(19)
(11)EP 2 850 867 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
04.12.2019 Bulletin 2019/49

(21)Application number: 14753755.9

(22)Date of filing:  19.02.2014
(51)Int. Cl.: 
H04W 16/18  (2009.01)
H04W 24/02  (2009.01)
H04W 40/24  (2009.01)
(86)International application number:
PCT/US2014/017019
(87)International publication number:
WO 2014/130496 (28.08.2014 Gazette  2014/35)

(54)

MOBILE BACKHAUL TOPOLOGY PLANNING AND/OR OPTIMIZATION

MOBILE BACKHAUL-TOPOLOGIEPLANUNG UND/ODER -OPTIMIERUNG

PLANIFICATION ET/OU OPTIMISATION DE TOPOLOGIE DE RACCORDEMENT MOBILE


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(30)Priority: 20.02.2013 US 201361766981 P
18.02.2014 US 201414182479

(43)Date of publication of application:
25.03.2015 Bulletin 2015/13

(73)Proprietor: NEC Corporation
108-8001 Tokyo (JP)

(72)Inventor:
  • CVIJETIC, Neda
    Plainsboro, New Jersey 08536 (US)

(74)Representative: Betten & Resch 
Patent- und Rechtsanwälte PartGmbB Maximiliansplatz 14
80333 München
80333 München (DE)


(56)References cited: : 
KR-A- 20090 076 675
US-A1- 2012 195 219
US-A1- 2008 008 093
US-A1- 2012 294 179
  
  • JULIUS ROBSON: 'Small Cell Backhaul Requirements' NGMN ALLIANCE 04 June 2012, XP055174163 Retrieved from the Internet: <URL:https://www.ngmn.org/uploads/media/ NGMN_Whitepaper_Small_ Cell _Backhaul_Requirements.pdf>
  
Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


Description

BACKGROUND OF THE INVENTION



[0001] The present invention relates to mobile backhaul (MBH) topology and, more particularly, to MBH transmission topology planning and/or optimization.

[0002] Due to the massive proliferation of smart mobile devices, the mobile backhaul (MBH) and access networks of many leading global operators are evolving on a rapid, almost day-to-day basis. Due to the heterogeneity of MBH technologies (e.g. point-to-point microwave, copper, optical fiber), topologies within a given technology (e.g. line-of-sight versus non-line-of-sight wireless; free-space optical; point-to-point versus passive optical network (PON) fiber; etc..), and very different operator-specific legacy network investments and architectures, the optimization and future-proofing of the emerging high-speed MBH network emerges as a very difficult problem. We attempt to solve this problem by translating it into a malleable data rate optimization problem that incorporates physical layer parameters and can be solved in a software-defined way on a per-case, per-operator basis, and can thus act as both an a priori network planning tool, as well as a posteriori network upgrade/optimization tool.

[0003] Previous studies have considered optimal wireless access base station site placement in networks with randomly distributed mobile users [1-3], assuming that the wireless access points are either not connected or are a priori interconnected with a backhaul infrastructure. However, the optimization of the backhaul infrastructure itself was not considered; it was simply taken to be either present a priori, or absent altogether. In [4], a highly theoretical study arguing for a posteriori backhaul infrastructure optimization was presented (i.e. building a backhaul infrastructure after learning mobile user traffic patterns and distribution), but such an optimization was not performed, and the theoretical study was moreover not related to concrete physical-layer parameters. In most recent studies [5], physical-layer parameters have likewise not been explicitly taken into account.

References



[0004] 
  1. [1] S. Toumpis and A. J. Goldsmith, "Capacity regions for wireless ad hoc networks," IEEE Trans. Wireless Commun., vol. 2, pp. 189-205, July 2003.
  2. [2] M. Franceschetti, O. Dousse, D. Tse, and P. Thiran, "Closing the gap in the capacity of wireless networks via percolation theory," IEEE Trans. Inf. Theory, vol. 53, pp. 1009-1018, March 2007.
  3. [3] P. Li, C. Zhang, and Y. Fang, "Capacity and delay of hybrid wireless broadband access networks," IEEE J. Sel. Areas Commun., vol. 27, no. 2, pp. 117-125, Feb. 2009.
  4. [4] A. Reznik, S. R. Kulkarni, and S. Verdu, "A small world approach to heterogeneous networks," Commun. Inf. Systems, vol. 3, pp. 325-348, September 2003.
  5. [5] M. Xia, Y. Owada, M. Inoue, H. Harai, "Optical and Wireless Hybrid Access Networks: Design and Optimization," OSA Jnl. Opt. Commun. Netw. (JOCN), vol. 4, no. 10, Oct. 2012.
  6. [6] N. Yoshimoto, "Next-Generation Access for Mobile Backhaul Application," Proc. The 17th OptoElectronics and Communications Conference (OECC 2012), paper 6A1-1.
  7. [7] N. Cvijetic, A. Tanaka, M. Cvijetic, Yue-Kai Huang, E. Ip, Y. Shao, and T. Wang, "Novel Optical Access and Digital Processing Architectures for Future Mobile Backhaul," J. Lightwave Technol., vol. 31, No. 4, pp. 621-627, Feb. 2013.


[0005] The document US 2012/195219 A1 describes an apparatus and method for precoding by limited cooperation in a wireless communication system. A method of data precoding by a Base Station (BS) in a wireless communication system includes determining a Channel Quality Indicator (CQI) parameter indicating channel size information between the BS and a User Equipment (UE) served by the BS; transmitting the determined CQI parameter through a backhaul to a central unit; receiving, from the central unit, a weight factor based on a CQI parameter from each cell; and determining a precoding vector based on the weight factor. The weight factor represents a relative ratio of an inter-cell environment that is to be interfered by the BS.

[0006] The document: JULIUS ROBSON, "Small Cell Backhaul Requirements", NGMN ALLIANCE, (20120604), XP055174163; relates to small cell backhaul requirements.

BRIEF SUMMARY OF THE INVENTION


The invention is defined in independent claims 1, 11, 12.



[0007] An objective of the present invention is to solve the problem of optimization and future-proofing of the emerging high-speed MBH network.

[0008] An aspect includes a method implemented in a network controller used in a mobile backhaul (MBH) network. The method comprises receiving one or more physical layer parameters from MBH transceivers in the MBH network, computing an admissible set of rate matrices, computing a capacity region based on the admissible set of rate matrices, deciding MBH transmission topology, and transmitting the decision to the MBH transceivers.

[0009] Another aspect includes a network controller used in a mobile backhaul (MBH) network. The network controller comprises a reception means for receiving one or more physical layer parameters from MBH transceivers in the MBH network, a first computing means for computing an admissible set of rate matrices, a second computing means for computing a capacity region based on the admissible set of rate matrices, a decision means for deciding MBH transmission topology, and a transmission means for transmitting the decision to the MBH transceivers.

[0010] Still another aspect includes a mobile backhaul (MBH) network comprising a network controller, MBH transceivers connected to the network controller, wherein the network controller receives one or more physical layer parameters from the MBH transceivers in the MBH network, computes an admissible set of rate matrices, computes a capacity region based on the admissible set of rate matrices, decides MBH transmission topology, and transmits the decision to the MBH transceivers.

[0011] Means herein can comprise one or more of various kinds of components such as, for example, software, a computer program, an electronic device, a computer, and/or a dedicated controller.

BRIEF DESCRIPTION OF THE DRAWINGS



[0012] 

Fig. 1 depicts a software-defined MBH network optimization architecture (a) and two possible optimized topology options (b) and (c). SD = software-defined; MS = mobile station; Tx = transmitter; Rx = receiver.

Fig. 2 depicts a detailed operational block diagram of the software-defined network optimizer (100).

Fig. 3 depicts SD throughput optimization architecture for optical MBH.

Fig. 4 depicts per-cell throughput rate, rmax, versus bandwidth, Wij (MHz); Case I: pre-aggregation = omni-directional wireless, aggregation = pt-to-pt wireless (θij=θji=30°); Case II: pre-aggregation = pt-to-pt wireless (θij=θji=30°), aggregation = pt-to-pt wireless (θij=θji=15°); Case III: pre-aggregation = pt-to-pt wireless (θij=θji=30°), aggregation = 1Gb/s fiber; Case IV: pre-aggregation = pt-to-pt wireless (θij=θji=30°), aggregation = 2.5Gb/s fiber; Case V: pre-aggregation = pt-to-pt wireless (θij=θji=30°), aggregation = 10Gb/s fiber; Case VI: pre-aggregation = pt-to-pt wireless (θij=θji=15°), aggregation = 2.5Gb/s fiber.


DETAILED DESCRIPTION



[0013] Unlike in previous work, the problem is solved for practical cases with finitely-many network nodes, while also including physical parameters relevant to a typical MBH deployment which can include both wireless (microwave and free-space optical) and wired (e.g. fiber) links. Specifically, the MBH network planning problem is defined as a linear optimization problem in which the achievable capacity regions and/or uniform capacity of the network are computed for the adopted set of transmission protocols. Such an optimization framework enables the direct inclusion of physical-layer parameters into MBH network planning, while remaining sufficiently generic and malleable to cover a wide range of practical cases. It provides a software-defined way to both plan the MBH network topology prior to deployment, optimize and upgrade already existing networks, and gain insight into data rate scaling for future-proofing practical systems.

[0014] The solution enables the computationally-efficient, physically accurate, and software-reconfigurable determination of a heterogeneous MBH network topology which, out of all possible admissible topology combinations, enables the highest backhaul data rate in the network. Prior art focused either on all-wireless solutions and neglected optimization of the backhaul topology, and/or considered highly-theoretical cases with abstracted physical layer characteristics. The SD operation can enable cost-efficient, flexible MBH network planning, optimization and upgrades that can be readily customized to the network scenario at hand by the direct inclusion of relevant physical-layer parameters into the data rate maximization problem.

[0015] The SD MBH network architecture optimization approach is illustrated in Fig.1(a): at the core of the approach is the SD MBH network optimizer (100) or controller, including a software-defined algorithm used to plan and optimize the network architecture in order to maximize the data rate that the MBH network, i.e. MBH transceivers (Transmitter (Tx)/Receiver (Rx)) (300)-(700) in Fig. 1(a), can deliver to a set of mobile stations (MS), i.e. users, denoted by (800)-(810) in Fig. 1(a), given a fixed set of underlying transmission protocols that govern MBH Tx/Rx behavior (e.g. time division multiple access (TDMA)). Both the information about the physical-layer specifics of each MBH Tx/Rx (300)-(700) and the resulting decisions of the SD MBH network optimizer (100) are communicated from/to the remote MBH transceivers (Tx/Rx) (300)-(700) via generic parallel control messaging (201), for example using the OpenFlow protocol (200). A point addressed by the SD MBH network optimizer is that, depending on the type of the MBH links interconnecting the MBH Tx/Rx (300)-(700), i.e. fiber vs. point-to-point microwave, for example, a different route might be selected to maximize the MBH data rate between two users, e.g. MS (800) and MS (810). This is illustrated in Figs. 1(b) and 1(c), where the bold solid arrows denote the active OpenFlow control messaging (201) between the SD network optimizer (100) and the MBH Tx/Rx (300), and the dashed arrows denote the MBH transmission topology selected to maximize the data rate between MS (800) and MS (810), as decided by the SD network optimizer (100) based on the specific physical-layer conditions and transmission protocol requirements at hand. It is noted that both options in Fig. 1(b) and Fig. 1(c) contain four hops between the source and destination MSs, (800) and (810) respectively, highlighting the inability of generic shortest-path algorithms to discriminate between physical-layer heterogeneity features of the network. In addition to thus maximizing the MBH data rate between MS source-destination pairs in a way that accounts for physical-layer specifics of the network, the SD controller can also be used to answer the following network planning, optimization, and upgrade questions:
  • If there are no wired (e.g. fiber) links in the network, where should these be placed? If fiber links exist, are they being used in a maximally efficient way? If not, what route/approach would be more efficient (i.e. resulting in a higher backhaul data rate)?
  • Given a legacy network topology, are there more optimal ways to route MBH traffic so as to increase the supported backhaul data rate between users?


[0016] It is also noted that, by addressing the second question above, in this approach, the MBH transmission topology selected by (100) is allowed to change as the traffic demands and patterns of the MSs (800)-(810) change; on the physical layer, this can be supported by point-to-point microwave or free-space optical links with adaptive directivity and spatial orientation. For wired links (e.g. fiber), topology morphing to accommodate changing network conditions is not feasible, such that the wired link subset would be exempt from topology re-optimization (i.e. rate matrices corresponding to such scenarios would be excluded from the capacity region computation outlined below; it is this computation that would determine the optimal MBH transmission topology). Finally, it is noted that while the illustrative case in Fig. 1(a) assumes the most flexible and general scenario in which no fixed a priori links exist between MBH Tx/Rx (300)-(700), i.e. that they can all be configured according to the decisions of the SD MBH network optimizer (100), this scenario can readily be constrained to cases with legacy fixed links, without loss of generality, by simply constraining the set of admissible rate matrices for the network as briefly described above, and as will be discussed and shown below.

[0017] Fig. 1 shows the SD MBH network optimizer (100), which operates by computing the capacity region of the underlying network given its physical-layer link parameters and set of transmission protocols. Specifically, the capacity region of a network is defined using a set of rate matrices that capture all needed information about the network activity at a given time. Each rate matrix is a function of an underlying transmission scheme that specifies which nodes are transmitting/receiving (Tx/Rx) whose data, and at what rate. The notation (i, k, j) expresses this succinctly, where (i) denotes the source, (k) denotes the relay, and (j) denotes the destination, respectively.

[0018] Thus, given a network of n nodes and a transmission scheme S, where S is defined as the full description of information flow between different nodes in the network at a given point in time (i.e. all transmit-receive node pairs in operation at the given time instant and their transmission rate), its rate matrix R(S) is a n × n matrix populated by entries data rate entries, rij, such that: rij = rij if j receives information at rate rij from i; rij = -rij if j transmits information at rate rij from i; and otherwise rij = 0.

[0019] Moreover, to account for MBH traffic variability over time, it is assumed that the network will operate under different transmission schemes as active node pairs change due to different traffic demands. Such temporal evolution can be organized into slots of fractional length ai where ai ≥ 0 and

The resulting time-division schedule of the network is given by

Since each transmission scheme Si corresponds to a rate matrix R(S), the rate matrix of the time-division schedule is given by



[0020] It is thus the weighted ensemble of admissible rate matrices Ri that defines the overall network time-division matrix R(T), and ultimately the network's capacity region. Mathematically, if

[0021] {R1, ..., Rk} denote the set of admissible rate matrices, the capacity region of the network is defined as


where Co({Ri}) denotes the convex hull of the set of {Ri} admissible rate matrices. The word admissible is intentionally used to indicate that all rate matrices must abide by the imposed transmission protocol. For example, a given transmission protocol may only allow single-hop routing and may treat all interference as noise, while another may allow multi-hop routing and use successive interference cancellation. By determining the pool of admissible rate matrices, the transmission protocol directly governs the shape of the capacity region, which is in this way directly tied to concrete network topology, physical-layer parameters, as well as the adopted transmission protocol. The dimensionality of the capacity region can be as large as n×(n-1) since the network has n source nodes, each with up to (n-1) possible destinations. For easier visualization, two-dimensional slices of C may be computed when only 2 of possible n nodes are active. Alternately, a uniform capacity metric, Cu = rmaxn×(n-1) may be computed, corresponding to the largest transmission rate that all n×(n - 1) communication pairs can support under time-division routing and a given transmission protocol. Furthermore, since the rate matrices for a n-node network are isomorphic to vectors with length n×(n-1), the resulting problem of computational geometry may be solved by linear programming as an optimization problem in N-1 Euclidean space as



[0022] By iteratively solving the linear problem of (C), the vertices of the capacity regions, corresponding to optimal network modes of operation (i.e. MBH transmission topology), may be determined. However, to formally solve (C), the set of all admissible rate matrices must first be calculated, which may increase quite rapidly with the number of nodes, n. Significant speed gains may thus be obtained by recognizing that not all rate matrices significantly contribute to the capacity regions, and constructing smaller subsets of admissible rate matrices that accurately reproduce most of the capacity region. Cases where fixed a priori MBH links already exist are natural examples of this simplification, as mentioned above.

[0023] Finally, to relate the MBH transmission topology optimization with physical-layer parameters, the rate matrix entries, rij, which denote the data rate between source node (i) and destination node (j), are defined as a direct function of the effective signal-to-noise ratio (SNR), ξij, on the source-destination link; in other words, rij = f(ξij), where both the ξij computation and the choice of the function f(ξij) can be varied in accordance with both the physical-layer technology employed (e.g. microwave vs. free-space optical vs. fiber transmission), as well as the desired data rate criterion. For example, rij can be obtained as the maximum data rate that satisfies a given bit error rate (BER) requirement under a selected coding/modulation scheme for a given SNR, ξij, which may be computed from the adopted coding, modulation, signal and noise power parameters. In an Additive White Gaussian Noise (AWGN) channel, which is the most common case for all of the MBH physical-layer technologies, closed form expressions based on the Gaussian Q-function may be used for f(ξij) From this it also follows that improvements in ξij via physical-layer techniques such as modulation, detection and coding thus improve not only maximum link transmission rates, but also the achievable rates and the connectivity of the entire network. In the most ideal case, f(ξij) can also be equated to the Shannon capacity of the link, which, for an AWGN channel, is given by

where in (D) ξij is the link SNR and Wij is the physical bandwidth of the link. As mentioned above, depending on the physical-layer technology employed, the link SNR may be computed in customized ways, incorporating the crucial physical-layer specifics into the topology planning and optimization problem. For microwave links, for example, ξij can be calculated as

where in (E), Pi is the transmitted power, Gij is the channel gain coefficient from source (i) to destination (j), σj2 is the AWGN variance, and the second term in the denominator of (E) accounts for any microwave interference on the link that is also treated as noise. For direction RF or free-space optical transmission, SNR ξij may be computed as

where in (F), η denotes the AWGN power spectral density, W is the bandwidth, P is transmitted power, θij is the one-dimensional beam width of the signal, dij is the Euclidean distance between the source and destination, d0 is the reference distance based on which the effect of the path loss parameter α is computed, and the channel gain coefficient Gij accounts for any fading or shadowing effects. Finally, for rij computation on optical fiber links, the SNR and optical SNR (OSNR) may be related as

where in (G), ρ = 1, 2 for the cases of single-polarization and polarization-multiplexed optical signals, respectively, Rs is the symbol rate, and Bref is the reference bandwidth for OSNR measurement (commonly taken as 12.5GHz).

[0024] By using (D)-(G) to compute rij entries in each admissible rate matrix, and substituting the rate matrices into a linear optimization routine to solve (A)-(C), the SD MBH network optimizer (100) computes the MBH transmission topology that supports the highest backhaul data rate, and enables flexible, software-reconfigurable network planning and optimization in a way that also incorporates the physical-layer specifics of every MBH link in the overall network. Finally, in cases where experimental measurements of ξij are available, these may also be readily used in the rij computation in place of the analytical values given by (D)-(G).

[0025] Fig. 2 illustrates operational principles of the architecture. Fig. 2 shows the SD network optimizer (100) and its communication with the MBH Tx/Rx, denoted by (300)-(700) in Fig. 1; it is noted that the mobile users, i.e. MSs (800)-(810) in Fig. 1, do not participate in this process such that the MBH transmission topology optimization is abstracted from the end user for simplicity and ease of management. As shown in Fig. 2, the SD network optimizer (100) first collects the needed physical-layer data (101) from the MBH Tx/Rx (300)-(700) by an exchange of generic control messaging (201), which may be implemented via the OpenFlow interface (200), for example. The collected physical-layer parameters can include, but are not restricted to, the Tx/Rx technology type, link SNR/OSNR, adopted network transmission protocol, traffic load (i.e. number and data rate demands of MSs assigned to each MBH Tx/Rx), etc.. Based on the physical-layer information compiled in (101), the network optimizer (100) computes the admissible set of rate matrices (102) that can be used for the capacity region and/or uniform capacity calculation. To account for any network constraints (e.g. fixed links, and/or computational complexity limitations), the SD network optimizer (100) can also prune the set of admissible rate matrices to a smaller sub-set (103). Based on the results of (103), the capacity computation is performed via linear optimization outlined in (A)-(C), to determine the MBH transmission topology which maximizes the backhaul data rate of the network (104). Finally, the MBH transmission topology results are communicated by the SD network optimizer (100) to the MBH Tx/Rx (300)-(700) via generic control messaging (201). To implement the decisions of (100), the MBH Tx/Rx (300)-(700) locally adjust the required physical layer parameters (301), and confirm the resulting modification and new operational parameters to the SD network optimizer (100) via generic control messaging (201), e.g. using OpenFlow (200). The updated physical-layer values can then be used by the SD optimizer (100) to re-optimize topology as needed, e.g. due to network upgrades and/or traffic changes.

[0026] The software-defined centralized controller (100) and its two-way communication and management of the remote MBH Tx/Rx (300)-(700) include the sequence of operations given by (101)-(104) in Fig. 2. By compiling and exploiting the physical-layer specifics of the network MBH Tx/Rx (300)-(700), the SD network optimizer (100) is able to compute a MBH transmission topology that, out of all possible admissible topology combinations, enables the highest backhaul data rate in the network. This is done through the customized linear optimization formulation of (101)-(104), through which the capacity regions and/or uniform capacity of the network can be accurately computed. Prior art focused either on purely wireless solutions and highly-theoretical cases with abstracted physical layer characteristics. Moreover, through the SD network optimization of (101)-(104) in Fig. 2, customized network planning and optimization are enabled in a software-reconfigurable way that supports physical-layer accuracy and is sufficiently malleable to accommodate an arbitrary mix of heterogeneous physical-layer MBH technologies and architectures.

[0027] The function blocks (101)-(104) enable the operation of the SD network optimizer (100) in Fig. 2. These features enable the efficient, accurate, and software-reconfigurable computation of a MBH network topology which, out of all possible admissible topology combinations, enables the highest backhaul data rate in the network. Prior art focused either on all-wireless solutions and neglected optimization of the backhaul topology, and/or considered highly-theoretical cases with abstracted physical layer characteristics. The SD operation can enable cost-efficient, flexible MBH network planning, optimization and upgrades that can be directly customized to the network scenario at hand by the direct inclusion of relevant physical-layer parameters into the data rate maximization problem.

Further system details



[0028] We propose a software-defined algorithm for dynamic, physical-layer-aware throughput maximization in next-generation mobile backhaul (MBH) networks. Results confirm >100Mb/s end-to-end per-cell throughputs with ≥2.5Gb/s optical backhaul links deployed at legacy cell sites.

I. INTRODUCTION



[0029] The rapid proliferation of 4G and beyond mobile technologies is dramatically changing mobile backhaul (MBH) and access networks [6, 7]. High-density small cell deployments are transforming legacy last-mile MBH segments into pre-aggregation networks, with legacy macro cell sites evolving into aggregation points for a dozen or more small cells. Moreover, fiber-optic MBH is emerging as a highly-attractive complement to point-to-point wireless backhaul, particularly at high-density aggregation sites (e.g. legacy macro cells). Maximizing throughput of the resulting heterogeneous point-to-point wireless and optical MBH network thus becomes a critical technical and economic challenge. First, if fiber links already exist, are they being used with maximal efficiency? Secondly, if additional optical links need to be deployed the MBH network, where should these be placed to maximize throughput? In previous studies for legacy networks [1,3,5], throughput optimization via optimal cell site placement for randomly distributed mobile users was considered, yet the optimization of the cell site interconnections (i.e. MBH infrastructure) was not examined. In this paper, we introduce and evaluate a software-defined (SD) MBH throughput optimization algorithm which uses global knowledge of physical-layer MBH transmission topology (which can include both point-to-point wireless and optical links) to maximize per-cell end-to-end throughput, rmax, identify per-link bottlenecks, and assess rmax gains that would arise by upgrading the bottleneck links. The algorithm is evaluated for a heterogeneous optical/wireless last-mile small cell backhaul network, confirming 4G LTE-compatible per-cell rmax =141.1Mb/s with 2.5Gb/s fiber links at legacy macro cell sites. The approach is thus attractive for optical MBH planning/optimization.

II. SOFTWARE-DEFINED (SD) OPTICAL MBH THROUGHPUT OPTIMIZATION ALGORITHM



[0030] As shown in Fig. 3, the SD algorithm seeks to optimize throughput in an evolving MBH network composed of both legacy (macro) and emerging (small cell) sites. Specifically, given a MBH network of n nodes and global knowledge of physical-layer link parameters (block i in Fig. 3), communicated via generic protocols (e.g. OpenFlow, block iii in Fig. 3), the SD algorithm computes the maximum achievable throughput rate, rmax, between all n/2 communication pairs (block ii in Fig. 3), where the same rmax is required for all pairs for fairness. (Based on rmax, the aggregate network throughput can readily be computed as rmax × n/2.) As shown in blocks i, ii, and iii in Fig. 3, the optimal value of rmax will depend on both the physical per-link rates in the network, as well as on optimized scheduling/routing of the MBH data flows. To translate both parameters into the throughput optimization problem, we define a set of rate matrices, R, for the n-node network, wherein each R is a n × n matrix populated by physical per-link data rates, rij, such that: rij = rij if node j receives information at rate rij from node i; rij = -rij if j transmits information at rate rij from i; and rij = 0 otherwise. At any given time, T=t, the corresponding rate matrix, R(t), thus contains all data flow information in the network. The rate matrix entries, rij, can then be computed based on the effective signal-to-noise ratio (SNR), ξij, on each MBH link. For ideal AWGN channels, rij = Wijlog2(1+ξij), where ξij denotes link SNR and Wij is physical bandwidth. For non-ideal wireless MBH links, ξij can be computed as

where Pi is the transmitted power, Gij is the channel gain coefficient,

is the AWGN noise variance, and the second term in the denominator accounts for any interference that is also treated as noise [1]. For point-to-point wireless links, ξij may be computed as:

where θij and θji are the one-dimensional antenna beam widths, dij is the Euclidean distance between nodes i and j, and d0 is the reference distance based on which the effect of the free-space path loss parameter α is computed [1]. Finally, for rij computation on optical fiber links, the SNR and optical SNR (OSNR) may be related as

where ρ = 1, 2 for single- and dual- polarization signals, respectively, Rs is the symbol rate, and Bref is the OSNR reference bandwidth. To then account for MBH traffic variability over time, T, and dynamically optimize scheduling/routing, the temporal evolution is modeled by fractional time slots

where N denotes the number of slots. The rate matrix of the dynamic network, RDN, is then given by

wherein each Rk in (4) is populated by physical per-link rates rij of (1)-(3). The optimal RDN is then found by computing the schedule, {tk, k=1, ...,N}, which enables the highest effective throughput rate, rmax, between all communication pairs in the network. Mathematically, for a n-node network, the dimensionality of RDN in (4) is isomorphic to vectors with length n(n-1), such that (4) may be solved as a linear optimization problem in N-1 Euclidean space as



[0031] The result of (5) can then be exploited by the SD MBH controller (Fig. 3) to make the traffic routing/scheduling decisions that maximize overall throughput, as well as to identify network bottlenecks (i.e. those links/hops that require longer scheduling slots, tk), and evaluate throughput effects of cell site MBH upgrades to fiber connectivity. Moreover, such optimization can be done in a software-reconfigurable way, accounting for real-time changes in traffic demands and patterns.

III. SD MBH THROUGHPUT OPTIMIZATION RESULTS AND ANALYSIS



[0032] The SD throughput maximization algorithm was evaluated on the MBH network of the inset of Fig. 4, formed by n = 20 nodes randomly distributed on a 3.75km2

[0033] 1.5mi2 (e.g. last mile) rectangular area and mapped into n/2 = 10 source-destination pairs. As shown in the inset of Fig. 4, each dark square denotes a new small cell ("pre-aggregation") site targeted for point-to-point wireless backhaul, while each light circle indicates a co-located legacy macro/new small-cell ("aggregation") site, already featuring legacy point-to-point wireless backhaul and potentially requiring an upgrade to fiber connectivity. In other words, for the legacy MBH sites, which both originate traffic and must route traffic to/from other cells, the SD algorithm seeks to evaluate throughput effects of MBH link upgrades to optical fiber. To compute rmax for each target MBH case I-VI in Fig. 4, all per-link rates, rij, for the network in the inset of Fig. 4 were first calculated for the according to (1)-(3), wherein, unless otherwise noted Pi =0.1Watts, dij values were taken from the geometry of the inset of Fig. 4, d0 = 70m, α = 3,

ρ = 1, and Bref = 12.5GHz. A linear optimization routine was implemented in MATLAB to compute rmax using the rate matrix formulation of (4)-(5), with the per-cell maximum throughput, rmax, plotted in Fig. 4 versus wireless MBH bandwidth, Wij (MHz); as shown in Fig. 4, for all cases, low Wij values created severe local MBH bottlenecks and limited rmax ≤ 20Mb/s despite fiber connectivity, highlighting the importance of this parameter. As also shown in Fig. 4, while the upgrade of legacy sites to 1Gb/s fiber backhaul (Case III) improved rmax compared to both all-wireless MBH scenarios (Cases I and II), an additional upgrade of legacy sites to 2.5Gb/s fiber links (Case IV) was required to alleviate routing/scheduling bottlenecks and enable rmax = 100Mb/s, corresponding to 4G LTE data rates. We note that this numerical optimization result (i.e. 2.5Gb/s fiber-optic backhaul rate for n = 20 cells in a last-mile scenario) is in line with previous analytical and experimental evaluations of 24-25Gb/s/λ optical MBH rates for n ≈ 200 cell last-mile backhaul [7]. Moreover, as shown in Fig. 4 for Case V, a further upgrade of legacy sites of the inset of Fig. 4 to 10Gb/s fiber backhaul provided virtually no rmax gains because, in this case, end-to-end throughput was limited by the substantially lower rates on the wireless backhaul links. This was confirmed by the results of Case VI, where rmax = 141.1Mb/s was achieved by upgrading legacy MBH sites to 2.5Gb/s fiber links and by adopting more advanced antenna directionality to increase the point-to-point wireless pre-aggregation segment MBH rates.

CONCLUSIONS



[0034] We have proposed and evaluated a software-defined throughput maximization algorithm for next-generation optical MBH. Results show that judiciously placed 2.5Gb/s optical backhaul links can notably increase throughput in heterogeneous 4G MBH systems and support software-reconfigurable network planning.

[0035] The foregoing is to be understood as being in every respect illustrative and exemplary, but not restrictive, and the scope of the invention disclosed herein is not to be determined from the Detailed Description, but rather from the claims as interpreted according to the full breadth permitted by the patent laws. It is to be understood that the embodiments shown and described herein are only illustrative of the principles of the present invention and that those skilled in the art may implement various modifications without departing from the scope of the invention. Those skilled in the art could implement various other feature combinations without departing from the scope of the invention.


Claims

1. A method implemented in a network controller (100) used in a mobile backhaul (MBH) network, the method comprising:

receiving one or more physical layer parameters from MBH transceivers (300, 400, 500, 600, 700) in the MBH network;

computing an admissible set of rate matrices;

computing a capacity region C based on the admissible set of rate matrices, defined as

wherein ai is a fractional length of a time slot and Ri is a rate matrix;

deciding MBH transmission topology which enables the highest backhaul data rate in the network; and

transmitting the decision to the MBH transceivers;

wherein rate matrix entry rij, which denote a data rate between source node i and destination node j, is expressed as a function of signal-to-noise ratio (SNR) ξij on a source-destination link:

wherein the SNR is calculated according to a formula that can be expressed as

for direction radio frequency (RF) or free-space optical transmission,
where η denotes an Additive White Gaussian Noise (AWGN) power spectral density, W is a bandwidth, P is transmitted power, θij is a one-dimensional beam width of a signal, dij is a Euclidean distance between source node i and destination node j, d0 is a reference distance based on which an effect of path loss parameter α is computed, and channel gain coefficient Gij accounts for a fading or shadowing effect.


 
2. The method as in claim 1, wherein said one or more physical layer parameters include at least one of a signal-to-noise ratio (SNR), an optical SNR, a transceiver technology type, a transmission protocol, and a traffic load.
 
3. The method as in claim 1, further comprising:
pruning the admissible set of rate matrices to account for a physical-layer constraint or a computation complexity limitation.
 
4. The method as in claim 1, wherein the MBH transmission topology is decided by



where Ri denotes an admissible rate matrix, {Ri}={R1, ... , Ri, ... , RN} is a set of admissible rate matrices, and R is a rate matrix of a time-division schedule.
 
5. The method as in claim 1, wherein rij is obtained as a maximum data rate that satisfies a given bit error rate (BER) requirement under a selected coding or modulation scheme for given SNR ξij.
 
6. The method as in claim 5, wherein the given SNR is computed from at least one of adopted coding, adopted modulation, and a signal and noise power parameter.
 
7. The method as in claim 1, wherein the function can be expressed as

where Wij is a physical bandwidth of the source-destination link.
 
8. The method as in claim 1, wherein the SNR is calculated according to a formula that can be expressed as

for a microwave link,
where Pi is transmitted power, Gij is a channel gain coefficient from source node i to destination node j, and σj2 is an Additive White Gaussian Noise (AWGN) variance.
 
9. The method as in claim 1, wherein the SNR for an optical fiber link (optical SNR, OSNR) is calculated according to a formula that can be expressed as

where ρ = 1, 2 for single-polarization and polarization-multiplexed optical signals, respectively, Rs is a symbol rate, and Bref is a reference bandwidth for OSNR measurement.
 
10. The method as in claim 1, wherein the decision is transmitted to the MBH transceivers (300, 400, 500, 600, 700) via OpenFlow control messaging.
 
11. A network controller (100) used in a mobile backhaul (MBH) network, the network controller comprising:

a reception means for receiving one or more physical layer parameters from MBH transceivers (300, 400, 500, 600, 700) in the MBH network;

a first computing means for computing an admissible set of rate matrices;

a second computing means for computing a capacity region C based on the admissible set of rate matrices, defined as

wherein ai is a fractional length of a time slot and Ri is a rate matrix;

a decision means for deciding MBH transmission topology; and

a transmission means for transmitting the decision to the MBH transceivers

wherein rate matrix entry rij, which denote a data rate between source node i and destination node j, is expressed as a function of signal-to-noise ratio (SNR) ξij on a source-destination link:

wherein the SNR is calculated according to a formula that can be expressed as

for direction radio frequency (RF) or free-space optical transmission,
where η denotes an Additive White Gaussian Noise (AWGN) power spectral density, W is a bandwidth, P is transmitted power, θij is a one-dimensional beam width of a signal, dij is a Euclidean distance between source node i and destination node j, d0 is a reference distance based on which an effect of path loss parameter α is computed, and channel gain coefficient Gij accounts for a fading or shadowing effect.


 
12. A mobile backhaul (MBH) network comprising:

a network controller (100);

MBH transceivers (300, 400, 500, 600, 700) connected to the network controller (100),

wherein the network controller (100)

receives one or more physical layer parameters from the MBH transceivers (300, 400, 500, 600, 700) in the MBH network,

computes an admissible set of rate matrices,

computes a capacity region C based on the admissible set of rate matrices , defined as

wherein ai is a fractional length of a time slot and Ri is a rate matrix,

decides MBH transmission topology, and

transmits the decision to the MBH transceivers

wherein rate matrix entry rij, which denote a data rate between source node i and destination node j, is expressed as a function of signal-to-noise ratio (SNR) ξij on a source-destination link:

wherein the SNR is calculated according to a formula that can be expressed as

for direction radio frequency (RF) or free-space optical transmission,
where η denotes an Additive White Gaussian Noise (AWGN) power spectral density, W is a bandwidth, P is transmitted power, θij is a one-dimensional beam width of a signal, dij is a Euclidean distance between source node i and destination node j, d0 is a reference distance based on which an effect of path loss parameter α is computed, and channel gain coefficient Gij accounts for a fading or shadowing effect.


 


Ansprüche

1. Verfahren, das in einer Netzwerksteuerung (100) implementiert ist, die in einem mobilen Backhaul-(MBH)-Netzwerk verwendet wird, wobei das Verfahren umfasst:

Empfangen eines oder mehrerer physikalischer Layer-Parameter von MBH-Transceivern (300, 400, 500, 600, 700) im MBH-Netzwerk;

Berechnen eines zulässigen Satzes von Ratenmatrizen;

Berechnen eines Kapazitätsbereichs C basierend auf dem zulässigen Satz von Ratenmatrizen, definiert als

wobei ai is a Bruchteilslänge eines Zeitschlitzes ist und Ri eine Ratenmatrix ist;

Entscheiden der MBH-Übertragungstopologie, die die höchste Backhaul-Datenrate im Netzwerk ermöglicht; und

Übertragen der Entscheidung an die MBH-Transceiver;

wobei Ratenmatrixeingabe rij, ie eine Datenrate zwischen Quellknoten i und Zielknoten bezeichnet, als Funktion des Signal-Rausch-Verhältnisses (SNR) ξij auf einer Quell-Zielverbindung ausgedrückt wird:

wobei die SNR nach einer Formel berechnet wird, die ausgedrückt werden kann als

für Richtungs-Hochfrequenz (RF) oder optische Freiraumübertragung,
wobei η eine spektrale Leistungsdichte des additiven weißen Gaußschen Rauschens (AWGN) bezeichnet, W eine Bandbreite ist, P übertragene Leistung, θij eine eindimensionale Strahlbreite eines Signals ist, dij eine euklidische Entfernung zwischen Quellknoten i und Zielknoten j ist, d0 ine Referenzdistanz ist, auf deren Grundlage ein Effekt des Wegverlustparameters α berechnet wird, und der Kanalgewinnkoeffizient Gij einen Fading oder Shadowing-Effekt darstellt.
 
2. Verfahren nach Anspruch 1, worin der eine oder die mehreren physikalischen Layerparameter mindestens eines von einem Signal-Rausch-Verhältnis (SNR), einem optischen SNR, einem Transceiver-Technologietyp, einem Übertragungsprotokoll und einer Verkehrslast beinhalten.
 
3. Verfahren nach Anspruch 1, ferner umfassend:
Beschneiden des zulässigen Satzes von Ratenmatrizen, um einer Beschränkung der physikalischen Layer oder einer Begrenzung der Komplexität der Berechnung Rechnung zu tragen.
 
4. Verfahren nach Anspruch 1, worin die MBH-Übertragungstopologie bestimmt wird durch



wobei Ri eine zulässige Ratenmatrix ist, {Ri}={R1, ... , Ri, ... , RN} ein Satz von zulässigen Ratenmatrizem ist, und R eine Ratenmatrix eines Time-division schedules ist.
 
5. Verfahren nach Anspruch 1, wobei rij als maximale Datenrate erhalten wird, die eine gegebene Bitfehlerratenanforderung (BER) unter einem ausgewählten Kodierungs- oder Modulationsschema für gegebenes SNR ξij Erfüllt.
 
6. Verfahren nach Anspruch 5, worin der gegebene SNR aus mindestens einer der angenommenen Kodierungen, der angenommenen Modulation und einem Signal- und Rauschleistungsparameter berechnet wird.
 
7. Verfahren nach Anspruch 1, worin die Funktion ausgedrückt werden kann als

wobei Wij ine physikalische Bandbreite der Quell-Zielverbindung ist.
 
8. Verfahren nach Anspruch 1, worin der SNR nach einer Formel berechnet wird, die ausgedrückt werden kann als

für eine Mikrowellenverbindung,
wobei Pi die übertragene Leistung ist, Gij ein Kanalverstärkungskoeffizient vom Quellknoten i zum Zielknoten j ist, und σj2 eine Additiv-Weiß-Gauß-Rausch-Varianz (AWGN) ist.
 
9. Verfahren nach Anspruch 1, worin das SNR für eine Glasfaserverbindung (optisches SNR, OSNR) nach einer Formel berechnet wird, die ausgedrückt werden kann als

Wobei jeweils ρ = 1, 2 für einfach polarisierte bzw. polarisationsmultiplexierte optische Signale, Rs eine Symbolrate und Bref ine Referenzbandbreite für die OSNR-Messung ist.
 
10. Verfahren nach Anspruch 1, wobei die Entscheidung über OpenFlow Control Messaging an die MBH-Transceiver (300, 400, 500, 600, 700) übertragen wird.
 
11. Netzwerksteuerung (100), die in einem mobilen Backhaul-(MBH)-Netzwerk verwendet wird, wobei die Netzwerksteuerung umfasst:

ein Empfangsmittel zum Empfangen eines oder mehrerer physikalischer Layerparameter von MBH-Transceivern (300, 400, 500, 600, 700) im MBH-Netzwerk;

ein erstes Rechenmittel zum Berechnen eines zulässigen Satzes von Ratenmatrizen;

ein zweites Berechnungsmittel zum Berechnen eines Kapazitätsbereichs C basierend auf dem zulässigen Satz von Ratenmatrizen, definiert als

wobei ai eine Bruchstücklänge eines Zeitfensters ist und Ri eine Ratenmatrix ist;

ein Entscheidungsmittel zum Bestimmen der MBH-Übertragungstopologie; und

ein Übertragungsmittel zum Übertragen der Entscheidung an die MBH-Sende-Empfänger

wobei Ratenmatrixeingabe rij, die eine Datenrate zwischen Quellknoten i und Zielknoten j bezeichnet, als Funktion des Signal-Rausch-Verhältnisses (SNR) ξij auf einer Quell-Zielverbindung ausgedrückt wird:

wobei die SNR nach einer Formel berechnet wird, die ausgedrückt werden kann als

für Richtungs-Hochfrequenz (RF) oder optische Freiraumübertragung,
wobei η eine spektrale Leistungsdichte des additiven weißen Gaußschen Rauschens (AWGN) bezeichnet, W eine Bandbreite ist, P übertragene Leistung, θij eine eindimensionale Strahlbreite eines Signals ist, dij eine euklidische Entfernung zwischen Quellknoten i und Zielknoten j ist, do ine Referenzdistanz ist, auf deren Grundlage ein Effekt des Wegverlustparameters α berechnet wird, und der Kanalgewinnkoeffizient Gij einen Fading oder Shadowing-Effekt darstellt.
 
12. Mobiles Backhaul-(MBH)-Netzwerk, umfassend:

eine Netzwerksteuerung (100);

MBH-Transceiver (300, 400, 500, 600, 700), die mit der Netzwerksteuerung (100) verbunden sind,

wobei die Netzwerksteuerung (100)

einen oder mehrere physikalische Schichtparameter von den MBH-Transceivern (300, 400, 500, 500, 600, 700) im MBH-Netzwerk empfängt,

einen zulässigen Satz von Ratenmatrizen berechnet,

einen Kapazitätsbereich C basierend auf dem zulässigen Satz von Ratenmatrizen berechnet, definiert als



wobei ai eine Bruchstücklänge eines Zeitfensters ist und Ri eine Ratenmatrix ist; die MBH-Übertragungstopologie entscheidet und

die Entscheidung an die MBH-Transceiver überträgt,

wobei die Eingabe der Ratenmatrix rij, die eine Datenrate zwischen Quellknoten i und Zielknoten j bezeichnet, als Funktion des Signal-Rausch-Verhältnisses (SNR) ξij auf einer Quell-Zielverbindung ausgedrückt wird:

wobei die SNR nach einer Formel berechnet wird, die ausgedrückt werden kann als

für Richtungs-Hochfrequenz (RF) oder optische Freiraumübertragung,
wobei η eine spektrale Leistungsdichte des additiven weißen Gaußschen Rauschens (AWGN) bezeichnet, W eine Bandbreite ist, P übertragene Leistung, θij eine eindimensionale Strahlbreite eines Signals ist, dij eine euklidische Entfernung zwischen Quellknoten i und Zielknoten j ist, d0 ine Referenzdistanz ist, auf deren Grundlage ein Effekt des Wegverlustparameters α berechnet wird, und der Kanalgewinnkoeffizient Gij einen Fading oder Shadowing-Effekt darstellt.
 


Revendications

1. Procédé mis en œuvre dans un contrôleur de réseau (100) utilisé dans un réseau de backhaul mobile (MBH), le procédé comprenant :

la réception d'un ou plusieurs paramètres de couche physique en provenance d'émetteurs-récepteurs MBH (300, 400, 500, 600, 700) dans le réseau MBH ;

le calcul d'un ensemble admissible de matrices de débits ;

le calcul d'une région de capacité C sur la base de l'ensemble admissible de matrices de débits, selon

ai est une longueur fractionnelle d'un créneau temporel et Ri est une matrice de débits ;

la décision d'une topologie de transmission MBH permettant le plus haut débit de données de backhaul dans le réseau ; et

la transmission de la décision aux émetteurs-récepteurs MBH ;

dans lequel une entrée de matrice de débits rij, indiquant un débit de données entre un nœud de source i et un nœud de destination j, est exprimée en fonction d'un rapport de signal sur bruit (SNR) ξij sur une liaison de source à destination :

dans lequel le SNR est calculé selon une formule pouvant être exprimée par

pour une transmission radiofréquence (RF) directionnelle ou une transmission optique en espace libre,
η indique une densité spectrale de puissance de bruit blanc additif gaussien (AWGN), W est une largeur de bande, P est une puissance transmise, θij est une largeur de faisceau unidimensionnel d'un signal, dij est une distance euclidienne entre un nœud de source i et un nœud de destination j, d0 est une distance de référence sur la base de laquelle un effet de paramètre d'atténuation de propagation a est calculé, et un coefficient de gain de canal Gij tient compte d'un effet d'évanouissement ou d'occultation.


 
2. Procédé selon la revendication 1, dans lequel lesdits un ou plusieurs paramètres de couche physique comportent au moins l'un parmi un rapport de signal sur bruit (SNR), un SNR optique, un type de technologie d'émetteur-récepteur, un protocole de transmission et une charge de trafic.
 
3. Procédé selon la revendication 1, comprenant en outre :
l'élagage de l'ensemble admissible de matrices de débits pour tenir compte d'une contrainte de couche physique ou d'une limitation de complexité de calcul.
 
4. Procédé selon la revendication 1, dans lequel la topologie de transmission MBH est décidée par



Ri indique une matrice de débits admissible, {Ri) = {R1,...,Ri,...,RN} est un ensemble de matrices de débits admissibles, et R est une matrice de débits d'un ordonnancement par répartition dans le temps.
 
5. Procédé selon la revendication 1, dans lequel rij est obtenu en tant qu'un débit de données maximal satisfaisant une exigence de taux d'erreur binaire (BER) donnée pour un schéma de codage ou de modulation sélectionné, à SNR ξij donné.
 
6. Procédé selon la revendication 5, dans lequel le SNR donné est calculé à partir d'au moins l'un d'un codage adopté, d'une modulation adoptée, et d'un paramètre de puissance de signal et de bruit.
 
7. Procédé selon la revendication 1, dans lequel la fonction peut être exprimée par

Wij est une largeur de bande physique de la liaison de source à destination.
 
8. Procédé selon la revendication 1, dans lequel le SNR est calculé selon une formule pouvant être exprimée par

pour une liaison hyperfréquence,
Pi est une puissance transmise, Gij est un coefficient de gain de canal du nœud de source i au nœud de destination j, et

est une variance de bruit blanc additif gaussien (AWGN).
 
9. Procédé selon la revendication 1, dans lequel le SNR pour une liaison de fibre optique (SNR optique, OSNR) est calculé selon une formule pouvant être exprimée par

ρ = 1, 2 pour respectivement des signaux optiques à polarisation unique et à multiplexage de polarisation, Rs est un débit de symboles, et Bref est une largeur de bande de référence pour une mesure OSNR.
 
10. Procédé selon la revendication 1, dans lequel la décision est transmise aux émetteurs-récepteurs MBH (300, 400, 500, 600, 700) par l'intermédiaire d'une messagerie de commande OpenFlow.
 
11. Contrôleur de réseau (100) utilisé dans un réseau de backhaul (MBH), le contrôleur de réseau comprenant :

un moyen de réception pour la réception d'un ou plusieurs paramètres de couche physique en provenance d'émetteurs-récepteurs MBH (300, 400, 500, 600, 700) dans le réseau MBH ;

un premier moyen de calcul pour le calcul d'un ensemble admissible de matrices de débits ;

un deuxième moyen de calcul pour le calcul d'une région de capacité C sur la base de l'ensemble admissible de matrices de débits, selon :

ai est une longueur fractionnelle d'un créneau temporel et Ri est une matrice de débits ;

un moyen de décision pour la décision d'une topologie de transmission MBH ; et

un moyen de transmission pour la transmission de la décision aux émetteurs-récepteurs MBH

dans lequel une entrée de matrice de débits rij, indiquant un débit de données entre un nœud de source i et un nœud de destination j, est exprimée en fonction d'un rapport de signal sur bruit (SNR) ξij sur une liaison de source à destination :

dans lequel le SNR est calculé selon une formule pouvant être exprimée par

pour une transmission de radiofréquences (RF) directionnelle ou optique d'espace libre,
η indique une densité spectrale de puissance de bruit blanc additif gaussien (AWGN), W est une largeur de bande, P est une puissance transmise, θij est une largeur de faisceau unidimensionnel d'un signal, dij est une distance euclidienne entre un nœud de source i et un nœud de destination j, d0 est une distance de référence sur la base de laquelle un effet de paramètre d'atténuation de propagation a est calculé, et un coefficient de gain de canal Gij tient compte d'un effet d'évanouissement ou d'occultation.


 
12. Réseau de backhaul (MBH) comprenant :

un contrôleur de réseau (100) ;

des émetteurs-récepteurs MBH (300, 400, 500, 600, 700) connectés au contrôleur de réseau (100),

dans lequel le contrôleur de réseau (100)

reçoit un ou plusieurs paramètres de couche physique en provenance des émetteurs-récepteurs MBH (300, 400, 500, 600, 700) dans le réseau MBH,

calcule un ensemble admissible de matrices de débits,

calcule une région de capacité C sur la base de l'ensemble admissible de matrices de débits, selon

ai est une longueur fractionnelle d'un créneau temporel et Ri est une matrice de débits,

décide une topologie de transmission MBH, et

transmet la décision aux émetteurs-récepteurs MBH,

dans lequel une entrée de matrice de débits rij, indiquant un débit de données entre un nœud de source i et un nœud de destination j, est exprimée en fonction d'un rapport de signal sur bruit (SNR) ξij sur une liaison de source à destination :

dans lequel le SNR est calculé selon une formule pouvant être exprimée par

pour une transmission radiofréquence (RF) directionnelle ou pour une transmission optique en espace libre,
η indique une densité spectrale de puissance de bruit blanc additif gaussien (AWGN), W est une largeur de bande, P est une puissance transmise, θij est une largeur de faisceau unidimensionnel d'un signal, dij est une distance euclidienne entre un nœud de source i et un nœud de destination j, d0 est une distance de référence sur la base de laquelle un effet de paramètre d'atténuation de propagation α est calculé, et un coefficient de gain de canal Gij tient compte d'un effet d'évanouissement ou d'occultation.


 




Drawing















REFERENCES CITED IN THE DESCRIPTION



This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

Patent documents cited in the description




Non-patent literature cited in the description