LPM-Based NLI Estimation Over Heterogeneous Links: Analysis and Comparison

Scheduling precedence constrained stochastic tasks on heterogenous cluster systems is an important issue which impact the performance of clusters significantly. Different with deterministic tasks, stochastic task model assumes that the workload of task and quantity of data transmission between tasks are stochastic variables, which is more realistic than other task models. Scheduling model and algorithms of precedence constrained stochastic tasks attract a large number of researchers' attention recently. An algorithm SDLS Stochastic Dynamic Level Scheduling has been proved performing well in scheduling stochastic tasks on heterogenous clusters. However, the assumption about communication time between tasks in SDLS is much simpler than its assumptions about task computing time, which makes it cannot depict the communication cost among heterogenous links well. In this paper, it is assumed that the quantity of data communication between tasks is a stochastic variable of normal distribution, instead of assuming communication time among heterogenous links a same stochastic variable immediately. Moreover, a modified scheduling model and algorithm SDLS-HC Stochastic Dynamic Level Scheduling on Heterogenous Communication links are proposed. Work in this paper focus on considering much more detailed communication cost in task scheduling based on SDLS. Evaluation on many random generated tasks experiments demonstrates that SDLS-HC achieves better performance than SDLS on cluster systems with heterogenous links.

Aggregating the bandwidth of multiple low-cost physical communication links to form a single, higher capacity logical link is a common and inexpensive approach to increasing network performance. Aggregating or bonding links has traditionally been used in relatively static settings such as LAN-WAN interconnection, where the component links to be aggregated were typically identical (e.g., multiple T1 links) with very similar communication characteristics. However the emergence of multi-homed hosts and diverse access technologies (e.g., cable modem, wireless DSL) compel us to study the performance of aggregating the bandwidth of increasingly heterogeneous links.In this paper we examine the end-to-end performance of a single TCP connection with packets transmitted over multiple heterogeneous links. We describe a testbed we constructed, present empirical results, and introduce a new Linux-based traffic scheduler called wTEQL that we have developed to optimize the performance of TCP over multiple heterogeneous communication links. Finally, we present a novel analytical approach based on linear systems theory to characterize the performance of TCP over an inverse multiplexed channel with WRR scheduling over heterogeneous links. We believe that our model is the first to demonstrate analytically how TCP throughput on two links with a poorly matched packet scheduling policy can fall below the achievable throughput of the slower of the two links when that link is used alone.1

Communication over geostationary satellite links is improved by introducing end-to-end Forward Error Correction (FEC) and simultaneous transmissions over two links (channel bundling). The main objective of this work is to investigate to which degree the goodput and the reliability can be enhanced using the mentioned techniques. The performances of the two FEC schemes Reed-Solomon Codes and Random Linear Network Coding are compared. Uncorrelated and correlated packet errors are considered, the latter with a Gilbert-Elliot channel model. Experiments are conducted in a testbed consisting of a single PC with virtual network interfaces to determine the influence of various parameter settings on performance. Results are compared against a scenario with one link offering the same capacity as the two links together. It is concluded that using two heterogeneous links is beneficial for the goodput and losses for generation sizes larger than 20 for three correlated lost packets on average.

The satellite data relay networks (SDRNs) hold great promise in 6G communications for the timely offloading of the global traffic. Since the delay performance is regarded as one of the most important metrics reflecting the offloading efficiency, studying its relationship with network parameters becomes really essential to the development and application of the SDRN. However, the complex data offloading process and heterogeneity of traffic arrivals and transmission links pose many challenges to the stochastic delay analysis. To accurately model the data offloading process in SDRNs, we build a series-parallel queuing model with through and cross traffic while considering the propagation delay. On this basis, we respectively propose a propagation delay embedded min-plus convolution method based on stochastic network calculus and a Markov chain method based on Monte Carlo to depict the leftover services of the heterogeneous links received by the per-flow traffic in an aggregate. To eliminate the impacts of the heterogeneity, we uniformly characterize the arrivals and leftover services by their moment generating functions (MGFs) which contain the full moment information, and shield the heterogeneity by deriving the envelopes of the arrivals and leftover services with the help of MGFs, Chernoff bound and union bound. Then, in the light of the geometric relationship between the envelopes of the arrivals and leftover services, we analyze the upper bounds of the stochastic delay, which provides the guidance to the network configuration. Eventually, simulation results verify the effectiveness of the theoretical analysis and further reveal maximum four times the delay difference between the heterogeneous links influenced by traffic type, burstiness, and access number.

We present a novel feedback protocol for wireless broadcast networks that use linear network coding. We consider transmission of packets from a single source to many receivers over a single-hop broadcast erasure channel with heterogeneous links. We propose a predictive model to minimize feedback as well as extraneous data transmissions by the source. The predictive model will schedule feedbacks only when there is a significant probability that all receivers have completed the download. We demonstrate analytically as well as empirically that reliable multicast can provide a good completion time characteristic for all users, if the initial feedback is delayed until the expected download completion time. We show that with SMART, counter to conventional wisdom the average user's download completion time improves slightly as the number of users increases. Furthermore, we show that completion time of the worst user in a multicast session is not very sensitive to the number of users; however it is very sensitive to imbalanced effective rate and heterogeneity among users. Moreover, we show that SMART performs nearly as well as an omniscient transmitter that requires no feedback.

We describe an approach to improving network performance in networks with heterogeneous links - for example, a communications network constructed from satellite links, airborne relays, and line-of-site radios - using disruption tolerant networking. DTN defines an overlay network that uses storage at each hop to make end-to-end communications resilient to delay, loss, network disruption, temporary network partitioning, and congestion. DTN further improves performance by allowing separate transport layers at each hop of the overlay that can be tuned for the local network environments. We present recent work interfacing one such transport layer, the NACK oriented reliable multicast (NORM) protocol, to DTN; performance experiments from tactical radios; and simulation results based on models of deployed systems.

Bonding multiple low-cost physical communication links to form a single, higher capacity logical link is a common and inexpensive approach to increasing WAN access bandwidth. Though bonded links have traditionally been homogeneous, the emergence of multi-homed hosts and diverse access technologies (e.g., cable modem, DSL, connection sharing via 802.11x) compel us to study the performance of bonding increasingly heterogeneous links. We examine the end-to-end performance of a single TCP connection with packets transmitted over multiple links. We describe a testbed we constructed, and present throughput results for bonded links with heterogeneous component links

An Analysis of Speeding Multicast by Acknowledgment Reduction Technique (SMART) with Homogeneous and Heterogeneous Links - A Method of Types Approach

A network social interaction model with heterogeneous links

Macroscopic traffic flow models have been widely used to provide reliable prediction inputs for active transportation and demand management (ATDM). The representative models include METANET, FREFLOW, and KRONOS. In recent development of ATDM, METANET-based prediction models have started to more attention with their convenience of formulating several important macroscopic control variables, such as speed limit and ramp flow. However, applying METANET models on real-world observations (e.g., loop detector data) may experience significant performance issues, especially when applied to a corridor with heterogeneous links. Heterogeneous links refer to adjacent links that have different macroscopic traffic flow characteristics such as free flow speed and capacity. The heterogeneity is usually caused by changes in road geometry and environment, which commonly exist in urban freeway networks. This article provides an in-depth study to investigate the error mechanism of this problem and proposes several effective methods to improve the performance of METANET model over heterogeneous links. It is found in this article that applying some dynamic formulations to the convention term of the speed equation in METANET can significantly reduce the impact of link heterogeneity issue. The proposed method was evaluated using field loop detector data collected at I-894 corridor in Milwaukee, WI.

In this paper, we have proposed an extended version of MPTCP with heterogeneous communication links to solve the out-of-order packet problem. We have confirmed that the proposed method can improve the end-to-end throughput performance up to 72.7% when RTTs of subflows are different. In addition, we have also confirmed that the proposed method does not affect the end-to-end throughput performance in the situation when RTTs of subflows are almost the same.

Based on the requirement to ensure the efficient and reliable operation of power wireless heterogeneous network, the intelligent management strategy of power wireless heterogeneous network link based on traffic balance is proposed. The strategy consists of four processes which are traffic collection process, traffic prediction analysis process, traffic diversion process and traffic balance management process. The algorithm of traffic balancing degree of the whole network is put forward, which is conducive to balancing the network traffic by adjusting the threshold of the link. This strategy has been widely used in many areas, which effectively improves the efficiency of link intelligent management in power wireless heterogeneous network.

Link prediction is a fundamental task in statistical analysis of network data. Though much research has concentrated on predicting entity-entity relationships in homogeneous networks, it has attracted increasing attentions to predict relationships in heterogeneous networks, which consist of multiple types of nodes and relational links. Existing work on heterogeneous network link prediction mainly focuses on using input features that are explicitly extracted by humans. This paper presents an approach to automatically learn latent features from partially observed heterogeneous networks, with a particular focus on entity-attribute networks (EANs), and making predictions for unseen pairs. To make the latent features discriminative, we adopt the max-margin idea under the framework of maximum entropy discrimination (MED). Our maximum entropy discrimination joint relational model (MED-JRM) can jointly predict entity-entity relationships as well as the missing attributes of entities in EANs. Experimental results on several real networks demonstrate that our model has improved performance over state-of-the-art homogeneous and heterogeneous network link prediction algorithms.

Bandwidth aggregation of heterogeneous wireless links faces numerous challenges in maximizing available connectivity services for mobile applications. One serious issue is the out-of-sequence arrival of packets at receiver that is amplified by distribution of packets of a flow over multiple paths. These packets experience asymmetric path characteristics that results in diverse congestion states for traffic flow. This paper presents an adaptive mechanism for estimation of path suitability for a specific type of flow on the basis of end-to-end path statistics of delay, packet drop and available bandwidth. The proposed mechanism is also supported by a stochastic model that estimates delay with lesser computational complexity. The results of proposed approach have been validated through simulation as well. The results have shown robust performance of proposed mechanism in achieving acceptable quality-of-service levels during mobility with minimized out-of-sequence reception and reduced buffer occupancy.

Recently, broadband access service is desired even in mobile environment. However, Internet connection service is provided in only a small number of railways because it is difficult to continuously reserve enough bandwidth by single medium. In this paper, we propose a method to utilize heterogeneous wireless links, one of which provides continuous slow connection and the others provide intermittent fast connections. Simulation results show that sufficient quality is achievable by predictive scheduling by proxy servers both on the ground and the train.