networks(VNs), isolated from each other, share the underlying resources of one or multiple substrate networks(SNs) according to the resource allocation strategy. This kind of resource allocation strategy is commonly known as so called Virtual Network Embedding(VNE) algorithm in network virtualization. Owing to the common sense that VNE problem is NP-hard in nature, most of VNE algorithms proposed in the literature are heuristic. This paper surveys and analyzes a number of representative heuristic solutions in the literature. Apart from the analysis of representative heuristic solutions, a taxonomy of the heuristic solutions is also presented in the form of table. Future research directions of VNE, especially for the heuristics, are emphasized and highlighted at the end of this survey.
基金supported by the National Basic Research Program of China (973 Program) under Grant 2013CB329104the National Natural Science Foundation of China under Grant 61372124 and 61427801the Key Projects of Natural Science Foundation of Jiangsu University under Grant 11KJA510001
文摘Network virtualization(NV) is widely considered as a key component of the future network and promises to allow multiple virtual networks(VNs) with different protocols to coexist on a shared substrate network(SN). One main challenge in NV is virtual network embedding(VNE). VNE is a NPhard problem. Previous VNE algorithms in the literature are mostly heuristic, while the remaining algorithms are exact. Heuristic algorithms aim to find a feasible embedding of each VN, not optimal or sub-optimal, in polynomial time. Though presenting the optimal or sub-optimal embedding per VN, exact algorithms are too time-consuming in smallscaled networks, not to mention moderately sized networks. To make a trade-off between the heuristic and the exact, this paper presents an effective algorithm, labeled as VNE-RSOT(Restrictive Selection and Optimization Theory), to solve the VNE problem. The VNERSOT can embed virtual nodes and links per VN simultaneously. The restrictive selection contributes to selecting candidate substrate nodes and paths and largely cuts down on the number of integer variables, used in the following optimization theory approach. The VNE-RSOT fights to minimize substrate resource consumption and accommodates more VNs. To highlight the efficiency of VNERSOT, a simulation against typical and stateof-art heuristic algorithms and a pure exact algorithm is made. Numerical results reveal that virtual network request(VNR) acceptance ratio of VNE-RSOT is, at least, 10% higher than the best-behaved heuristic. Other metrics, such as the execution time, are also plotted to emphasize and highlight the efficiency of VNE-RSOT.
基金Supported by the National 863 High-tech Program of China (No.2003AA121560) and High-tech Project of Jiangsu Province (No.BG2003001).
文摘This paper investigates the maximal achievable multi-rate throughput problem of a multicast ses-sion at the presence of network coding. Deviating from previous works which focus on single-rate network coding, our work takes the heterogeneity of sinks into account and provides multiple data layers to address the problem. Firstly formulated is the maximal achievable throughput problem with the assumption that the data layers are independent and layer rates are static. It is proved that the problem in this case is, unfortunately, Non-deterministic Polynomial-time (NP)-hard. In addition, our formulation is extended to the problems with dependent layers and dynamic layers. Furthermore, the approximation algorithm which satisfies certain fair-ness is proposed.
文摘We address the problem of optimizing a distributed monitoring system and the goal of the optimization is to reduce the cost of deployment of the monitoring infrastructure by identifying a minimum aggregating set subject to delay constraint on the aggregating path. We show that this problem is NP-hard and propose approximation algorithm proving the approximation ratio with lnm+1, where is the number of monitoring nodes. At last we extend our modal with more constraint of bounded delay variation. Key words network - distributed monitoring - delay constraint - NP-hard CLC number TP 393 Foundation item: Supported by the National Natural Science Foundation of China (60373023)Biography: LIU Xiang-hui(1973-), male, Ph. D. candidate, research direction: algorithm complexity analysis, QoS in Internet.
基金Project supported by the National Science Foundation of U.S.A.(Nos.DMS-1555072,DMS-2053746DMS-2134209)+1 种基金the Brookhaven National Laboratory of U.S.A.(No.382247)U.S.Department of Energy(DOE)Office of Science Advanced Scientific Computing Research Program(Nos.DESC0021142 and DE-SC0023161)。
文摘This paper develops a deep learning tool based on neural processes(NPs)called the Peri-Net-Pro,to predict the crack patterns in a moving disk and classifies them according to the classification modes with quantified uncertainties.In particular,image classification and regression studies are conducted by means of convolutional neural networks(CNNs)and NPs.First,the amount and quality of the data are enhanced by using peridynamics to theoretically compensate for the problems of the finite element method(FEM)in generating crack pattern images.Second,case studies are conducted with the prototype microelastic brittle(PMB),linear peridynamic solid(LPS),and viscoelastic solid(VES)models obtained by using the peridynamic theory.The case studies are performed to classify the images by using CNNs and determine the suitability of the PMB,LBS,and VES models.Finally,a regression analysis is performed on the crack pattern images with NPs to predict the crack patterns.The regression analysis results confirm that the variance decreases when the number of epochs increases by using the NPs.The training results gradually improve,and the variance ranges decrease to less than 0.035.The main finding of this study is that the NPs enable accurate predictions,even with missing or insufficient training data.The results demonstrate that if the context points are set to the 10th,100th,300th,and 784th,the training information is deliberately omitted for the context points of the 10th,100th,and 300th,and the predictions are different when the context points are significantly lower.However,the comparison of the results of the 100th and 784th context points shows that the predicted results are similar because of the Gaussian processes in the NPs.Therefore,if the NPs are employed for training,the missing information of the training data can be supplemented to predict the results.
