Journal of Network and Computer Applications 116 (2018) 24–34
Contents lists available at ScienceDirect
Journal of Network and Computer Applications journal homepage: www.elsevier.com/locate/jnca
Performance based user-centric dynamic mode switching and mobility management scheme for 5G networks
T
Sanjay Kumar Biswasha,b,∗, Dushantha Nalin K. Jayakodya a b
Department of Software Engineering, School of Computer Science and Robotics, National Research Tomsk Polytechnic University, Tomsk, Russia Department of Computer Science and Engineering, NIIT University, Neemrana, Rajasthan 301705, India
A R T I C LE I N FO
A B S T R A C T
Keywords: User-centric communication Mobility management 5G networks QoE Cellular network
User-centric communication in fifth generation (5G) network enables wireless peer-to-peer network interface between mobile users in order to improve the data rate and offload the traffic for improved QoE as compared to traditional legacy base station centric (network centric or eNB centric) architecture. In this paper, we introduce a user-centric performance based cooperative cellular communication architecture and device mobility management procedure for 5G networks. Due to the exponential growth of connected devices, the users are deployed very densely. It motivate the researcher for user-centric communication, and it is a feasible approach, where the user can communicate via a relay node with minimum network infrastructure support. The proposed approach is central to mode switching and supporting a high degree of user mobility during the communication. The mode switching techniques is depends on quality parameters (such as link utilization, delay, and energy consumption). When the network switches the communication link from traditional mode (network centric) to user-centric mode of communication, it resort to relays to sustained the quality parameters. The relay selection is a random process, the network selects an arbitrary node as a relay node without any negotiation on performance metrics and node mobility. In order to improve the network mobility management performance, a mobility management scheme is proposed, where the system computes the QoS/QoE and make a decision for mode switching between network-centric to user-centric or vice-versa. The proposed technique show better performance over the traditional cellular network and we compare our results with 4G/Long Term Evolution (LTE), with respect to link utilization, energy consumption, call-to-mobility ratio analysis and system scalability. The performance analysis and comparison demonstrates the superiority of proposed system in terms of QoE parameters as compared to LTE networks.
1. Introduction The 5G is an emerging area of modern cellular networks, and expected to deploy by 2020. It is assumed to be the future of next generation wireless cellular systems with massive advantages over current communication technology (Andrews et al., 2014; Thompson et al., 2014; Agyapong et al., 2014; Zhang et al., 2015). It has many complex and challenging tasks associated with its real time deployment (Mishra et al., 2016a; Biswash and Kumar, 2010). Now the academia, industry, regulatory organizations, and the standardization bodies are anticipating its implementation, as the current cellular technologies cannot fulfill the communication needs of forthcoming user expectations (Ameigeiras et al., 2015). The authors of (Boccardi et al., 2014a) proposed “Big Five” features for 5G networks such as: Massive Multiple Input Multiple Output (M-MIMO), Device-centric Communication,
Native support for heterogeneous communication, Millimeter wave, Smarter device-to-Smarter device, Native support for Machine-to-Machine communication system. The 5G and beyond network generations anticipate support for the concept of user-assistant green networking (Mishra et al., 2016b; Rowell et al., 2014). Resource allocation is an important challenge of 5G networks. The authors of (Mishra et al., 2016a; Hoang et al., 2016) show the importances of 5G networks and associate research challenges, including link efficiency, energy-efficiency, resource allocation etc. for performance enhancement. In this paper, we focus on performance improvement and QoE metrics analysis of user-assisted 5G networks. The network slicing is another a promising technology for 5G networks in order to provide services tailored for users’ specific QoS issues (Zhang et al., 2017a). It is driven by the increased data traffic from different applications, efficient resource allocation schemes should be exploited to improve the flexibility of
∗ Corresponding author. Department of Software Engineering, School of Computer Science and Robotics, National Research Tomsk Polytechnic University, Tomsk, Tomskaya oblast', 634050, Russia. E-mail addresses:
[email protected] (S.K. Biswash),
[email protected] (D.N.K. Jayakody).
https://doi.org/10.1016/j.jnca.2018.05.013 Received 2 December 2017; Received in revised form 29 April 2018; Accepted 9 May 2018 Available online 11 May 2018 1084-8045/ © 2018 Elsevier Ltd. All rights reserved.
Journal of Network and Computer Applications 116 (2018) 24–34
S.K. Biswash, D.N.K. Jayakody
(M2M) communication and mobile devices referred to as user equipment (UE). The LTE network architecture is divided into the RAN and Core Network (CoN). Here, Enhanced Node B (eNB) is the base station (BS) component of the LTE network. It provides PHY and MAC layer services to the broadband users, and is an intelligent network (Chen et al., 2017). Fig. 1 shoes the LTE network architecture. The eNB and Mobility Management Entity (MME) is directly connected with the S1 links and all eNBs are commented with X2 links. The MME is responsible for mobility management, network access control, roaming management etc. The combination of X2 and S1 link supported by U-TRAN (Abu-Ali et al., 2014). In this paper we consider the LTE is base line cellular network architecture and it follows the eNB centric communication between the mobile nodes. Where, the all traffic goes via eNB and network has full control over it in order to provide the better user experience, it refer as network-centric communication system (Chen et al., 2017; Soltanmohammadi et al., 2016).
network resource allocation and capacity of 5G networks based on network slicing. In this work, we are consider delay, link utilization and energy consumption as a QoE metrics. The delay is associated with overall communication delay between source node and destination mobile user. The modern cellular network requires less communication delay, and consider it a performance metrics. We also considers the link utilization as a quality metric because, the cellular network has limited bandwidth, channel capacity and dynamic channel allocation for multiple users. The proposed work help to reduce traffic over the network, because the associated traffic does not move via the eNB and network. The cellular network consume 60% of energy in signalling energy. The high energy consumption lead to frequent device power discharge, and it refer as a poor performance. The proposed work helps to reduce the overall signalling energy. The Table 1 shows the list of abbreviations used for formulation purpos. This paper is organized as follows: Section 2 provides the details of the domain specific literature survey. Section 3, presents the proposed mode switching procedure and mobility management. The complete system formulation is available in Section 4. The results and discussion are presented in Section 5 and followed by references.
2.2. User-centric communication in 5G and beyond The vision for the next generation 5G wireless networks lies in providing very high data rates (in Gbps order), extremely low latency, manifold increase in the eNB capacity, and a significant improvement in users’ perceived quality of service (QoS) as compared to the current 4G/ LTE networks (Agiwal et al., 2016). The 5G has the unique feature of Device-centric communication, and it supports minimum infrastructure utilization for mobile communication (Boccardi et al., 2014a). It is a promising area of research, with researchers working to design algorithms, protocols and systems design for device centric 5G networks (Agyapong et al., 2014; Costa-Perez et al., 2017; Schulz et al., 2017). Fig. 2 shows the user-centric communication system, where the mobile node (UE1) is communicating to target node (UE2) using relay node and UE1 can also communicate to UE3 without the relay node association. The user-centric communication is also shown in Fig. 7. The relay node selection is subject to network performance metrics. It is a prototype for user centric 5G network. The associated eNBs work as a intelligent network manager (NM) facilitator, it mange the MME and
2. Literature survey In this section, we provide the overview of current state-of-art for cellular communication and consider the LTE as recent cellular technologies (Soltanmohammadi et al., 2016; Araniti et al., 2013). 2.1. LTE networks The mobile network is upgraded to an LTE assists network. In the LTE network, mobile user have single-carrier frequency-division multiple access (SC-FDMA) for up-link and orthogonal frequency-division multiple access (OFDMA) for down-link channels (Soltanmohammadi et al., 2016). In the LTE network, the mobile system coverage interface facilitates Radio Access Networks (RAN) and support Host-to-Host (H2H) and Machine-to-Machine Table 1 List of abbreviations and symbols. Notations
Descriptions
eNB UE CN n N p A Rc∕Ru D Bdc∕Bdu MC H R ξ Sreq / S{rep}
Enhanced Node B User Equipment Crossroading Node (Target Node) Number of users within a cell Number of active eNB within the service area User movement-probability Coverage area of service provider Transmission rate in CM/UM Power spectrum density Bandwidth allocation for CM/UM Movement coefficient Mobile user's movement directions Network update energy Hop Counts Request and Reply packet
η ϕ w T{prop} T{pros} CMR d Bd(⋅) D(⋅) S(⋅) P(⋅) UT CT v Z
Proportionality constant Liner coefficient for link utilization Per unit association lookup cost Propagation time Processing time Call-to-Mobility ratio Straight line distance from current/origin position Bandwidth function Delay function Signal Strength function Performance function Link round trip time between UE and eNB Link round trip time between CN and eNB The index for UE or CN Metric value between source and target using relay node
Fig. 1. LTE network architecture. 25
Journal of Network and Computer Applications 116 (2018) 24–34
S.K. Biswash, D.N.K. Jayakody
communicate to each other with minimum network support and it provide the heterogeneous network support. 2.3. Motivation and contributions The authors of (Tehrani et al., 2014), suggest a device terminal relay scheme for devices-assisted communication. This represents a dramatic change over the conventional cellular architecture and it brings unique technical challenges. In (Mustafa et al., 2016), a study has been proposed for a separate framework to enable cooperative and device-centric communication for 5G networks. The above discussion indicates that the current state-of-the-art of mode switching (between device to the network), and strategic planning for efficient mobility management with user movement freedom is not being well investigated. Thus, we need a schema to implement the basic features of 5G networks without any compensation for QoS/QoE parameters. Most of the work suffers from poor implementation networks, as well as non-robust and non mobility assisted communications (Fan and ZhaoChih-Lin, 2016). In addressing the presented issues, we offer the following fivefold contribution.
• A QoE aware mode switching algorithm is presented. It switches Fig. 2. User-centric architecture for 5G networks.
local data center responsibility (Chen et al., 2016a). The authors of (Zhang et al., 2018) proposed user-centric network solution to organizes the dynamic femto access point group, in order to meet the exponential data demands in 5G networks, but a limited attention is given to node mobility. In (Tang et al., 2017), author is applying the sequential convex programming technique to achieve a reasonable approximation to find the joint admission control and resource allocation in D2D based 5G networks, but they are not considering the run-time network performance. The authors of (Liu et al., 2018) proposed a architecture of 5G new radio, where the wireless access has further evolved from the traditionally cell-centric radio access to a more flexible beam-based user-centric radio access, but a Network and Transport layer based QoE analysis is missing. In 5G the eNB is upgraded to next generation eNB (gNB). The objective for 5G solutions is to support ondemand mobility, such as on high-speed trains and aeroplanes, and low mobility like stationary devices (Fan and ZhaoChih-Lin, 2016). The device-assisted network faces several open research challenges such as mobility management, network performance, network relay selection with security, resource allocations etc. (Thompson et al., 2014; Zhang et al., 2016; Chen et al., 2016b). The authors of (Giust et al., 2015) introduce a distributed mobility management scheme for 5G networks and a valid framework for future mobile network architecture. User movement and device velocity is an important factor for densified mobile cellular network, and its association with mobility management is introduced in (Arshad et al., 2017). They claim their proposed scheme enables the cooperative eNB service and the strongest interference cancellation to compensate for losing the best connectivity. In (Zhang et al., 2017b), author suggest that the fog radio access network is a feasible approach for mobility management and resource optimization. In (Arshad et al., 2017), authors do not deal with the run-time user mobility during the communication, which will degrade the user performance. In (Yazlcl et al., 2014), authors propose an all-SDN network architecture with hierarchical network control capabilities to allow different grades of performance and complexity in offering core network services and service differentiation for 5G systems. However, the important measure aspects of user mobility and the call-to-mobility ratio has been neglected in the literature to date. The user centric 5G network is a one of the promising solution for next generation networks, and it will be a departure of traditional network centric architecture such as LTE network. In 5G network, the mobile user can
•
• • • •
calls by User-Mode (UM) from the Cellular-Mode (CM) and viseversa. It deals with user and network performance. When the UM has poor performance, then the call will transfer to CM, and throughout the procedure, eNB (or gNB1) monitors the network performance metrics. The cellular network has a dynamic communication topology. Thus, a mobile node (user mode), target node (correspondent node) and intermediate relay node (RN) can change their position frequently. If it affects the QoS/QoE of the communication. In order to overcome these issues, we introduce a mobility management procedure in user-centric communication and during the procedure, eNB will observe the network communication and user performance. During the communication, the eNB will observe the performance of the network. If it is less than a pre-set threshold (defined by the service provider) then the network will add a relay node (RN). Thus, it monitors the network performance. The network can add more relay nodes or replace the relay node for better user experience. We introduce an analytical model and parameter based formulation techniques for proposed mode switching and mobility management procedure with a high degree of user and relay node mobility. Our proposed work is based on users QoE performance, such as link utilization, communication delay and latency, Call-to-Mobility ratio and energy consumption. We analyze these factors and formulate them with respect to several dependent factors. The performance parameters are analyzed over varying degrees of data and dependent factors, and efficiency is checked with respect to the LTE network. Here, we discuss the performance measurements, limitations, and pitfalls of the suggested methodologies.
2.4. Why this paper? The literature survey provides the current-state-of-art in relation to device-centric communication. It is demonstrated that a balance between network performance based mode switching is missing. The proposed work fills the gap between the dynamic device-centric communication and 5G network, without any negotiation for high mobility management. 1. A dynamic user-centric relay based communications network for a 5G system.
1 In this work, we consider eNB is gNB, but most of the time using eNB for better readers understanding.
26
Journal of Network and Computer Applications 116 (2018) 24–34
S.K. Biswash, D.N.K. Jayakody
3.2. Mode switching algorithm
2. A high mobility based network management procedure for user-assisted communication.
The proposed mode switching is discussed in Algorithm 1. Here function S(⋅) and Bd(⋅) represents the signal strength and bandwidth utilization, respectively. These functions will take into account the parameters such as UE, and CN to compute the signal strength and bandwidth utilization between UE to eNB, and CN to eNB. The P(⋅) returns the performance utilization for UE and CN with respect the source and target eNBs. If the performance function metrics are better than the service providers threshold value (in this work, we set the preset threshold values as the LTE network). The call transfers to the usercentric mode of communication, otherwise it will continue with the traditional cellular mode. For better network performance, the network computes the performance metrics using P(⋅) and will be computed using Algorithm 2. The network calculates the current position of each UE and CN is using C(⋅) by an incoming signal from eNBs. The network computes signal energy (signal power) by function E(⋅), and delay time to reach the eNB by D(⋅). If it is greater than the pre-set threshold value, then calls will transfer in to user-mode. The computation of D(⋅) and E (⋅) will be illustrated in Sections 4.3 and 4.4, respectively. In this work, we consider the threshold values (Thrs) is similar to LTE network parameters (Cisco 4G/E). It is a dedicated multi-mode LTE wireless networks and it has backwards compatible with these technologies: Universal Mobile Telecommunications System (UMTS), High Speed Packet Access + (HSPA+), HSPA, Global System for Mobile communications (GSM), Exchanged Data rates for GSM Evolution (EDGE), General Packet Radio Services (GPRS. The frequency bands are For LTE: 700 MHz (band 17) AWS (band 4) 2100 MHz (band 1); For UMTS, HSPA+ and HSPA: 800, 850, 1900, 2100 MHz; For GSM, EDGE and GPRS: 850, 900, 1800, 1900 MHz (Cisco 4G/E).
3. Mode switching procedure In this work, we proposed a dynamic mode switching procedure. Where, network switches the mode of communication form network centric mode to user centric mode for reliable communication. The user centric communication techniques has minimum dependency of network infrastructure and associated eNB has observation over the performance metrics. Where, device-to-device communication is eNB independent communication. The proposed methodology has performance based mode switching scheme, mobility management procedure and relay node management procedure. The subsequent sections demonstrates them in detail.
3.1. Dynamic mode switching and QoE analysis In this section, we provide the overall description of our proposed mode switching procedure. It contains the following steps and Fig. 3 shows the schema of methodology. 1. The UE initiates a call, and it passes through associated eNB. 2. The eNB computes the cell location of the caller and finds the target node location. 3. The eNB computes signal-strength, bandwidth, energy consumption and delay for each UE and CN. In this paper, corresponding node (CN) refer as called mobile user or target mobile user. 4. If the QoE parameters are above the pre-set threshold value (as compared to LTE) then the call will transfer to User-Mode (UM)of communication. 5. Otherwise, the system follows traditional Cellular-Mode (CM) of communication. 6. Within the UM, the eNB estimates the QoS/QoE for each mobile subscriber. If these values are less than the LTE threshold value, then it adds a relay node in order to improve the network performance. (a) The eNB observes the performance of the relay node and the user-mode pair, as well as network performance. (b) If performance is poor, then the network adds an another relay node, replaces the relay node or otherwise continues with usermode. 7. The eNB supervises the overall performance of the system (including user-centric pairing, relay node, relay management, etc.) to ensure acceptable QoE/QoS.
Algorithm 1 Mode Selection Algorithm Data: S(UE), S(CN), Bd(UE), Bd(CN) Result: Device mode switching Compute: P(UE, eNB), P(CN, eNB) while P(UE,eNB), P(CE, eNB) > Thrs do | User-Mode(); end Cellular-Mode();
Algorithm 2 Performance Calculation Algorithm
Fig. 3 shows the schema of our proposed QoE-aware mode switching and mobility management procedure. It has two parts; observation and associated actions. The network observes the performance and takes actions for better services to mobile subscribers. This module has two sub phases. First, the network will take the information from the caller and perform an analysis on them. Next, it will make a decision for mode switching based on performance metrics (switching between the UM or CM). In order to improve the network performance, the system may select the relay nodes and it lead to relay section and the relay management.
Data: C(UE),C(CN) Result: P(x) Compute: E(UE,eNB), E(CN,eNB), D(UE,eNB), D(CN,eNB), Bd(v) if E(UE,eNB)&& E(CN,eNB)&& D(UE,eNB)&& D(CN,eNB) && Bd(v) > Thrs User-Mode(); 3.3. Performance calculation and relay selection procedure To enhance the network performance, the system selects a relay node, which could be any node within the proximity area. The Algorithm 3 will take care of relay node selection and its management. The relay node is a normal mobile node within the coverage area. The eNB selects the relay nodes, because the eNB has the complete network service information (from 4G and beyond network specification, the eNBs are intelligent network equipment). If, the UEs are connected to eNB, then the network knows the complete information about the mobile user as well as the network performance utilization by nodes. It is very common scenario when, all mobile users are not participate in
Fig. 3. Process and mode switching description. 27
Journal of Network and Computer Applications 116 (2018) 24–34
S.K. Biswash, D.N.K. Jayakody
node-to-node communication, but they utilize the full network capacity in terms of multimedia services and other network applications.2 Therefore we consider a network based relay selection technique rather than node based relay selection method. The network based relay selection has advantage of security of mobile nodes. This algorithm will take the input of signal strength, and tenure the channel utilization between the UE and eNB for the target and source nodes. The number of relay nodes will continuously increase until the performance reaches above the threshold value. In Algorithm 3, the S(⋅), CH(⋅) represents the signal strength and channel utilization for communication. If they are not upto mark the system add relay nodes(RN). Algorithm 3 Relay Selection Algorithm Data: S(UE),S(CN) Relay Node RN while (CH(UE → eNB)&& CH(CN → eNB)&& S(UE → CN)&& CH(CN → UE)) < Thrs do RN = RN++;
Fig. 4. The 5G network architecture with smart-home services.
4.1. User mobility model Node mobility is a fundamental aspect of cellular communication, allowing a user to move any direction for better QoS/QoE. In this paper, we consider a general seven cell based network topology. The related seven cells based Markov model is shown in Fig. 6. User movement probability is p. A mobile user moves within the coverage area A, The initial location of the user in a 3D plane is p1(x1, y1, z1) the current location is p2(x2, y2, z2), the distance d = p1 p2 = (x2 − x1)2 + (y2 − y1)2 + (z2 − z1)2 .The hop counts ξ as a dependent factor for distance, i is the initial position indicator of user. The movement coefficient of the user is
3.4. Mobility management technique In the cellular communication system, mobile users are free to move in any direction subject to communication environment. In this paper, we enable users to move in all six-directions of a hexagonal cell, following the i.i.d mobility pattern (as shown in Fig. 6). The Algorithm 4 computes the distance between the current and last location of the mobile users using C(⋅). Based on current location of UE, CN and RN the network computes the performance metrics P(⋅), signal strength S(⋅), delay D(⋅) and energy consumptions is E(⋅). The respective methodology is available in section 4.7. Here, Z is the metric between the source node to target node using relay node.
H
MC = ξ ∑ pi i
∫0
d
tdt (1)
This movement coefficient deal with user-movement, displacement, distance and movement probability. The proposed model provides better QoS/QoE with user mobility, thus the mobility coefficient has an important role. With user mobility, a call can arrive and departure from mobile user. The call arrival rate is λ and departure rate is μ. Fig. 5 shows various call arrivals and departure conditions. When a mobile user is moving far from its origin/ current position, the new location can be a neighboring cell or neighbor-toneighbor cell, respectively. The respective call arrival and departure rates are λ1, λ2, λ3 and μ1, μ2, μ3 for origin-to-neighbor, neighbor-to-neighbor and neighbor-to-origin. In Fig. 5, L1, L2, L3 is the call arrival and M1, M2 and M3 is call departure rates in each cells. With every movement of the user, the associated probability will change and it will affect system performance. These probabilities will depend on call arrival and departure values, pI, pi+1, pI−1. The mobile moves between the cells, and call arrival rate in a
Algorithm 4 Mobility Management Procedure Data: d for UE, CN,RN if(d > Thrs) Compute C(UE),C((CN), C(RN)) The nodes send the Beacon singles to associated eNBs. The eNB computes S(Z),D(Z), E(Z). P(Z) ⇒ S(Z)&&D(Z)&&E(Z) > Thrs User-mode() otherwise Cellular-mode().
• • •
cell is λj + 1 =
Cj + 1 (λi + μj + 1 ) λj
pj + 1 and μj =
Ci + 1 (λj + μj − 1 ) μj − 2
pi . The value of i, j, C
represents the index of positions and call-arrival and call-departure states in a mobile environment.
4. System formulation
4.2. Link utilization
Fig. 4, shows the system architecture for 5G networks. It consist of eNB, WLAN and UEs, and is managed by several other wireless network devices such as IEEE 802.11, Femto-cells etc. A geographical area is served by eNB and it provides the network connectivity of mobile users. In this (5G) architecture, a mobile node can communicate to another device with partial/with-out involution of mobile network infrastructure such as eNB. This methodology is referred as node-centric/ device-centric/user-centric communication in 5G networks. To improve the network and users’ performance, the user-centric pairs can select the relay nodes, called relay assistant user-centric communication in 5G domain. The Table 1 shows the list of abbreviations used for formulation purpose.
In cellular networks, mobile users are served by eNB, with up-link and down-link connections. A service provider has N number of active eNB with in the service area (A). A UEs location at position x, and x ⊂ A, and all eNBs are randomly deployed based on population density. Let, gc and gu are signal gain, Rc and Ru is the transmission rate, αc and αu is the percentage of resource allocation for CM and UM mode of communication, respectively. The Bdc and Bdu is the bandwidth utilization during CM and UM connections. Then link utilization in cellularmode is
Ec R c ⎞ πc = max αc, Bdc ⎛ , ⎝ Pc + Pmn ⎠ ⎜
2
It cause the high link utilization. The proposed work help to reduce the overall network traffic through eNB.
⎟
the link utilization in user-mode is 28
(2)
Journal of Network and Computer Applications 116 (2018) 24–34
S.K. Biswash, D.N.K. Jayakody
Fig. 5. A three cell based call arrival and departure.
Eu Ru ⎞ πu = max αu, Bdu ⎛ . ⎝ Pu + Pmn ⎠ ⎜
⎟
(3)
In this work, there will be two state changes, CM to UM and vice-versa. The state change Markov model is shown in Fig. 8. The state probability puc pcu change is Πuc = pcu + pc and Πuc = puc + pu are respectively for state change. The value of pcu = puc because they are reversible and interchangeable events. The overall link utilization is
π = Πuc πu + Πuc πc .
Fig. 7. User-centric pairs & mobility analysis.
(4)
where, Pmn is minimum power allocation for communication.
(
R c = α c Bd log 2 1 +
gc Pc, Bdc α c DBdc
(
),
Ru = αu Bd log 2 1 +
gu Pu, Bdu αu DBdu
)
it provides
the maximum data rate and its upper bound will set by the service provider. Pc ≤ Pcmax and Pu ≤ Pumax . In the proposed work, network link utilization will also vary from UM and CM. The UM has less link initialization, but eNB has observed over communication. Mean link utilization during the communication in CM mode is
πCM =
∫0
H ,t
pi λi μj Πuc πu Sdt +
∫0
H /2, δt
pi λi ± 1 μj ± 1 Πuc πc Sdt .
(5)
The link utilization in user mode is
Fig. 8. State change markov model for CM & UM
πUM = pi λi μj Πuc πu Sdt + pi λi ± 1 μj ± 1 Πuc πc Sdt .
(6) while user mobility and hop counts will affect the T. During the communication, the mobile user is selecting the mode and total delay in CM is
Mean link utilization during the communication in LTE is
πLTE =
∫0
H , t , δt
(πc + πu)2S dt .
(7)
ξ
⎛ TCM = MC pi ⎜Πuc ∑ ⎝ i=1
The value of t > δt and λ ± 1, μ ± 1 is forward and backward user-cell movement. The S is per hop link utilization cost (Biswash and Kumar, 2010, 2011), where. S =
(
NUE 2(h − 1 + η) w + Tp d
+ NUE NCN
2(h − 1 + η) ϕ d
2 + NUE NCN
ξλi pi St
S{req} S{rep}
∫0
δt
Tdt + Πuc
∫0
δt
⎞ Tdt ⎟. ⎠
(8)
This is the average packet delivery time. Where the UM and CM are fluctuation is available. Then the delay in UM is
)
ϕ
4.3. Delay calculation
⎛ TUM = MCpi ⎜Πuc ∑ ⎝ i=1
In the proposed work, the eNB will select the mode of communication subject to network performance. The communication delay is based on processing and propagation time and it creates the delay constraint. So, T = T{pros} + T{prop}. The mobile user is independent,
∫0
δt
Tdt + Πuc
∫0
δt
⎞ Tdt ⎟, ⎠
(9)
where the maximum value of is ϕ = 2 because it is only two hop communication. The communication time for the LTE based communication network is
Fig. 6. A seven cell user mobility model. 29
Journal of Network and Computer Applications 116 (2018) 24–34
S.K. Biswash, D.N.K. Jayakody
ξ ,U
TLTE = MCpi
T ⎛ ⎜ ∑ ⎝ i = 1, j = 0
∫0
δt
CT
Tdt +
∑ ∫0
δt
j=0
⎞ Tdt , ⎟ ⎠
by (Zhang et al., 2017c). In the area-level bandwidth allocation, the average intensity or arrival rate of potential traffic in different service areas should be obtained either via theoretic analysis or actual data collection. Without loss of generality, an M∕M∕m(m) queuing model is used to model the average amount of traffic in a certain service area in
(10)
where, UT = CT = 2(T{Prop} + T{Proc}) is link round-trip time is packet transmission between UE-to-eNB and CN-to-eNB, respectively. It is twice of propagation delay and processing delay.
a=
The energy consumption in cellular networks divides into two parts; network-energy and hardware-energy. In this work we focus on network-energy only. The network-energy depends on three components, the network-update energy, the packet binding energy and the packets transmission energy. These factors are proportional to hop count(ξ) and user mobility factor (MC). Thus the total energy consumption is cellular network (Ψ) is ξ
Bd (v) =
ξ
MC
(12)
MC
∑
i=1 j=0
θ
j=0
(13)
The energy consumption in LTE is ξ
(14)
4.5. The mode switching & performance calculation
δt
δt
δt
δt
δt δt DR = Fpi (Πuc ∫0 Tdt + Πuc T ) + Fpi ⎜⎛Πuc ∑ ∫ Tdt + Πuc ∫0 Tdt ⎞⎟ 0 ⎠ ⎝
(15)
The energy consumption for mode switching(ET) is MC , pi
ET = ∑0
MC , h, pi
(R + β + θ) + ∑0
∫0
H , δt
pi λi μj Πuc πSdt +
∫0
H /2, δt
(16)
F , pi
ER = ∑0
pi λi ± 1 μj ± 1 Πuc πSdt .
γ (j ) . Bton
F , h, pi
(R + β + θ) + ∑0
(R + β + θ) + pi λj μj Bd.
(24)
The signal strength associated with the relay-based communication is expressed as
(17)
The b = maxx ∈ BtonGT GR Pb , where Bton is the number of active eNB at a given time t, GT, GR and Pb is the gain of a transmission, receiver and transmission power of eNB, respectively. The call arrival per unit area is λ (j ) λ(j), service rate per unit area μ(j), thus the traffic rate γ (j ) = μ (j) . The eNB load density is B (ℓ) =
(23)
The energy consumption for relay assisted communication is given by
(R + β + θ) + pi λj μj Bd
The signal strength for UE or CN-to-eNB, during the communication is
S(v) =
(22)
where PU(RN) is the performance metric for relay node, DR is delay associated with relay node, ER is the energy consumption by relay node, SRN signal strength for relay node and BdRN shows the bandwidth consumption of relay node. Where, P(v) > PU(RN). To prove the robustness of work, we are taking the larger values as P(v).
DT = MCpi (Πuc ∑1 ∫0 Tdt + Πuc ∫0 Tdt ) + MCpi (Πuc ∫0 Tdt + Πuc ∫0 Tdt )
(20)
(21)
PU (RN ) = DR ∧ ER ∧ SRN ∧ BdRN ,
The proposed work is performance calculation and its association with the mode switching scheme. The network performance calculates in order to energy consumption, and delay related to CN-to-eNB and UE-to-eNB. The delay for the mode switching (DT) is δt
if P (v ) > Thrs ⇒ UM () Cellular−Mode()
Thrs is the threshold value (it is decided by service provider). This relay improves the system performance in order to delay, energy, signal strength, bandwidth. Here, the performance is estimated in one direction (a post connection performance estimation)
ΨLTE = 2 ∑ ∑ (R + β + θ) i=1 j=0
(18)
(19)
P (v ) < Thrs ⇒ Add Relay Node ,
MC
H
pi λi μj Πuc BdSAdt .
In the 5G cellular network, the eNB can start the device centric communication (Boccardi et al., 2014b). Where, the associated eNB can initiate the communication, and its enable the intelligent network management. A mobile user associate with eNB, and node receive the network signal. The node revert back the signal with individual subscriber identity. Thus, eNB has sufficient information and network in order to estimates the performance of called and caller mobile users. Therefore the associated eNB determine the relay node or initiate the relay node section procedure.
The Energy consumption in UM is
ΨUM = Πuc ∑ ∑ (R + β + θ) + Πuc
H /2, δt
4.6. Relay selection procedure
MC
i=1 j=0
∫0
In this work, P(v) represents the performance metric for UE or CN. It is the commutative value with delay, energy utilization and bandwidth consumption.
ΨCM = Πuc ∑ ∑ (R + β + θ) + Πuc ∑ ∑ θ i=1 j=0
λi μj Πuc BdSAdt +
True, Mode Switching = ⎧ ⎨ False , ⎩
where R = 2ξUb + C , is the unit energy for binding vary from various transmission media and device manufactures model. C is constant h startup energy for individual devices. The θ = ∑i = 1 (Dp + S{req} + S{rep} ) is the singling cost for location update incurs with the transmission cost and processing costs. Then the cost per unit of distance can be computed as We consider the network in a time-slotted fashion and denote the length of a time slot by t. Let Ei,j denote the energy consumption on link Li,j and it will depend on distance, user mobility, bandwidth utilizations, hop count, distance and user mobility. Then the energy consumption in CM is MC
H , δt
where, ET and DT is the energy consumption and associated delay during the transmission of data packet between source and destination nodes, respectively. The S(v) and Bd(v) is signal strength and bandwidth consumption between the source and target node, respectively.
(11)
i=1 j=0
∫0
P (v ) ⇒ ET ∧ DT ∧ Sv ∧ Bd v
MC
ξ
SA
Thus, the overall performance metric function P(CN∕UE) is
∑ ∑ (R + β + θ),
ξ
, where m is index of network performance indicator. The
Bandwidth allocation is Bd = ∑a = 1 bct , δt . Then Bandwidth consumptions between the source and target node is
4.4. Energy consumption
Ψ=
mγ (m) m! γ (m) T ∑m = 0 m!
∑T m=1
S(RN ) =
∫0
2δt
pi ∨ i ± 1 λi μj πSdt +
∫0
2, δt
pi ∨ i ± 1 λi ± 1 μj ± 1 πSdt.
(25)
The bandwidth utilization for relay based communication is
Bd (RN ) =
Hence, Bandwidth utilization is given 30
∫0
δt
λi μj BdSAdt +
∫0
2, δt
pi λi μj BdSAdt.
(26)
Journal of Network and Computer Applications 116 (2018) 24–34
S.K. Biswash, D.N.K. Jayakody
the important factors for mobility management. The associated delay in the mobility (D(MM)) is the sum of the delay in the link UE-to-RN-to-CN and twice the communication time between UE or CN-to eNB. Then the delay for mobility management is
F is the two directional mobility function, because during the relay selection procedure, the node will make the connections with node and T , RN , δt 2 tdt and TR is the minimum destination only. F = ξ ∑i = 1 pi ∫0 threshold value for UM communication.
D (MM ) = D (Z ) + 2D (Ā).
4.7. Mobility management procedure
where D(Z) is the communication time between the UE to CN with the RN (within the UM of communication). D (Ā is the communication time between the UE or CN to the associated enB.
The efficient mobility management procedure help to improve the network performance metrics. During the communication, if any node (source, destination and relay node) changes their position, then the associated performance may degrade, and it affect the QoS and QoE. In the proposed work, the mobility management techniques is responsible for the manage of UM mode and relay node based communication, and if it is less than the pre-set threshold value (as decided by the network service provider), then the communication will revert back into the traditional mode. The Algorithm 4 help for mobility management procedure. During the mobility management procedure, the signal strength, energy consumption and delay will be calculated between the source node, destination node and relay node, because the nodes are following the user-centric communication. These outcome will additive to the associated signal strength, energy consumption and delay between the eNB-to-CN and UE-to-eNB, because the eNB has the observation over the communication. Here, we are calculating the overall performance metrics. Sometime, the performance values between the source and target nodes is good, but the performance metric between the UE-to-eNB is poor then it reduces the overall metrics. Thus we are calculating additive metrics.
P (Z ) > Thr ⇒ UM ()
3
H
⎞ ξ + θ + θξ − 1 (ξ − 2). ⎠
3
ξ
γ (x )
(29)
CMR =
0
⎠
(30)
(31)
The energy consumption for the mobility management (E(MM)) is consumed energy between the UE-to-RN-to-CN and two times the UE or CN-to-eNB, because it provides the up-link and down-link connections. Then we have
E (MM ) = E (Z ) + 2E (Ā),
E (Z ) = EC =
(32)
δt
max γx , Ux (λ x ) max eNB (ℓ), t (μx )
(37)
(38)
and
1 1 μA1 = max Si + x , Bdi + x ⎛ UEx MC + CNx MC ⎞. 2 ⎝2 ⎠
MC
(39)
When the user is moving back from the adjacent cell to its own cell, the basic call information will be associated with the previous visited cell. The current cell will fetch the information from the last visited cell and the rate is
∑ ∑ (R + β + θ) i=1 j=0
δt
δt , B on
1 1 λA1 = max Si, Bdi ⎛ UEx MC + CNx MC ⎞, 2 ⎝2 ⎠
The E(Z) and E(Ā) represents energy consumption between UE to CN with relay nodes, and between UE or CN to associated eNB. where the δt generic value of E is Ei = ξBdi (Dp + S{req} + R {rep} ) ∫0 (T ) . h
H
The λx and μx represents the actual call arrival departure ratio at a given time and location x, the x ⊂ A. It is depends on mobility, population density, speed and bandwidth utilization. Here, we can write λ x = max Si, Bdi (Πuc UEx MC + Πuc CNx MC ) and μx = max Si, Bdi (Πuc UEx MC + Πuc CNx MC ) , we are setting CMR < 1. Let, λA1 λA2 λA3 and muA1 μA2 μA3 three stage of actual arrival and departure rates as: initials, indeterminate and last/threshold/final stage value. In the 5G network, the cell area is very small, following the extreme densification of eNB techniques. The call arrival and departure rate of a mobile user at its first neighbor (the mobile user moving to next cell from its own cell) cell is
⎞
∑ T ⎟.
δ
eNB (ℓ) = B on , the link utilization is U = ∫0, δt b. eNB (ℓ) dt , where δt is t call duration. Now, we can compute the performance of the network with call arrival, call departure, link utilization as well energy consumptions. Then CMR is
The signal strength between UE or CN-to-eNB is
⎛ S (Ā) = Fpi ⎜∑ T 2 + ⎝ i=1
3
The cellular network provide the any time any where feature to mobile subscriber. The mobility management procedure is responsible to provide the pervasive network infrastructure to mobile user. If a mobile user change the associated eNB, then it need a fresh location update and location registration with new eNB, and it is manage by mobility management. The mobility management is classified into two types, network initiated and user initiated. The modern cellular system follows the network initiated procedure, and in this work we are considering the same for mobility management. The call-to-mobility (CMR) has a important role in network based mobility management. In this section, we provide the analysis of proposed work with the CMR in order to test the scalability of contribution. In the cellular network, the call arrival departure rate is dependent on signal coverage user mobility rate, bandwidth utilization and available services to users. Then the λ (x ) traffic rate at location x is γ (x ) = μ (x ) . The eNB load density
⎟
Bd
δt
4.8. Call-to-mobility ratio & robustness
ξ
⎜
(35)
(36)
where F′ is the associated factor for signal strength, depending on the size of the data packet, hop count, processing time, bandwidth consumption etc.
F=⎛ ⎝
δt
D (Ā) = Fpi (∑1 ∫0 Tdt + ∑0 ∫0 t Tdt ) + Fpi (∑1 ∫0 Tdt + ∫0 Tdt )
Where, Ā is signal strength between UE or CN to associated eNB. The S (MM) and S(Z) is signal strength for mobility management and between UE-to-RN-to-CN, respectively. The S(Z) is
Dp + S{req} + S{rep}
3
Similarly, the communication delay between the UE or CN-to-eNB is
(28)
3
δt
+ Fpi (Πuc ∑1 ∫1 Tdt + Πuc ∫0 Tdt ).
P(Z) represents the performance metrics during the communication between the UE and CN with relay nodes. In order to maintain node mobility and network performance, the system checks the performance for better QoE. Then the signal strength for mobility management is
⎛ ⎞ S (Z ) = F ′pi ⎜Πuc ∑ T 2dt + Πuc ∑ T ⎟, 0 ⎝ i=1 ⎠
δt
H
D (Z ) = DT = Fpi (Πuc ∑1 ∫0 Tdt + Πuc ∫0 Tdt )
(27)
S (MM ) = S (Z ) + 2S (Ā),
(34)
(33)
Node movement and efficient network service during the movement are 31
Journal of Network and Computer Applications 116 (2018) 24–34
S.K. Biswash, D.N.K. Jayakody Πu
c ⎛ λA2 = max Si, Bdi Πuc UEx ∑ MC + Πuc CNx ⎜ 1 ⎝
Πuc
⎞
∑ MC ⎟, 1
(40)
⎠
and Πu
c ⎛ μA2 = max Si + x , Bdi + x ⎜Πuc UEx ∑ MC + Πuc CNx 1 ⎝
Πuc
⎞
∑ MC ⎟. 1
⎠
(41)
Here, the user is moving from neighboring cell to neighboring cell, so the related call arrival and departure rate will depends on the previously associated cell and its mobility factors. Πu
c ⎛ λA3 = max Si, Bdi Πuc UEx ∑ MC + Πuc CNx ⎜ 1 ⎝
Πuc
⎞
∑ MC ⎟, 1
(42)
⎠
and
Fig. 10. The link utilization in bit/sec verses user movement probability pi. Πuc
⎛ μA3 = max Si + 2x , Bdi + 2x ⎜Πuc UEx ∑ MC + Πuc CNx 1 ⎝
Πuc
⎞
∑ MC ⎟. 1
⎠
(43)
5. Results and discussions In this section we provide the numerical results of the proposed user-centric 5G communication network. We consider the initial parameters as follows. The hop counts is 2, user movement directions is 6, initial movement probability 0.01, call duration is 120 s, distance from last location is 20 m, and the λi = 0.01, μi = 0.001. The network antenna gain for transmitting and receiving is 45 dBm and 43 dBm, respectively. The active eNB is 20, and least 10 active users per cell, and cell-radius is 50 m, user-mobility-distance threshold is 5 cells crossing, the average interference margin is 5.5 dB. Fig. 9 shows the link utilization in bit/sec verses hop counts, where pi = 0.001, d = 20 m. The proposed work has less link utilization because the CM/UM has frequent mode switching based on performance observation. The proposed model has less link utilization for any number of active users in a cell, as compared to an LTE network. Fig. 10 shows the link utilization in bit/ sec verses user movement probability. The d vary from 20 to 2400 m. The proposed work has significant advantages over LTE networks. In LTE, the call should always transfer to eNB and it has observation between the communication. In the proposed model, there is less link utilization than in LTE, because the high movement probability has less influence. The node mobility leads to relay node selection and relay based communication. This results in less link utilization. Fig. 11 shows the communication delay in microsecond versus user movement probability. In the proposed work, the delay will be minimum because the UM has only two hops of communication as compared to LTE networks. The UM follows the direct communication or relay based
communication. The user-mode has less signal exchange between the eNB and the network. Fig. 12 shows the associated delay for communication in microsecond versus user hop counts during the communications. In the UM, the delay is minimum because the it has least hop communication, high degree of movement probability, and less signalling overheads during the communication. The CM mode of communication refers to switching between the traditional and user-centric mode. After the mode switching procedure the call (between the UE-toCN) is only two hops or three hops(in relay assisted communication). The proposed work does not follows the eNB-centric communication,
Fig. 9. The link utilization in bit/sec verses hop counts.
Fig. 12. The associated delay for communication verses hops count.
Fig. 11. The associated delay for communication verses hops count.
32
Journal of Network and Computer Applications 116 (2018) 24–34
S.K. Biswash, D.N.K. Jayakody
Fig. 16. Link Utilization(bit/sec)vs call-to-mobility ratio.
Fig. 13. The Energy consumption in Joules verses user movement probability.
Fig. 14. The Hop counts vs energy consumption in Joules.
Fig. 17. Link Utilization(bit/sec)vs call-to-mobility ratio.
minimum energy consumption. Call-to-Mobility is an important factor to achieve QoE and scalability of the proposed system. Fig. 15 represents the energy consumption for communication vs the Call-toMobility. The range of CMR varies from 0.12 to 1.56. High user mobility and heavy network utilization leads to maximum energy consumption. In the proposed work, the system has less network dependency and always aims for user centric communication. Thus the proposed UM achieves lower energy consumption than the LTE network. Finally, the proposed work is more robust than the current legacy cellular network. Fig. 16 shows the link utilization in CM and UM vs CMR. The proposed work has less link utilization than LTE, with a high degree of mobility and user-network-utilization. Because the proposed work has only partial dependency on eNB, it leads to less signal overheads and network utilization. Fig. 17 shows call arrival and call departure rates in a cell vs CMR vs link utilization. It illustrates the performance of the proposed work, where the user is very dynamic and busy-in-network. The UM and CM method has better performance, because it does not guarantees the QoE/QoS limited to an eNB or network.
Fig. 15. Consumed Energy in Joules vs Call-to-Mobility Ratio.
thus it has less delay than LTE networks. Fig. 13 shows energy consumption (in joules) for the communication vs user movement probability. The proposed work employs user-centric communication, has less dependency on network infrastructure and follows the minimum hops communication. These factors lead to lower signalling costs, overheads and energy consumption. Fig. 14 represents the energy consumption (in joules) vs traveled hops counts. This figure demonstrates that the proposed work achieves lower energy consumption. Minimum hops of communication lead to lower signalling costs and
6. Conclusion In this paper, we introduced a user-centric communication methodology for 5G networks. It incorporates a dynamic performance-based 33
Journal of Network and Computer Applications 116 (2018) 24–34
S.K. Biswash, D.N.K. Jayakody
mode switching and mobility management scheme. The network computes the QoS/QoE parameters such as link utilization, delay and energy consumption for each mobile user. If these parameters are above the pre-set threshold value (equivalent to LTE network), the call will be switched to user-mode communication, and associated eNB has a performance observation over the communication. If the performance is not up to the mark, the network can add relay nodes in order to improve the user experience. Where the user-centric communication network performance is extremely poor the moved the traditional mode of communication. We analyzed the proposed system performance over various factor and observed it has major improvement over the LTE network in order to achieve robustness and scalability.
framework: an enabler for cooperative and D2D communication for future 5G networks. IEEE Commun. Surv. Tutor. 18 (1), 419–445 Firstquarter. Rowell, C.L.I.C., Han, S., Xu, Z., Li, G., Pan, Z., February 2014. Toward green and soft: a 5G perspective. IEEE Commun. Mag. 52 (2), 66–73. Schulz, P., Matthe, M., Klessig, H., Simsek, M., Fettweis, G., Ansari, J., Ashraf, S.A., Almeroth, B., Voigt, J., Riedel, I., Puschmann, A., Mitschele-Thiel, A., Muller, M., Elste, T., Windisch, M., February 2017. Latency critical iot applications in 5G: perspective on the design of radio interface and network architecture. IEEE Commun. Mag. 55 (2), 70–78. Soltanmohammadi, E., Ghavami, K., Naraghi-Pour, M., Dec 2016. A survey of traffic issues in machine-to-machine communications over LTE. IEEE Internet Things J. 3 (6), 865–884. Tang, R., Zhao, J., Qu, H., Zhang, Z., 2017. User-centric joint admission control and resource allocation for 5g d2d extreme mobile broadband: a sequential convex programming approach. IEEE Commun. Lett. 21 (7), 1641–1644. Tehrani, M.N., Uysal, M., Yanikomeroglu, H., May 2014. Device-to-Device communication in 5G cellular networks: challenges, solutions, and future directions. IEEE Commun. Mag. 52 (5), 86–92. Thompson, J., Ge, X., Wu, H.C., Irmer, R., Jiang, H., Fettweis, G., Alamouti, S., February 2014. 5G wireless communication systems: prospects and challenges [guest editorial]. IEEE Commun. Mag. 52 (2), 62–64. Yazlcl, V., Kozat, U.C., Sunay, M.O., Nov 2014. A new control plane for 5G network architecture with a case study on unified handoff, mobility, and routing management. IEEE Commun. Mag. 52 (11), 76–85. Zhang, H., Chu, X., Guo, W., Wang, S., 2015. Coexistence of wi-fi and heterogeneous small cell networks sharing unlicensed spectrum. IEEE Commun. Mag. 53 (3), 158–164. Zhang, C., Ge, J., Li, J., Gong, F., Ding, H., 2016. Complexity-aware relay selection for 5G large-scale secure two-way relay systems. IEEE Trans. Veh. Technol. PP (99) 1–1. Zhang, H., Liu, N., Chu, X., Long, K., Aghvami, A.H., Leung, V.C.M., 2017a. Network slicing based 5g and future mobile networks: mobility, resource management, and challenges. IEEE Commun. Mag. 55 (8), 138–145. Zhang, H., Qiu, Y., Chu, X., Long, K., Leung, V.C.M., 2017b. Fog radio access networks: mobility management, interference mitigation, and resource optimization. IEEE Wireless Commun. 24 (6), 120–127. Zhang, Q., Fu, B., Feng, Z., Li, W., Jan 2017c. Utility-maximized two-level game-theoretic approach for bandwidth allocation in heterogeneous radio access networks. IEEE Trans. Veh. Technol. 66 (1), 844–854. Zhang, H., Chen, Y., Liu, Y., 2018. Spatial correlation based analysis of power control in user-centric 5g networks. IET Commun. 12 (3), 326–333.
Acknowledgment This work was funded, in part, by the Ministry of Education and Science of the Russian Federation Grant No. #2.3649.2017/4.6 and, in part, by the framework of Competitiveness Enhancement Program of the National Research Tomsk Polytechnic University No. TPU CEP_IC_110\2017. References Abu-Ali, N., Taha, A.E.M., Salah, M., Hassanein, H., 2014. IEEE Commun. Surv. Tutor. 16 (3), 1239–1265 Third. Agiwal, M., Roy, A., Saxena, N., 2016. Next generation 5G wireless networks: a comprehensive survey. IEEE Commun. Surv. Tutor. 18 (3), 1617–1655 thirdquarter. Agyapong, P.K., Iwamura, M., Staehle, D., Kiess, W., Benjebbour, A., Nov 2014. Design considerations for a 5G network architecture. IEEE Commun. Mag. 52 (11), 65–75. Ameigeiras, P., Ramos-munoz, J.J., Schumacher, L., Prados-Garzon, J., Navarro-Ortiz, J., Lopez-soler, J.M., March 2015. Link-level access cloud architecture design based on SDN for 5G networks. IEEE Network 29 (2), 24–31. Andrews, J.G., Buzzi, S., Choi, W., Hanly, S.V., Lozano, A., Soong, A.C.K., Zhang, J.C., June 2014. What will 5G be? IEEE J. Sel. Area. Commun. 32 (6), 1065–1082. Araniti, G., Campolo, C., Condoluci, M., Iera, A., Molinaro, A., May 2013. Lte for vehicular networking: a survey. IEEE Commun. Mag. 51 (5), 148–157. Arshad, R., ElSawy, H., Sorour, S., Al-Naffouri, T., Alouini, M., 2017. Velocity-aware handover management in two-tier cellular networks. IEEE Trans. Wireless Commun. PP (99) 1–1. Biswash, S.K., Kumar, C., 2010. Multi home agent and pointer-based (MHA–PB) location management scheme in integrated cellular-wlan networks for frequent moving users. Comput. Commun. 33 (18), 2260–2270. Biswash, S.K., Kumar, C., 2011. An efficient metric-based (EM-B) location management scheme for wireless cellular networks. J. Netw. Comput. Appl. 34 (6), 2011–2026. Boccardi, F., Heath, R.W., Lozano, A., Marzetta, T.L., Popovski, P., February 2014a. Five disruptive technology directions for 5G. IEEE Commun. Mag. 52 (2), 74–80. Boccardi, F., Heath, R.W., Lozano, A., Marzetta, T.L., Popovski, P., 2014b. Five disruptive technology directions for 5g. IEEE Commun. Mag. 52 (2), 74–80. Chen, S., Qin, F., Hu, B., Li, X., Chen, Z., April 2016a. User-centric ultra-dense networks for 5g: challenges, methodologies, and directions. IEEE Wirel. Commun. 23 (2), 78–85. Chen, Z., Li, T., Fan, P., Quek, T.Q.S., Letaief, K.B., Aug 2016b. Cooperation in 5G heterogeneous networking: relay scheme combination and resource allocation. IEEE Trans. Commun. 64 (8), 3430–3443. Chen, B., Chen, J., Gao, Y., Zhang, J., 2017. Coexistence of LTE-LAA and wi-fi on 5 ghz with corresponding deployment scenarios: a survey. IEEE Commun. Surv. Tutor. 19 (1), 7–32 Firstquarter. “Cisco 4G/LTE descriptions”. https://www.cisco.com. Costa-Perez, X., Garcia-Saavedra, A., Li, X., Deiss, T., de la Oliva, A., di Giglio, A., Iovanna, P., February 2017. A. Moored. 5G-Crosshaul: an SDN/NFV integrated fronthaul/backhaul transport network architecture. IEEE Wirel. Commun. 24 (1), 38–45. Fan, P., Zhao, J., Chih-Lin, I., 2016. 5G high mobility wireless communications: challenges and solutions. China Commun. 13 (Suppl. 2), 1–13. Giust, F., Cominardi, L., Bernardos, C.J., January 2015. Distributed mobility management for future 5G networks: overview and analysis of existing approaches. IEEE Commun. Mag. 53 (1), 142–149. Hoang, T.D., Le, L.B., Le-Ngoc, T., Oct 2016. Resource allocation for D2D communication underlaid cellular networks using graph-based approach. IEEE Trans. Wireless Commun. 15 (10), 7099–7113. Liu, J., Au, K., Maaref, A., Luo, J., Baligh, H., Tong, H., Chassaigne, A., Lorca, J., 2018. Initial access, mobility, and user-centric multi-beam operation in 5g new radio. IEEE Commun. Mag. 56 (3), 35–41. Mishra, P.K., Pandey, S., Biswash, S.K., 2016a. Efficient resource management by exploiting D2D communication for 5G networks. IEEE Access 4, 9910–9922. Mishra, P.K., Pandey, S., Biswash, S.K., 2016b. A Device-Centric scheme for relay selection in a dynamic network scenario for 5G communication. IEEE Access 4, 3757–3768. Mustafa, H.A.U., Imran, M.A., Shakir, M.Z., Imran, A., Tafazolli, R., 2016. Separation
Sanjay Kumar Biswash received his Ph. D. degree from the Indian Institute of Technology (Indian School of Mines), Dhanbad, 2012. Dr. Biswash currently working as a Assistant Professor, NIIT University, Neemrana, Rajasthan, India and a Research Scientist, at Department of Software Engineering, School of Computer Science and Robotics, National Research Tomsk Polytechnic University, Tomsk, Russia. He held Postdoc position at the National Laboratory of Scientific Computing (LNCC), RJ Brazil, and San Diego State University, CA, USA. He was a visiting scientist to the University of Coimbra, Portugal. He was an Assistant Professor (on contract) at the Motilal Nehru National Institute of Technology, Allahabad, India. His research interests lies on Network Management, Mobility Management, 5G, Device-centric Networks, Wireless and Mobile Networks, and Bio-inspired Networks. He serves as an editorial board member for Int. Journal of Wireless Personal Communications (Springer Verlag), Int. Journal of Mobile and Network Application (Springer Verlag) and reviewer of many reputed international journals/conferences.
Dushantha Nalin K. Jayakody (S’09, M’14) received the Ph. D. degree in Electronics, Electrical, and Communications Engineering in 2014, from the University College Dublin, Ireland. He received his MSc degree in Electronics and Communications Engineering from the Department of Electrical and Electronics Engineering, Eastern Mediterranean University, Turkey (under the University full graduate scholarship) and ranked as the first merit position holder of the department, and B. E. electronics engineering degree (with first-class honors) from Pakistan and was ranked as the merit position holder of the University (under SAARC Scholarship.). From 2014 - 2016, he was a Postdoc Research Fellow at the Institute of computer science, University of Tartu, Estonia and Department of Informatics, University of Bergen, Norway. From summer 2016, he is a Professor at the School of Computer Science & Robotics, National Research Tomsk Polytechnic University, Russia, where he also serves as the Director of Tomsk Infocomm Lab. Dr Jayakody has received the best paper award from the IEEE International Conference on Communication, Management and Information Technology (ICCMIT) in 2017. Dr. Jayakody has published over 80 international peer reviewed journal and conference papers. His research interests include PHY layer prospective of 5G communications, Cooperative wireless communications, device to device communications, LDPC codes, Unmanned Ariel Vehicle etc. Dr. Jayakody is a Member of IEEE and he has served as workshop chair, session chair or technical program committee member for various international conferences, such as IEEE PIMRC 2013/2014, IEEE WCNC 2014-2018, IEEE VTC 2015-2018 etc. He currently serves as a Area Editor the Elsevier Physical Communications Journal, MDPI Information journal and Wiley Internet of Technology Letters. Also, he serves as a reviewer for various IEEE Transactions and other journals.
34
