WAN Environment Research Articles

Optimizing Secure Decision Tree Inference Outsourcing

Outsourcing decision tree inference services to the cloud is highly beneficial, yet raises critical privacy concerns on the proprietary decision tree of the model provider and the private input data of the client. In this paper, we design, implement, and evaluate a new system that allows highly efficient outsourcing of decision tree inference. Our system significantly improves upon prior art in the overall online end-to-end secure inference service latency at the cloud as well as the local-side performance of the model provider. We first present a new scheme which securely shifts most of the processing of the model provider to the cloud, resulting in a substantial reduction on the model provider&#x0027;s performance complexities. We further devise a scheme which substantially optimizes the performance for secure decision tree inference at the cloud, particularly the communication round complexities. The synergy of these techniques allows our new system to achieve up to <inline-formula><tex-math notation="LaTeX">$8 \times$</tex-math></inline-formula> better overall online end-to-end secure inference latency at the cloud side over realistic WAN environment, as well as bring the model provider up to <inline-formula><tex-math notation="LaTeX">$19 \times$</tex-math></inline-formula> savings in communication and <inline-formula><tex-math notation="LaTeX">$18 \times$</tex-math></inline-formula> savings in computation.

Fundamentals of SD WAN Communication Technologies

Let us take a look at these new high-tech communications and how they shape the industry. Local area-defined software is a software system that manages wide area networks, provides ease of use, single site management and minimizes the cost, and can enhance communication with branch offices and the cloud. There are been some changes in local networks nothing has been more relevant than software-defined WAN or SD-WAN, which changes the way network professionals think about improving multimedia connectivity such as changing the Multi-protocol label. (MPLS), independent transfer and Digital Subscriber Line (DSL). The separation of hardware and control plane is similar to how software-defined networks use visual technology to improve the management and operation of their data centers. In fact, SD-WAN are set up and managed using proprietary protocols such as Cisco iOS. That is less hardware and less control over the hardware. A leading SD-WAN application that enables organizations to build highly efficient WANs with cost-effective turnkey Internet access. This allows organizations to phase out or completely replace low-cost WAN communication technologies such as MPLS. So customers can easily manage their network regardless of the connection provider. SD-WAN is currently one of the hottest topics with a real impact on CC services and WAN environment. SD-WAN it affects the way we think about how we use the web services yet. More importantly, it has great potential to change the way we use Communication Services in future. There are several industries that are interesting SD-WAN deployment conditions.

Investigating the Throughput Performance of the MPT-GRE Network Layer Multipath Library in Emulated WAN Environment

Nowadays there is a growing demand for a much faster and more secure communication without borders through the internet, which is provoking more and more both network designers and manufacturers of communication devices. Thanks to the BYOD trend, our communication devices can be easily carried anywhere in the world. They generally have several built-in network interfaces (e.g. Ethernet, Wi-Fi, 4G). Theoretically, using these cards in parallel, we could speed up data transmission, and thus communication, by aggregating the channel capabilities of the interfaces. On the other hand, we could make data transmission more secure by applying redundancy to the system. Unfortunately, traditional IP-based communications do not allow the use of parallel interfaces in a given communication session, leaving the hardware capabilities of our communications devices virtually untapped. To address this issue, we have developed a multipath communication solution called MPT-GRE, which we have already tested in several laboratory environments. The measurement results were published in our previous articles. In this paper we are going to test it in a much more realistic environment, using the Dummynet WAN emulation software. The measurement results confirmed that the MPT-GRE multipath solution is able to aggregate the performance of physical connections efficiently in the emulated Fast Ethernet IPv4 WAN environment as well.

IoTSim-SDWAN: A simulation framework for interconnecting distributed datacenters over Software-Defined Wide Area Network (SD-WAN)

Software-defined networking (SDN) has evolved as an approach that allows network administrators to program and initialize, control, change and manage networking components (mostly at L2-L3 layers) of the OSI model. SDN is designed to address the programmability shortcomings of traditional networking architectures commonly used in cloud datacenters (CDC). Deployment of SDN solutions have demonstrated significant improvements in areas such as flow optimization and bandwidth allocation in a CDC. However, the benefits are significantly less explored when considering Software-Defined Wide Area Networks (SD-WAN) architectures in the context of delivering solutions by networking multiple CDCs. To support the testing and bench-marking of data-driven applications that rely on data ingestion and processing (e.g., Smart Energy Cloud, Content Delivery Networks) across multiple cloud datacenters, this paper presents the simulator, IoTSim-SDWAN. To the best of our knowledge, IoTSim-SDWAN is the first simulator that facilitates the modeling, simulating, and evaluating of new algorithms, policies, and designs in the context of SD-WAN ecosystems and SDN-enabled multiple cloud datacenters. Finally, IoTSim-SDWAN simulator is evaluated for network performance and energy to illustrate the difference between classical WAN and SD-WAN environments. The obtained results show that SD-WAN surpasses the classical WAN in terms of accelerating traffic flows and reducing power consumption.

Improving End-Users Utility in Software-Defined Wide Area Network Systems

Software Defined Networks (SDNs) has brought a new form of network architecture that simplifies network management through innovations and programmability. But, the distributed control plane of SD-Wide Area Network is challenged by load imbalance problem due to the dynamic change of the traffic pattern. The packet_in messages are one of the major contributors of the control’s load. When such packet rate exceeds a certain threshold limit, the response time for control request increases non-linearly. In order to achieve better end-user experience, most of the previous works considered the optimal switch to controller association with an objective to minimize the response time on LAN environment but ignores the consequence of large scale network. In this regard, the proposed work realizes the necessity of layer-2 and layer-3 controller in LAN and WAN environment separately. A load prediction based alertness approach has been introduced to reduce the burden of the controllers. This approach may create an additional delay for the initial packets of the flow entry that lead to more prediction error. However, the proposed method reduces the error by selecting an optimal timeout value of the flow. Further, minimization of the response time between router to the controller has been taken care of. An extensive simulation shows the efficacy of the proposed scheme.

A novel framework for data acquisition and ubiquitous communication provisioning in smart cities

Abstract Gigantic progressions in heterogeneous communication technologies have empowered smart cities (SCs) to communicate with each other, while ensuring network connectivity. These innovations cannot give impeccable network availability in SCs, because of the coexistence of thousands of gadgets, which realize some issues. In this paper, to resolve the issue of ubiquitous communication provisioning, the MADM techniques are employed for network selection in proposed scenarios under phase 1. Data rate, transmission power, latency, cost, reliability, coverage range, channel bandwidth and security attributes of diverse available networks are utilized for optimal network selection. Cohesive AHP-Entropy weights of respective attributes are derived from subjective weights generated by AHP and objective weights calculated by Entropy. Cohesive weights are employed by TOPSIS algorithm for ranking of the networks available in the vicinity of level 1, 2 and 3 of the communication networks hierarchy. The network with the highest preference level has been selected for communication. The implementation of proposed criterion showed that 3G (HSDPA), WLAN (802.11p), Bluetooth (802.15.1), WiMAX (802.16), Wi-Fi haLow (802.11 ah), 4G have the highest preference level in proposed scenarios UPN, HN, HAN, CR, NAN/MAN and WAN environments respectively. The continuous monitoring of the connected networks and time-to-time selection of optimal networks is required for provisioning of ubiquitous communication in diverse applications (e.g. smart buildings and infrastructure services, smart traffic and transportation system, smart metering of electricity, water, gas and emergency calling services) of SCs. In phase 2, we propose a novel framework for data acquisition and provisioning of ubiquitous communication network in SCs. According to the suitability of applications, the data is acquired by proposed diverse acquisition techniques. Computing at diverse levels 1, 2 and 3 is described in phase 3 for provisioning of offloading, reducing delay and complexity. The overall framework resulted in real-time system scenario for smart cities in forthcoming days.

Statistical user behavior detection and QoE evaluation for thin client services

Remote desktop connection (RDC) services offer clients the ability to access remote content and services, commonly in the context of accessing their working environment. With the advent of cloud-based services, an example use case is that of delivering virtual PCs to users in WAN environments. In this paper, we aim to detect and analyze common user behavior when accessing RDC services, and use this as input for making Quality of Experience (QoE) estimations and subsequently providing input for effective QoE management mechanisms. We first identify different behavioral categories, and conduct traffic analysis to determine a feature set to be used for classification purposes. We propose a machine learning approach to be used for classifying behavior, and use this approach to classify a large number of real-world RDCs. We further conduct QoE evaluation studies to determine the relationship between different network conditions and subjective end user QoE for all identified behavioral categories. Results show an exponential relationship between QoE and delay and loss degradations, and a logarithmic relationship between QoE and bandwidth limitations. Obtained results may be applied in the context of network resource planning, as well as in making QoE-driven resource allocation decisions.

Heterogeneous Convergence Model of Intelligent Perception Network Using Fiber Optic Sensing

Fiber Bragg Grating sensor network plays an important role in the field of large-scale engineering and major equipment condition monitoring and diagnosis. The heterogeneous convergence between FBG sensor network and the Internet constitutes the FBG sensor network of Things which can future expand the network applications range and enhance the condition perception and monitoring capabilities. In this paper, the heterogeneous convergence model between FBG sensor network and the Internet is built. Based on the model, heterogeneous convergence principle and method between FBG sensor network and the Internet is introduced from three layers: perception layer, convergence processing layer and application layer. According to the theory and method, a condition monitoring and fault diagnosis system for the harbor machinery in the WAN environment is presented. The system physical running effect in the working field demonstrates the effectiveness and the rationality of the convergence method.

Optimizing Paxos with batching and pipelining

Paxos is probably the most popular state machine replication protocol. Two optimizations that can greatly improve its performance are batching and pipelining. However, tuning these two optimizations to achieve high-throughput can be challenging, as their effectiveness depends on many parameters like the network latency and bandwidth, the speed of the nodes, and the properties of the application. We address this question by presenting an analytical model of the performance of Paxos, and showing how it can be used to determine configuration values for the batching and pipelining optimizations that result in high-throughput. We then validate the model, by presenting the results of experiments where we investigate the interaction of these two optimizations both in LAN and WAN environments, and comparing these results with the prediction from the model. The results confirm the accuracy of the model, with the predicted values being usually very close to the ones that provide the highest performance in the experiments. Furthermore, both the model and the experiments give useful insights into the relative effectiveness of batching and pipelining. They show that although batching by itself provides the largest gains in all scenarios, in most cases combining it with pipelining provides a very significant additional increase in throughput, with this being true not only on high-latency WAN but also in many LAN scenarios.

A transport-layer approach for improving thin-client performance in a WAN environment

The performance of thin-client systems based on TCP depends on network quality, so it becomes worse in a WAN environment; however, the effects of TCP mechanisms have not been clarified. In this paper, we first describe the download traffic of thin-client systems as a two-state model with interactive data flows in response to keystrokes and bulk data flows related to screen updates. Since users are more sensitive to the keystroke response time, our next objective is to minimise the latency of interactive data flows, especially when the network is congested. Through detailed simulation experiments, we reveal that the main delays are queuing delay in the bottleneck router and buffering delay in the server. We then enhance two TCP mechanisms: retransmission timeout calculation and SACK control, which negate the negative impacts of existing options and increase the interval between occurrences of large delays by about four times.

A cost-effective load-balancing policy for tile-based, massive multi-core packet processors

Massive multi-core architectures provide a computation platform with high processing throughput, enabling the efficient processing of workloads with a significant degree of thread-level parallelism found in networking environments. Communication-centric workloads, like those in LAN and WAN environments, are fundamentally composed of sets of packets, named flows. The packets within a flow usually have dependencies among them, which reduce the amount of parallelism. However, packets of different flows tend to have very few or no dependencies among them, and thus can exploit thread-level parallelism to its fullest extent. Therefore, in massive tile-based multi-core architectures, it is important that the processing of the packets of a particular flow takes place in a set of cores physically close to each other to minimize the communication latency among those cores. Moreover, it is also desirable to spread out the processing of the different flows across all the cores of the processor in order to minimize the stress on a reduced number of cores, thus minimizing the potential for thermal hotspots and increasing the reliability of the processor. In addition, the burst-like nature of packet-based workloads render most of the cores idle most of the time, enabling large power savings by power gating these idle cores. This work presents a high-level study of the performance, power, and thermal behavior of tile-based architectures with a large number of cores executing flow-based packet workloads, and proposes a load-balancing policy of assigning packets to cores that minimizes the communication latency while featuring a hotspot-free thermal profile.

Communication Cost Reduction by Hierarchical Communication Pattern for FE Computation on Cluster-of-Clusters

With the rapid growth of the WAN infrastructures and enhancement of MPI libraries for WAN environment, Grid computing is expected to be a practical methodology for performing very large finite element analyses. In particular, the utilization of Grid enhanced MPI libraries allows one to execute legacy programs on WAN environments, which is very attractive for finite element application users. However, the high communication cost associated with WANs has discouraged its use as a parallel environment for finite element analysis. In this study, we focus on the utilization of Grid environments, in particular on cluster-of-clusters environments for the execution of FEA, and propose a hierarchical communication pattern to reduce the communication cost. By analytically estimating the communication cost, based on the decomposed FE mesh, we assess the effectiveness of the hierarchical communication pattern. We then show the obtained communication cost reduction by numerical experiments performed on cluster-of-clusters on WAN.

FaTLease: scalable fault-tolerant lease negotiation with Paxos

A lease is a token which grants its owner exclusive access to a resource for a defined span of time. In order to be able to tolerate failures, leases need to be coordinated by distributed processes. We present FaTLease, an algorithm for fault-tolerant lease negotiation in distributed systems. It is built on the Paxos algorithm for distributed consensus, but avoids Paxos' main performance bottleneck of requiring persistent state. This property makes our algorithm particularly useful for applications that can not dispense any disk bandwidth. Our experiments show that FaTLease scales up to tens of thousands of concurrent leases and can negotiate thousands of leases per second in both LAN and WAN environments.

An intermedia synchronisation mechanism for multimedia distributed systems

Distributed Multimedia Systems (DMS) deal with simultaneous geographically distributed sources by transmitting heterogeneous data. The preservation of temporal relations among different data types and simultaneous distributed sources is an open research area. This paper proposes a temporal synchronisation mechanism to be used at runtime in a DMS. One original aspect is that the present work avoids the use of a common reference by executing all multimedia temporal relations according to their causal dependencies. The mechanism is emulated considering a WAN environment and using MPEG-4 encoders. The results show that our mechanism is effective in reducing the intermedia synchronisation error.

Provider link state bridging

Wide area communications technology has been challenged to virtualize large numbers of Ethernet LAN segments. This is a consequence of a mismatch between the broadcast nature of the LAN segment and the extremely constrained connectivity implied by the p2p connections or tunnels available in the WAN environment, which have been combined to create virtual LAN segments. PBB-TE has been a practical demonstration of how filtering applied to a broadcast media can result in a connection. This article introduces provider link state bridging (PLSB), which adds a control plane to the PBB data plane in order to extend the techniques for manipulation of Ethernet bridges for unicast paths pioneered by PBB-TE. PLSB solves the problem of largescale virtualization of LAN segments over an Ethernet infrastructure by applying advances in computation performance to the multicast capabilities inherent in the Ethernet data plane. The result is that the fundamental primitives of connectivity today, the broadcast LAN segment and the connection, can be virtualized in a scalable manner on a common (but substantially larger and better utilized) Ethernet-based infrastructure.

SCALABLE CATALOG INFRASTRUCTURE FOR MANAGING ACCESS COSTS AND SOURCE SELECTION IN WIDE AREA NETWORKS

A WAN environment, such as the Internet, connects a federation of hundreds of servers with tens of thousands of clients, which poses a substantial scalability challenge. Clients may choose among sources that vary in both their content and quality as well as in their access latencies. At the same time, Internet accessible data sources exhibit transient behavior; the unpredictable behavior of a dynamic WAN results in a wide variability in access cost (end-to-end latency). This motivates a need for a source selection strategy that requires maintaining access cost distributions (latency profiles) for each client/server pair. However, in the presence of hundreds of servers and thousands of clients, managing latency profiles cannot scale. We present a scalable methodology to manage latency profiles that use non-random associations between client/server pairs. Such non-random associations may be identified by topology-independent measures such as correlation and mutual information. We propose a Catalog infrastructure that implements our methodology and utilize non-randomly associated latency profiles to estimate access cost distribution for client/server pairs. We perform an extensive experimental study demonstrating feasibility and efficiency of our approach.

THINC

Rapid improvements in network bandwidth, cost, and ubiquity combined with the security hazards and high total cost of ownership of personal computers have created a growing market for thin-client computing. We introduce THINC, a virtual display architecture for high-performance thin-client computing in both LAN and WAN environments. THINC virtualizes the display at the device driver interface to transparently intercept application display commands and translate them into a few simple low-level commands that can be easily supported by widely used client hardware. THINC's translation mechanism efficiently leverages display semantic information through novel optimizations such as offscreen drawing awareness, native video support, and server-side screen scaling. This is integrated with an update delivery architecture that uses shortest command first scheduling and non-blocking operation. THINC leverages existing display system functionality and works seamlessly with unmodified applications, window systems, and operating systems.We have implemented THINC in an X/Linux environment and compared its performance against widely used commercial approaches, including Citrix MetaFrame, Microsoft RDP, GoToMyPC, X, NX, VNC, and Sun Ray. Our experimental results on web and audio/video applications demonstrate that THINC can provide up to 4.8 times faster web browsing performance and two orders of magnitude better audio/video performance. THINC is the only thin client capable of transparently playing full-screen video and audio at full frame rate in both LAN and WAN environments. Our results also show for the first time that thin clients can even provide good performance using remote clients located in other countries around the world.

Next-generation optical storage area networks: the light-trails approach

Multiple architectures, technologies, and standards have been proposed for storage area networks, typically in the WAN environment. The transport aspect of storage signifies that optical communications is the key underlying technology. The contemporary SAN over optical network concept uses the optical layer for pure transport with minimal intelligence. This leads to high cost and overprovisioning. Future optical networks, however, can be expected to play a role in optimizing SAN extension into the WAN. An essential characteristic of SAN systems is tight coupling between nodes in a SAN network. Nodes in a SAN system have two critical functions that are presently emulated by data layers and can be offloaded to the optical layer. First, nodes need to signal among each other to achieve tasks such as synchronous and asynchronous storage. Second, to benefit from an optimized network, nodes need to allocate bandwidth dynamically in real time. In this article we show how the optical layer can be furthered from just pure transport to creating opportunities in provisioning as well as providing the mirroring function of SAN systems (multicasting) and consequently lead to reduction in cost. We demonstrate that the light-trail model is one way of efficiently utilizing the optical layer for SAN.

Parallel LAN/WAN heuristics for optimization

We present in this work a wide spectrum of results on analyzing the behavior of parallel heuristics (both pure and hybrid) for solving optimization problems. We focus on several evolutionary algorithms as well as on simulated annealing. Our goal is to offer a first study on the possible changes in the search mechanics that the algorithms suffer when shifting from a LAN network to a WAN environment. We will address six optimization tasks of considerable complexity. The results show that, despite their expected slower execution time, the WAN versions of our algorithms consistently solve the problems. We report also some interesting results in which WAN algorithms outperform LAN ones. Those results are further extended to analyze the behavior of the heuristics in WAN with a larger number of processors and different connectivities.

Parallel LAN/WAN heuristics for optimization

We present in this work a wide spectrum of results on analyzing the behavior of parallel heuristics (both pure and hybrid) for solving optimization problems. We focus on several evolutionary algorithms as well as on simulated annealing. Our goal is to offer a first study on the possible changes in the search mechanics that the algorithms suffer when shifting from a LAN network to a WAN environment. We will address six optimization tasks of considerable complexity. The results show that, despite their expected slower execution time, the WAN versions of our algorithms consistently solve the problems. We report also some interesting results in which WAN algorithms outperform LAN ones. Those results are further extended to analyze the behavior of the heuristics in WAN with a larger number of processors and different connectivities.