The spatial structure of base stations (BSs) in cellular networks plays a key role in evaluating the downlink performance. In this paper, different spatial stochastic models (the Poisson point process (PPP), the Poisson hard-core process (PHCP), the Strauss process (SP), and the perturbed triangular lattice) are used to model the structure by fitting them to the locations of BSs in real cellular networks obtained from a public database. We provide two general approaches for fitting. One is fitting by the method of maximum pseudolikelihood. As for the fitted models, it is not sufficient to distinguish them conclusively by some classical statistics. We propose the coverage probability as the criterion for the goodness-of-fit. In terms of coverage, the SP provides a better fit than the PPP and the PHCP. The other approach is fitting by the method of minimum contrast that minimizes the average squared error of the coverage probability. This way, fitted models are obtained whose coverage performance matches that of the given data set very accurately. Furthermore, we introduce a novel metric, the deployment gain, and we demonstrate how it can be used to estimate the coverage performance and average rate achieved by a data set.

/pdf/spatial-stochastic-models-and-metrics-for-the-structure-of-4hb8fo9zja.pdf

Spatial Stochastic Models and Metrics for the Structure of Base Stations in Cellular Networks

Being able to identify the groups of clients that are responsible for a significant portion of a Web site's requests can be helpful to both the Web site and the clients. In a Web application, it is beneficial to move content closer to groups of clients that are responsible for large subsets of requests to an origin server. We introduce clusters---a grouping of clients that are close together topologically and likely to be under common administrative control. We identify clusters using a ``network-aware" method, based on information available from BGP routing table snapshots.

On Network-Aware Clustering of Web Clients

The recent increasing interest in ad hoc networks has motivated the study and development of interference models capable of capturing the intrinsic characteristics of this kind of network, such as the lack of a central coordination and the consequent distributed nature of some of the network functions (e.g., medium access control and routing protocols). As a consequence, a myriad of interference models for ad hoc networks can be found in the literature, that describe the effects of interference at different layers and with different levels of detail. Some of these models describe interference as a random process whose statistics depend mainly on physical layer parameters, and are better suited for the analysis of issues related to the physical layer. Other models focus on the effects of interference at higher layers, and are more appropriate when the analysis of network-related issues is of interest. This paper presents a comprehensive survey of interference models for wireless ad hoc networks, emphasizing their domains of application, illustrated with examples. The interference models are presented in this survey from the perspective of the radio capture phenomenon, resulting in a unified view of interference models, which may be helpful when selecting the appropriate model for a given purpose.

Modeling Interference in Wireless Ad Hoc Networks

Continuum percolation: References

Matern hard core processes of types I and II are the point processes of choice to model concurrent transmitters in CSMA networks. We determine the mean interference observed at a node of the process and compare it with the mean interference in a Poisson point process of the same density. It turns out that despite the similarity of the two models, they behave rather differently. For type I, the excess interference (relative to the Poisson case) increases exponentially in the hard-core distance, while for type II, the gap never exceeds 1 dB.

/pdf/mean-interference-in-hard-core-wireless-networks-4wmbvcljn1.pdf

Mean Interference in Hard-Core Wireless Networks

As interference results from the summation of signals issued by concurrent transmitters, it directly depends on the transmitters location. The point process used to model concurrent transmitters location is thus fundamental in any multi-hop wireless network study. In this paper, we investigate original processes that comply with the CSMA/CA policies. We propose the use of the Simple Sequential Inhibition point process to model CSMA/CA networks where carrier detection depends on the strongest emitter only. We then extend this point process to a family of new point processes modeling a busy medium detection based on the strength of all concurrent signals. We study the impact of these different processes on the interference distribution and show that the results are very different with respect to the different point process. We finally show that the use of a Poisson process is generally inaccurate to model CSMA/CA networks.

Point processes for interference modeling in CSMA/CA ad-hoc networks

In analytical studies of multi-hop wireless networks, the spatial distribution of transmitters is typically modeled using a homogeneous Poisson point process. In this report, we show why such a modeling is inaccurate and leads to an inappropriate interference distribution in the case of CSMA/CA networks. We then study a more realistic model, the Matern point process, which still reveals some unexpected flaws such as an under-estimation of the transmitters density. To get round these limitations, we propose the use of an alternate model, referred to as the \emph{Simple Sequential Inhibition} (SSI) point process, which we assert being a valuable and more appropriate model for CSMA/CA networks. We present some analytical results on the Matern and the SSI model and study by simulation the interference distribution resulting from the different point processes.

https://hal.inria.fr/inria-00316029/document

Interference Modeling in CSMA Multi-Hop Wireless Networks

Flat ad hoc architectures are not scalable. In order to overcome this major drawback, hierarchical routing is introduced since it is found to be more effective. The main challenge in hierarchical routing is to group nodes into clusters. Each cluster is represented by one cluster head. Conventional methods use either the connectivity (degree) or the node Id to perform the cluster head election. Such parameters are not really robust in terms of side effects. In this paper we introduce a novel measure that both forms clusters and performs the cluster head election. Analytical models and simulation results show that this new measure for cluster head election induces less cluster head changes as compared to classical methods.

https://hal.archives-ouvertes.fr/hal-00383714/document

Self-organization in large scale ad hoc networks

In this paper, we study the impact of collisions and interference on a neighbor discovery process in the context of multi-hop wireless networks. We consider three models in which interference and collisions are handled in very different ways. From an ideal channel where simultaneous transmissions do not interfere, we derive an alternate channel where simultaneous transmissions are considered two-by-two under the form of collisions, to finally reach a more realistic channel where simultaneous transmissions are handled under the form of shot-noise interference. In these models, we analytically compute the link probability success between two neighbors as well as the expected number of nodes that correctly receive a Hello packet. Using this analysis, we show that if the neighbor discovery process is asymptotically equivalent in the three models, it offers very different behaviors locally in time. In particular, the scalability of the process is not the same depending on the way interference is handled. Finally, we apply our results to the dimensioning of a Hello protocol parameters. We propose a method to adapt the protocol parameters to meet application constraints on the neighbor discovery process and to minimize the protocol energy consumption.

https://hal.inria.fr/inria-00384147/document

Neighbor discovery in multi-hop wireless networks: evaluation and dimensioning with interferences considerations

In this paper, we are interested in the properties of vehicular ad hoc networks (VANET). We study the mean connection lifetimes to evaluate the feasibility to support non-safety applications. Because connection lifetimes are in the order of a few minutes, an infrastructure network is highly recommended to maintain connectivity between car clusters. Using an hybrid network, we perform comparisons of routing protocols used in the VANET. We show that reactive and geographic routing protocols are very close in terms of performance. However the reactive protocol are more accurate since it does not require neither geolocalization systems nor location services. We propose the use of hybrid networks and we give dimensioning rules to ensure connectivity of the network

Anthony Busson

Papers

Point processes for interference modeling in CSMA/CA ad-hoc networks

Interference Modeling in CSMA Multi-Hop Wireless Networks

Self-organization in large scale ad hoc networks

Neighbor discovery in multi-hop wireless networks: evaluation and dimensioning with interferences considerations

Inside VANET: Hybrid Network Dimensioning and Routing Protocol Comparison