Lattice based EPON energy-saving scheme analysis
Introduction
Over the coming decade, the total IP traffic will gain about threefold, and networked devices per capita will increase to 3.5 by 2021, up from 2.3 networked devices per capita in 2016 [1]. Telecommunication networks use near 5% of the total energy of industrialized countries [2]. The Internet consumes around 1%–2.5% of the whole electricity in developed countries [3]. Moreover, it has been approved that the access network consumes around 70% of the energy in the communication network because of the large number of user’s active elements [4], [5], whereas its energy efficiency is less than 15% [6]. To achieve the green communication networks, several efforts have been made to improve the energy efficiency of the access network.
Between different access network technologies, optical access networks (OAN) technology provides huge bandwidth, which is essential to meet the worldwide network operator's requirements[7], [60], [61]. Nowadays, many people prefer to use OAN for connecting their home to the Internet, which is called fiber to the home (FTTH). The total number of FTTH/B subscribers is rapidly increasing worldwide. The FTTH/B subscribers in the Asia-pacific and European countries (EU39) was 427.7 and 71 million respectively in 2019. The FTTH/B Homes Passed in the Asia-pacific and European countries (EU39) was more than 550 and 174 million respectively. It is predicted that until 2025, FTTH/B solutions will become mainstream, followed by strong fiber adoption in the world [8], [9]. Passive optical network (PON) is realized the FTTH [10] by reducing the infrastructure cost, providing high-volume bandwidth, operator coexistence, low transmission losses and low power consumption [11], [12], [13].
The Ethernet PON (EPON) is a time-division-multiple-access (TDMA) that consists of the optical line terminal (OLT) in the provider side and multiple optical network units (ONUs) placed at the users’ premises while linked via optical fibers and optical passive splitters. The feeder fiber is shared among ONUs, thus in the upstream direction TDMA and multi-point control protocol (MPCP) as the medium access control (MAC) algorithm is used to emulate a dedicated point-to-point channel from each ONU to the OLT. The OLT marks the packets with an appropriate logical link identifier (LLID) and broadcasts the packets in the downstream direction, so each ONU must filter out those packets not directed to itself.
To allocate efficient bandwidth among the ONUs in the upstream direction, the OLT executes the dynamic bandwidth allocation (DBA) algorithm that takes into account their different needs [14], [15], [16], [17], [18]. The MPCP uses the GATE and REPORT control messages for bandwidth allocation, i.e., the REPORT message carries the ONU’s buffer status to the OLT, and the OLT calculates the granted time and reports the allocated time to the ONU by the GATE message [19]. Many DBAs are well studied based on the cyclic polling (online) DBA and REPORT intensive (offline) DBA [20], [21].
To achieve green communication networks, the hardware and software-based techniques are developed in the PON [22], [23], [24], [25]. Because of the nature of the TDMA scheme, each ONU is active, upload its data to the OLT in the granted time, and has to wait for the next cycle granted time. In other words, the ONU is idle in the upstream direction when is prohibited to upload the data, which leads to wasting the energy [26]. The ONUs consume about 60%–70% of the PON systems, thus the ONUs is the best candidate to be the primary target for power-savings [12]. The sleep mode [5], [6], [12], [13], [14], [27], [28], [29], [30], [31], [32], [33], [34], [35], [36], [37], [38], [39], [40], [41],and doze [15], [17], [28], [42], [43] are two software-based techniques to reduce an ONU’s energy consumption. The transmitter and receiver are turned off for a substantial time in the sleep technique, while only the receiver is turned off in the doze technique. The IEEE 802.3ah and 802.3av are two EPON standard which does not support energy-saving, although the IEEE P1904.1 standard for service interoperability in Ethernet passive optical networks (SIEPON) provides the energy-saving framework. The SIEPON standard provides an ONU power-saving mechanism and protocols for cyclic sleep control and defines TRx (transmitter and receiver) sleep and Tx (transmitter) sleep mode, that similar to sleep and doze techniques respectively [17]. The “Tx” or “doze” disables the transmitter of the ONU during the sleep state while the receiver is always on. The “TRx” or “sleep” disables the transceiver of the ONU during the sleep state [16]. Definitely, “TRx” or “sleep” mode is more energy-efficient than the “Tx” or “doze” mode, but the ONU loses the OLT link synchronization, thus ONU requires a long clock recovery time when it wakes up [15]. Furthermore, some modification to the MPCP is required to implement the sleep mode.
The bounded QoS network parameters such as jitter, latency, and packet loss ratio is required by the emerging application (such as video, voice) [18]. Thus the doze mode can apply to save the energy and avoids the sleep mode delay problem with incorporate of DBA, control policy and frame coalescing [17]. The sleep/doze time duration is the main challenge issue in the PON energy-saving domain. Choosing the long doze/sleep time significantly decreases the QoS metrics, but it increases the energy-saving parameter and vice-versa [39].
A SIEPON-based energy-saving mechanism is proposed in which the ONU initiates and calculates the doze duration, forwards it to the OLT; then, the OLT calculates the sleep duration and decides the ONU’s enters to doze or sleep mode [6]. The energy-efficient dynamic bandwidth allocation (EE-DBA), uses the MPCP to investigate the eligibility of ONUs to enter a power-saving mode based on the current queue and packet delay factors [12]. The authors of [15] are proposed an energy-saving with “swift” and “laggy” doze mode; in which “swift” requires a very short transition time from the doze to normal mode, while “laggy” requires larger transition time. The cyclic sleep controller with a feedback control technique that observes a queue length (QL) of the downstream buffer and determines a sleep period proportional to the difference between a target QL and the observed QL is proposed in [16]. This paper clarifies the design methodology of the control parameter for QoS-aware cyclic sleep control with a QL-based proportional (P) controller and proposes a QL-based proportional-derivative (PD) controller to improve the power saving performance [16]. A new two-stage energy-saving mechanism for ONU doze mode is proposed in [17] that doze duration is calculated based on ONU’s queue state and incoming traffic rate, which can guarantee the QoS parameters. The sleep control scheme enables sleep mode with downstream traffic scheduling rules at the OLT and sleep control schemes at ONUs that ONUs can infer their downstream queue status and OLT can accurately infer the sleep status of ONUs. [27]. Furthermore, An adaptive delay-aware energy efficient (ADAEE) with aims of energy-saving and boundary delay requirements is proposed by redesigning ONU architecture and new algorithms to achieve maximum and minimum sleep interval values [31]; however, redesigning the hardware architecture introduces initial cost. The watchful sleep mode unifies the doze and cyclic sleep modes into a single power management mode that eliminates the extra control signaling and maximizes the amount of energy-saving [43]. The QoS aware tri-mode energy-saving scheme is proposed that allows the ONU to switch to one of the energy-saving modes whenever no upstream traffic exists [44]. Table 1 summarizes the recent studies and it’s characteristics in TDM-PON energy-saving. Unfortunately, many researchers have been done in TDM-PON energy-saving, but clearly, there is a lack of the mathematical model for finding the acceptable sleep/doze duration to achieve the energy-saving and at the same time guarantee the QoS metrics.
This paper introduces a new mathematical model based on the lattice theory for TDM-PON energy-saving. The model inputs different functions, parameters, and conditions, and computes the acceptable sleep/doze duration domain value. The model is built for all kinds of TDM-PON energy-saving, and the doze mode energy saving is analyzed and compared with the system without an energy-saving mechanism.
Lattices and related ordered structures are essential in algebra, analysis, logic, topology, computer science, geometry, combinatorics, and category theory. In computer science, to express a recursion when functions are in the lattice, the fixed point theory is employed [45]. Many nonlinear applications such as optimization theory, game theory, and differential equations can be solved using the fixed point theorems [46].
The rest of this paper is organized as follows. Section II describes the lattices and proposes a new mathematical model for finding the appropriate sleep/doze domain duration to achieve the optimum energy-saving without scarifying the QoS metrics. Section III presents the numerical result of section II for doze mode energy-saving. We conclude our work in Section IV.
Section snippets
Lattice model
To fully understand the lattice, the partial order on a set (Poset), meet, and join terminology is defined as follows.
“The binary relation on a set is a partial order on is reflexive, antisymmetric, and transitive. Therefore, the set together with a partial order on is called a partially ordered set (poset) [47].”
Let be a poset and let and be any pair of elements of the meet and join are define as:
- a)
The meet of and, denoted by, is the maximum of all lower bounds for and;
Conclusion
The common software-based techniques to improve energy-saving in the time-division multiplexing passive optical network (TDM-PON) are turning off the ONU’s transmitter/receiver for a substantial time, which is named sleep/doze mode. EPON and GPON standards introduced the energy-saving mechanism, but all of them leave the sleep/doze duration to provide flexibility in implementation and operation. Many studies have proposed to find the sleep time duration that is a critical value in the TDM-PON
CRediT authorship contribution statement
Maryam SafaeiSisakht: Data curation, Conceptualization, Writing - original draft, Investigation, Validation. AliAkbar Nikoukar: Conceptualization, Methodology, Software, Writing - original draft, Investigation, Validation. Hamidreza Goudarzi: Methodology, Formal analysis, Writing - original draft. I-Shyan Hwang: Writing - review & editing. Andrew Tanny Liem: Writing - review & editing, Software.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (61)
- et al.
A SIEPON based transmitter sleep mode energy-efficient mechanism in EPON
Optical Fiber Technol.
(2015) - et al.
Grant management procedure for energy saving TDM-PONs
Optical Fiber Technol.
(2018) - et al.
An adaptive two-stage energy-efficiency mechanism for the doze mode in EPON
Optical Fiber Technol.
(2016) - et al.
Genetic expression programming-based DBA for enhancing peer-assisted music-on-demand service in EPON
Optical Fiber Technol.
(2015) - et al.
Traffic-aware energy saving scheme with modularization supporting in TWDM-PON
Opt. Fiber Technol.
(2017) - et al.
Performance analysis of TCP traffic and its influence on ONU’s energy saving in energy efficient TDM-PON
Opt. Fiber Technol.
(2015) - et al.
Adaptive delay-aware energy efficient TDM-PON
Comput. Netw.
(2013) - et al.
Energy-saving scheme based on downstream packet scheduling in Ethernet passive optical networks
Opt. Fiber Technol.
(2013) - et al.
Building an energy-efficient uplink and downlink delay aware TDM-PON system
Opt. Fiber Technol.
(2016) - et al.
Energy-saving mechanism based on double-sleep-state algorithm and dynamic double-threshold receiver selection in EPON
Optik-Int. J. Light Electron Opt.
(2013)
Space-efficient classical and quantum algorithms for the shortest vector problem
Quantum Inf. Comput.
On fuzzification of Tarski's fixed point theorem without transitivity
Fuzzy Sets Syst.
A survey on energy efficiency for the future internet
Int. J. Comput. Commun. Eng.
Design and analysis of green optical line terminals for TDM passive optical networks
J. Opt. Commun. Network.
Energy-efficient load adaptive polling sequence arrangement scheme for passive optical access networks
J. Opt. Commun. Network.
Decentralized SIEPON-based ONU-initiated Tx/TRx energy-efficiency mechanism in EPON
J. Opt. Commun. Network.
Energy-efficient framework for time and wavelength division multiplexed passive optical networks
J. Opt. Commun. Network.
PON roadmap
J. Opt. Communicat. Network.
Sleep-Based Energy-Efficient DBA in an EPON Using the Controlling Mechanism for an ONU’s Buffer Occupancy and Packet Delay
J. Opt. Communicat. Network.
Early wake-up decision algorithm for ONUs in TDM-PONs with sleep mode
J. Opt. Commun. Network.
Multi-power-level energy saving management for passive optical networks
J. Opt. Commun. Network.
On the use of the doze mode to reduce power consumption in EPON systems
J. Lightwave Technol.
QoS-aware cyclic sleep control with proportional-derivative controllers for energy-efficient PON systems
J. Opt. Commun. Network.
Dynamic bandwidth allocation with SLA awareness for QoS in ethernet passive optical networks
J. Opt. Communicati. Network.
Integrated dynamic bandwidth allocation for low buffer aggregated passive optical network systems
J. Opt. Commun. Network.
Offline and online multi-thread polling in long-reach PONs: A critical evaluation
J. Lightwave Technol.
Joint resource allocation and software-based reconfiguration for energy-efficient OFDMA-PONs
J. Opt. Commun. Network.
Cited by (4)
A predictive TDM-PON resource allocation using the ANN method based on equilibrium points of discrete dynamical systems
2023, Optical Fiber TechnologySleep-aware wavelength and bandwidth assignment scheme for TWDM PON
2021, Optical and Quantum Electronics