3. Protection architectures for PON

Different protection architectures are proposed to facilitate fault detection and restoration in PON. This section discusses five pre-planned protection architectures, which vary in terms of topology, fiber duplication, and devices at both feeder level and ODN.

#### 3.1. ITU-T 983.1 type C architecture

since it is directly associated with the extent and cost of the network. Number of subscribers is primarily effected by the type of topology at both feeder and ODN along with devices at the CO and RN. For example, a typical tree-star topology can accommodate more subscribers as compared to a conventional ring-based architecture due to the use of 1 : N POC. Whereas, the

OLT 5:12e<sup>7</sup> 12100 20 W, 25 W (with EDFA), 2 W (standby)

**OLT OS FB OCP :N OCP ONU**

**Optical Components**

**99.9980**

Figure 1. Connection availability of basic optical components in PON.

Optical circulator 3e<sup>7</sup> 50 Optical switch 1:2e<sup>6</sup> 50 1:2, 2:2 POC 3e<sup>7</sup> 50 1:N, 2:N POC 7:2e<sup>7</sup> 800 Fiber (\$=Km) 1:3e<sup>5</sup> 160

Table 1. Components description, connection availability, and cost.

System components Unavailability Cost (\$) Energy consumption

ONU 1:54e<sup>6</sup> 350 1 W, 0.25 W (standby)

**99.9982**

**99.9984**

**99.9986**

**Connection**

**Availability**

 **(%)**

**99.9988**

**99.9990**

**99.9992**

**99.9994**

**99.9996**

**99.9998**

**100.0000**

342 Optical Fiber and Wireless Communications

ITU-T 983.1 type C is a pre-planned protection architecture, which provides fault detection and restoration throughout the network with redundant components at both feeder and the distribution levels [17]. The basic type C PON is shown in Figure 2, where each component of the network is duplicated to ensure high connection availability and fast restoration time. OLT is placed at the CO and consists of two transceiver modules, where one acts as primary ðOLTpÞ and another is set as a secondary ðOLTsÞ module. Under normal mode of operation, OLTp is responsible for originating and managing services across the network, whereas OLTs activates in the event of failure at OLTp module.

Each transceiver module at OLT is connected to a corresponding FF, namely primary ðFFpÞ and secondary ðFFsÞ. Both fibers extend the network toward the subscribers' premises. Under normal mode of operation, OLTp is connected to the FFp. Two FFs are used to provide maximum connection availability and fast restoration time, such that FFp is immediately replaced with FFs in the event of failure. The span of each fiber is about 20 km for a standard PON. Both FFs terminate into RN, which serves as chases for two 1 : N POC modules connected with the corresponding FF at the input port and N DFs at the output port, respectively, where N represents the number of subscribers.

Figure 2. Type C protection architecture.

Consequently, a total of 2N DFs are utilized to provide the required protection at the ODN. DFs terminate into ONU with two transceiver modules, ðONUp;ONUsÞ, in order to facilitate abrupt fault detection and restoration. Type C protection architecture duplicates every component including OFs to provide abrupt fault detection and restoration. Furthermore, a two-fiber tree-based topology is laid at the feeder level, whereas a star-based topology is adapted to implement the 2N DFs at the ODN.

Number of subscribers in type C protection architecture can be determined through power budget analysis from OLT toward ONU module. Power budgeting in PON ensures an efficient communication between the transmitting and receiving modules. Moreover, it also determines the POC splitting ratio, which translates the number of subscribers in PON, respectively. If Pi represents the power loss across each component i, α is attenuation=km in the OF medium, and Rsen represents the receiver sensitivity, then the number of subscribers for type C protection architecture can be determined as:

$$P\_T - P\_{\rm OUT} - aP\_{\rm FF} - P\_{\rm POC} - aP\_{\rm DF} - P\_{\rm ONU} - P\_{\rm mis} \ge R\_{\rm sen} \tag{1}$$

Equation (1) shows that major power loss occurs across the POC, which is determined by POPC ¼ 10Log10ðNÞ, where N represents the number of subscribers or splitting ration of 1 : N POC. Therefore, to maintain received power Pre ≥Rsen, the value of N must be adjusted to facilitate extended reach, fault detection, and restoration along with high bandwidth connectivity. For example, when the transmitter power is 10 dBm, <sup>α</sup> ¼ �0:<sup>25</sup> dB km, N ¼ 128, 25 km fiber, and Pmis ¼ �3 dBm, the approximate power received at the PIN photo-diode will be Pre ¼ �20 dBm. Consequently, for Rsen ¼ �25 dBm, type C architecture can efficiently support 128 subscribers simultaneously accessing the medium.

#### 3.2. Single ring architecture

In order to avoid extensive duplication of the OF medium at both feeder and ODN, ring-based topology is utilized to implement PON with desired connection availability and fault detection/ restoration between OLT and ONU modules [16]. The basic ring-based PON is shown in Figure 3, which contains OLT module at the CO. In order to avoid the high deployment cost of OLT module, this architecture employs a single unit, owing to the fact that failure per year of OLT module is minuscule [4, 18]. End-face of the OLT module is connected to a switching arrangement ðSAcoÞ that extends the OF medium in both clockwise (CW) and counter clockwise (CCW) directions. SAco serves as chases for a 1 : 2 POCsa, and 1 : 2 OSsa. POCsa sends and receives the traffic for OLT module. Port 1 splits the optical signal toward the clockwise feeder ring ðFRcwÞ, whereas port 2 extends the flow of traffic toward the counter clockwise FR ðFRccwÞ through a 1 : 2 OSsa. This arrangement recovers the flow of information in case of failures at the FR. Under normal working conditions, OSsa is at port a and both upstream and downstream traffic are carried on the FRcw.

FR contains multiple ONU modules ONU f g n; n ¼ 1; 2; 3…N , which are placed directly over the FR through individual RNs ðRNxÞ housing a 2 : 2 POC<sup>x</sup>;<sup>1</sup> and a 1 : 2 POC<sup>x</sup>;2, respectively. Consequently, the total number of ONU modules is equal to RNs on the FR X ¼ N. POC<sup>x</sup>;<sup>1</sup> is used to extend the FR to the neighboring ONU modules. Furthermore, it also connects individual ONU<sup>n</sup> with FR through POC<sup>x</sup>;<sup>2</sup> and controls the flow of traffic in and out of the ONU module.

Network capacity in single ring-based architecture is significantly effected by the utilization of a single POC per subscriber. If N represents the number of RNs, then the total subscribers accessing the medium simultaneously can be determined by:

$$P\_T - P\_{\text{OLT}} - P\_{\text{POC}\_{\text{un}}} - P\_{\text{OS}\_{\text{un}}} - 2 \,\text{XP}\_{\text{POC}\_{\text{van}}} - P\_{\text{ONU}} - P\_{\text{mis}} \ge R\_{\text{sen}} \tag{2}$$

X represents the number of RNs in Eq. (2), which is equal to ONUs N that can access the medium simultaneously. Consequently, for the transmitter power of 10, �3 dBm loss across the POC, Pmis ¼ �3 dBm, 25 km fiber, and Rsen ¼ �25 dBm, this architecture can scarcely support 16 subscribers simultaneously accessing the medium [6]. It is observed that fiber duplication is avoided at both levels through a ring-based structure; however, power drop across each POC

Figure 3. Single ring protection architecture.

Consequently, a total of 2N DFs are utilized to provide the required protection at the ODN. DFs terminate into ONU with two transceiver modules, ðONUp;ONUsÞ, in order to facilitate abrupt fault detection and restoration. Type C protection architecture duplicates every component including OFs to provide abrupt fault detection and restoration. Furthermore, a two-fiber tree-based topology is laid at the feeder level, whereas a star-based topology is adapted to

DFp

DFs

Number of subscribers in type C protection architecture can be determined through power budget analysis from OLT toward ONU module. Power budgeting in PON ensures an efficient communication between the transmitting and receiving modules. Moreover, it also determines the POC splitting ratio, which translates the number of subscribers in PON, respectively. If Pi represents the power loss across each component i, α is attenuation=km in the OF medium, and Rsen represents the receiver sensitivity, then the number of subscribers for type C protec-

Equation (1) shows that major power loss occurs across the POC, which is determined by POPC ¼ 10Log10ðNÞ, where N represents the number of subscribers or splitting ration of 1 : N POC. Therefore, to maintain received power Pre ≥Rsen, the value of N must be adjusted to facilitate extended reach, fault detection, and restoration along with high bandwidth connec-

fiber, and Pmis ¼ �3 dBm, the approximate power received at the PIN photo-diode will be Pre ¼ �20 dBm. Consequently, for Rsen ¼ �25 dBm, type C architecture can efficiently support

In order to avoid extensive duplication of the OF medium at both feeder and ODN, ring-based topology is utilized to implement PON with desired connection availability and fault detection/

tivity. For example, when the transmitter power is 10 dBm, <sup>α</sup> ¼ �0:<sup>25</sup> dB

128 subscribers simultaneously accessing the medium.

P<sup>T</sup> � POLT � αPFF � PPOC � αPDF � PONU � Pmis ≥ Rsen ð1Þ

km, N ¼ 128, 25 km

ONU1

ONUN

ONUNp ONUNs

ONU1p ONU1s

implement the 2N DFs at the ODN.

Figure 2. Type C protection architecture.

344 Optical Fiber and Wireless Communications

FFp

1 : N

O P C

1 : N

O P C

FFs

OLT p OLT s

tion architecture can be determined as:

3.2. Single ring architecture

introduces serious power budget issues. Consequently, capacity of the PON is effected. It is therefore imperative to utilize a topology that is capable of providing fault detection and restoration without compromising the network capacity and total cost.

#### 3.3. Dual ring protection architecture

Dual ring PON avoids fiber duplication at the feeder level through a single ring-based fiber ðFRÞ as shown in Figure 4 [9]. Furthermore, several small rings are deployed at ODN to negate the power budget issue. CO contains a single OLT module, which is fed into an optical circulator ðOCcoÞ and erbium-doped fiber amplifier (EDFA) module. To maintain passive nature of the PON, EDFA is placed inside the CO as shown in Figure 4. Switching arrangement SAco is placed at end-face of the EDFA, which serves as chases for 1 : 2 POCsa and 1 : 2 OSsa. Port 1 of POCsa splits the down-stream traffic towars FRcw, whereas port 2 extends the flow of traffic towars the FRccw through a 1 : 2 OSsa. In normal mode of operation, all traffic is carried through the FRcw. However, OLT medium access control (MAC) layer flips the switch position when a failure occurs at the FR. This new position of the switch converts the ringbased fiber in two trees-based fibers carrying traffic in clock- and counter clockwise direction between the PoF.

Both FRcw and FRccw are fed into multiple remote nodes forming several distribution rings ðDRsÞ at the ODN. Each remote node RN f g x; x ¼ 1; 2; 3…X houses two 2 : 2 bidirectional POCs, namely ðPOC<sup>x</sup>;<sup>1</sup>;POC<sup>x</sup>;<sup>2</sup>Þ as shown in Figure 4. The former extends the FR from CW toward CCW direction and the latter carries the traffic from feeder toward the ODN and vice versa. End-face of POC<sup>x</sup>;<sup>2</sup> splits the incoming traffic into clock- and counter clockwise DRs ð Þ DRcw and ð Þ DRccw , respectively. Both DRcw and DRccw are connected to individual ONU

Figure 4. Dual ring protection architecture.

ONU<sup>x</sup>;m; m <sup>¼</sup> <sup>1</sup>; <sup>2</sup>; <sup>3</sup>…<sup>M</sup> module through two 1:2 POConu<sup>x</sup>;<sup>m</sup> and OSonu<sup>x</sup>;<sup>m</sup> . Under normal working conditions, OSonu<sup>x</sup>;<sup>m</sup> is at position a and all traffic is handled through DRcw.

Dual ring architecture utilizes multiple rings at the distribution level to compensate the excessive power drop. Furthermore, EDFA is also employed at the CO to support high capacity transmission. Consequently, the power budget equation from OLT toward ONU module can be written as:

$$P\_T - P\_{\rm OUT} - P\_{\rm OC\_{wo}} + P\_{\rm EDFA} - P\_{\rm PCC\_u} - P\_{\rm OS\_u} - \alpha P\_{\rm FR} - 2XP\_{\rm PCC\_{\rm BS}} - \alpha P\_{\rm DR} - MP\_{\rm PCC\_{\rm own}} - P\_{\rm ON} \ge R\_{\rm sen} \tag{3}$$

Where X, in Eq. (3), represents the total number of RNs deployed at the FR and M represents the number ONUs per RN. Consequently, the total ONU modules accessing the medium simultaneously become N ¼ X � M. Now, at PT ¼ 10 dBm, 25 km fiber, and Pedfa ¼ 25 dB, the maximum value of N that can be achieved is 72 with X ¼ 9 and M ¼ 8 [9]. It is observed that the dual ring architecture supports more ONU modules in comparison with the single ring architecture. Nevertheless, power drops in the ring topology at the distribution level limits the overall capacity of the network.

#### 3.4. Tree-based hybrid protection architecture

introduces serious power budget issues. Consequently, capacity of the PON is effected. It is therefore imperative to utilize a topology that is capable of providing fault detection and

Dual ring PON avoids fiber duplication at the feeder level through a single ring-based fiber ðFRÞ as shown in Figure 4 [9]. Furthermore, several small rings are deployed at ODN to negate the power budget issue. CO contains a single OLT module, which is fed into an optical circulator ðOCcoÞ and erbium-doped fiber amplifier (EDFA) module. To maintain passive nature of the PON, EDFA is placed inside the CO as shown in Figure 4. Switching arrangement SAco is placed at end-face of the EDFA, which serves as chases for 1 : 2 POCsa and 1 : 2 OSsa. Port 1 of POCsa splits the down-stream traffic towars FRcw, whereas port 2 extends the flow of traffic towars the FRccw through a 1 : 2 OSsa. In normal mode of operation, all traffic is carried through the FRcw. However, OLT medium access control (MAC) layer flips the switch position when a failure occurs at the FR. This new position of the switch converts the ringbased fiber in two trees-based fibers carrying traffic in clock- and counter clockwise direction

Both FRcw and FRccw are fed into multiple remote nodes forming several distribution rings ðDRsÞ at the ODN. Each remote node RN f g x; x ¼ 1; 2; 3…X houses two 2 : 2 bidirectional POCs, namely ðPOC<sup>x</sup>;<sup>1</sup>;POC<sup>x</sup>;<sup>2</sup>Þ as shown in Figure 4. The former extends the FR from CW toward CCW direction and the latter carries the traffic from feeder toward the ODN and vice versa. End-face of POC<sup>x</sup>;<sup>2</sup> splits the incoming traffic into clock- and counter clockwise DRs ð Þ DRcw and ð Þ DRccw , respectively. Both DRcw and DRccw are connected to individual ONU

FR

RN1

RNx

a 2 O P C

b

1

CO

Figure 4. Dual ring protection architecture.

2

1

PO

C 2,1

RN2

POC2,1

C P O DR

cw

OS

b

a

DRccw

C P O Distribution Ring

ONU 2,2

ONU 2,1

ONU2,M

ONURX ONUTX

restoration without compromising the network capacity and total cost.

3.3. Dual ring protection architecture

346 Optical Fiber and Wireless Communications

between the PoF.

OLTTX

OLTRX

This architecture employs a hybrid topology through combination of tree and star-ring architectures at the feeder and ODN, respectively. OLT is placed at the CO, which is fed into an OCco, EDFA, and SAco as shown in Figure 5 [15]. SAco consists of a 1 : 2 OSsa, with port a connected to the FFp, whereas port b is fed into FFs. In normal working conditions, OSsa is at port a and all upstream and down-stream traffic is sent and received through the FFp. However, in case of failures or cuts at the FFp, OLT MAC layer flips the switch position and resumes the flow of traffic through the FFs.

Since a tree-based topology is adapted at the feeder level, the required protection is provided by duplicating the long-span fiber. End-face of both FFp and FFs terminates into RN that serves as chases for a 2 : N POC. Output of the 2 : N POC is connected to a series of dedicated DFs connecting N ONU transceiver modules as shown in Figure 5. In order to avoid extensive duplication of the DFs for protection, a ring-based topology is adapted by connecting intermediate ONU modules, respectively. Consequently, each ONU<sup>n</sup> module, where f g n ¼ 1; 2; 3…;N , consists of 1 : 2 POC<sup>n</sup> and 1:2 OSn. Port 1 of the POC is connected to respective DF, whereas port 2 is fed into a 1:2 OSn. Port 3 extends the DF toward a DR that provides the necessary protection in case of failure at the DF.

OS<sup>n</sup> is used to connect OC<sup>n</sup> and ONU<sup>n</sup> with transmission media at the distribution level. Port "a" of OS<sup>n</sup> connects ONU<sup>n</sup> with the DF, whereas port "b" is used to connect ONU<sup>n</sup> with the redundant DR in case of failure. Consequently, the required protection is achieved by avoiding the extensive duplication of the DFs. Under normal mode of operation, OS<sup>n</sup> is at port "a" and all services are delivered through the DF, whereas in case of failure the affected ONU<sup>n</sup> MAC

Figure 5. Tree-based hybrid protection architecture.

layer flips the switch position and transfers the flow of traffic to the DR between the adjacent ONU module.

The downstream power budget for tree-based hybrid protection architecture can be written as:

$$\begin{aligned} &P\_{\rm T} - P\_{\rm OLT} - P\_{\rm OC\_{\rm w}} + P\_{\rm EDFA} - P\_{\rm OS\_{\rm w}} - \alpha P\_{\rm FF} - 10 \text{Log}\_{10}N - \alpha P\_{\rm DF} \\ &- P\_{\rm POC\_{\rm w}} - P\_{\rm OS\_{\rm w}} - P\_{\rm OC\_{\rm w}} - P\_{\rm ONU} \ge R\_{\rm Sen} \end{aligned} \tag{4}$$

It is observed from Eq. (4) that for P<sup>T</sup> ¼ 10 dBm, 25 km fiber, Pedfa ¼ 25 dB, and Rsen ¼ �25 dBm, the tree-based hybrid protection architecture can efficiently support 128 subscribers simultaneously accessing the medium [15].

#### 3.5. Ring-based hybrid protection architecture

This architecture is formed by the combination of a single ring topology at the feeder level and a star-ring architecture at the ODN. A single OLT is placed at the CO, which is connected to an OCco, EDFA, and SAco as shown in Figure 6 [18]. This architecture utilizes a single ring-based fiber to provide the required high bandwidth connectivity and reliability at the feeder level. Port 1 and 2 of the POCsa extend traffic toward the FRcw and FRccw, respectively. A 1:2 OSsa is connected with port 2 of the POCsa to provide the required fault recovery in the event of failure. Port a of the OSsa is connected to ground, whereas port b is fed into the FRccw. Under normal mode of operation, OSsa is at port a and all traffic is sent and received through the FRcw.

FR is implemented at each RN by combining FRcw and FRccw through a special arrangement. This significantly reduces the power budget penalty and cost of the overall architecture. RN

Impact of Fiber Duplication on Protection Architectures Feasibility for Passive Optical Networks http://dx.doi.org/10.5772/intechopen.68237 349

Figure 6. Ring-based hybrid protection architecture.

layer flips the switch position and transfers the flow of traffic to the DR between the adjacent

2:N CPR

O P

C

2 : N

FFp

FFs

DFs DR

ONU1

ONUN

ONU N-1

ONU6

ONU2

ONU3

ONU4

a

DR

POC

DF b

3

2

1

ONU5

The downstream power budget for tree-based hybrid protection architecture can be written as:

P<sup>T</sup> � POLT � POCco þ PEDFA � POSco � αPFF � 10Log10N � αPDF

It is observed from Eq. (4) that for P<sup>T</sup> ¼ 10 dBm, 25 km fiber, Pedfa ¼ 25 dB, and Rsen ¼ �25 dBm, the tree-based hybrid protection architecture can efficiently support 128 subscribers

This architecture is formed by the combination of a single ring topology at the feeder level and a star-ring architecture at the ODN. A single OLT is placed at the CO, which is connected to an OCco, EDFA, and SAco as shown in Figure 6 [18]. This architecture utilizes a single ring-based fiber to provide the required high bandwidth connectivity and reliability at the feeder level. Port 1 and 2 of the POCsa extend traffic toward the FRcw and FRccw, respectively. A 1:2 OSsa is connected with port 2 of the POCsa to provide the required fault recovery in the event of failure. Port a of the OSsa is connected to ground, whereas port b is fed into the FRccw. Under normal mode of operation, OSsa is at port a and all traffic is sent and received through the FRcw. FR is implemented at each RN by combining FRcw and FRccw through a special arrangement. This significantly reduces the power budget penalty and cost of the overall architecture. RN

ð4Þ

ONUTx

ONU2

ONURx

� PPOConu � POSonu � POConu � PONU ≥Rsen

ONU module.

OLTRx

OLTTx

simultaneously accessing the medium [15].

OS

Figure 5. Tree-based hybrid protection architecture.

EDFA

348 Optical Fiber and Wireless Communications

b

a

3.5. Ring-based hybrid protection architecture

in the ring-based hybrid network RN f g x; x ¼ 1; 2; 3…X consists of two POCs, namely ðPOC<sup>x</sup>;<sup>1</sup>;POC<sup>x</sup>;<sup>2</sup>Þ. Where POCx;<sup>1</sup> and POCx;<sup>2</sup> are bidirectional couplers with 2 : 2 and 2:M ports respectively. FR is formed by connecting both CW and CCW paths through port 1 and 2 of the POC<sup>x</sup>;<sup>1</sup> as shown in Figure 6. POC<sup>x</sup>;<sup>1</sup> further extends the FR toward the ODN through a 2 : M POC<sup>x</sup>;2, where M is the number of ONUs connected to each POC<sup>x</sup>;<sup>2</sup> through dedicated DFs. If X represents the total number of RNs deployed over the FR and M is the number ONUs per RN, then the proposed system can support a total of N ¼ X � M ONUs.

ONU module, ONU<sup>x</sup>;m; m <sup>¼</sup> <sup>1</sup>; <sup>2</sup>; <sup>3</sup>…<sup>M</sup> , starts with a 1:2 POC<sup>x</sup>;<sup>m</sup> with three ports as shown in Figure 6. Port 1 is used to connect each ONU to its dedicated DF, whereas port 2 is fed into 1:2 OS<sup>x</sup>;<sup>m</sup>. Port 3 is used to form a backup DR, which ensures immediate survivability of traffic between ONUs and OLT. Port a of the OS<sup>x</sup>;<sup>m</sup> connects ONU<sup>x</sup>;<sup>m</sup> with corresponding DF, under normal working conditions, and port "b" is fed into the DR.

The number of subscribers in ring-based hybrid protection architecture can be determined as:

$$\begin{aligned} \left[P\_{\text{T}} - P\_{\text{OLT}} - P\_{\text{OC}\_{\text{ox}}} + P\_{\text{EDFA}} - P\_{\text{PCC}\_{\text{u}}} - P\_{\text{OS}\_{\text{u}}} - aP\_{\text{FR}} - XP\_{\text{PCC}\_{\text{x},1}} - 10 \text{Log}\_{10}(M) - aP\_{\text{DF}} \\ -P\_{\text{PCC}\_{\text{own}}} - P\_{\text{OC}\_{\text{own}}} - P\_{\text{ONU}} \end{aligned} \tag{5}$$

If N represents the total number of ONUs, then the value of X can be written as:

$$X \le \frac{P\_{\rm T} - R\_{\rm sen} - 16.25 - 10 \text{Log}\_{10} M + P\_{\rm EDFA}}{3} \tag{6}$$

Analysis of Eq. (6) shows that capacity of the network increases when more ONUs are placed per RN. Furthermore, it is observed that for P<sup>T</sup> ¼ 10 dBm, Pedfa ¼ 25 dB, and Rsen ¼ �25 dBm, the ring-based hybrid protection architecture can efficiently support 128 subscribers simultaneously accessing the medium [18].

## 4. Reliability analysis

Network reliability analysis is an important tool that is used to determine the overall connection availability for a given protection architecture. Reliability block diagrams (RBDs) are commonly used to determine the overall connection availability of PON, owing to its significant advantages like, accuracy, simplicity, visual impact, and flexibility. This section analyzes network reliability in terms of connection availability of the selected protection architectures through RBDs.

RBD represents each component including the OF medium as a functional block connected in series or parallel combination with adjacent blocks. Series and parallel connectivity represent the unprotected and protected components in protection architecture, respectively. The characteristic parameter of each RBD block is the asymptotic unavailability ðUiÞ of the components that represent their probability of failures. Consequently, if I represents the total number of components in PON, then overall connection availability A is given by:

$$A = 1 - \sum\_{i=1}^{I} \mathcal{U}\_i \tag{7}$$

where Ui is determined by

$$\mathcal{U}\_i = 1 - \frac{\text{MTBF}}{\text{MTBF } + \text{MTTR}} \tag{8}$$

MTBF in Eq. (8) is the mean time between failures, which is used to represent the number of failures per million hours for an optical component [19]. MTTR represents the mean time to repair, which is the time required for reparation or replacement of faulty hardware modules [5, 20, 21]. In order to determine the overall connection availability for each architecture, RBDs with series and parallel combination of functional blocks are extracted as shown in Figure 7.

If <sup>X</sup><sup>I</sup> i¼1 Ui is the summation of i components unavailability in a protection architecture, then connection availability for each survivable PON can be written as:

$$A\_{\left[\mathbb{C}\right]} = 1 - \left[ \left( \mathcal{U}\_{\text{OLT}} \times \mathcal{U}\_{\text{OLT}\_{\text{pf}}} \right) + \left( \mathcal{U}\_{\text{FF}} \times \mathcal{U}\_{\text{FF}\_{\text{pf}}} \right) + \left( \mathcal{U}\_{\text{1:N:POC}} \times \mathcal{U}\_{\text{1:N:POC}\_{\text{pf}}} \right) + \left( \mathcal{U}\_{\text{DF}} \times \mathcal{U}\_{\text{DF}\_{\text{pf}}} \right) \right]$$

$$+ \left( \mathcal{U}\_{\text{ONU}} \times \mathcal{U}\_{\text{ONU}\_{\text{pf}}} \right)$$

$$A\_{\rm [SR]} = 1 - \left[ U\_{\rm OUT} + U\_{\rm POC} + U\_{\rm OS} + \left( U\_{\rm FR} \times U\_{\rm FR\_{pt}} \right) + U\_{\rm POC} + U\_{\rm POC} + U\_{\rm ONU} \right] \tag{10}$$

Impact of Fiber Duplication on Protection Architectures Feasibility for Passive Optical Networks http://dx.doi.org/10.5772/intechopen.68237 351

$$\begin{aligned} A\_{\text{[DR]}} &= 1 - \left[ U\_{\text{OLT}} + U\_{\text{POC}} + U\_{\text{OS}} + \left( U\_{\text{FR}} \times U\_{\text{FR}\_{\text{pr}}} \right) + U\_{\text{POC}} + U\_{\text{POC}} + \left( U\_{\text{DR}} \times U\_{\text{DR}\_{\text{pr}}} \right) \right] \\\\ &+ \left( U\_{\text{POC}} \times U\_{\text{POC}\_{\text{pr}}} \right) + U\_{\text{OS}} + U\_{\text{ONU}} \Bigg] \end{aligned} \tag{11}$$
 
$$A\_{\text{[RT]}} = 1 - \left[ U\_{\text{OLT}} + U\_{\text{OS}} + \left( U\_{\text{FF}} \times U\_{\text{FF}\_{\text{pr}}} \right) + U\_{2:N\text{POC}} + \left( U\_{\text{DF}} \times U\_{\text{DR}\_{\text{pr}}} \right) + U\_{\text{POC}} + U\_{\text{OS}} + U\_{\text{ONU}} \right] \tag{12}$$

Analysis of Eq. (6) shows that capacity of the network increases when more ONUs are placed per RN. Furthermore, it is observed that for P<sup>T</sup> ¼ 10 dBm, Pedfa ¼ 25 dB, and Rsen ¼ �25 dBm, the ring-based hybrid protection architecture can efficiently support 128 subscribers simulta-

Network reliability analysis is an important tool that is used to determine the overall connection availability for a given protection architecture. Reliability block diagrams (RBDs) are commonly used to determine the overall connection availability of PON, owing to its significant advantages like, accuracy, simplicity, visual impact, and flexibility. This section analyzes network reliability

RBD represents each component including the OF medium as a functional block connected in series or parallel combination with adjacent blocks. Series and parallel connectivity represent the unprotected and protected components in protection architecture, respectively. The characteristic parameter of each RBD block is the asymptotic unavailability ðUiÞ of the components that represent their probability of failures. Consequently, if I represents the total number of

<sup>A</sup> <sup>¼</sup> <sup>1</sup> �<sup>X</sup>

Ui <sup>¼</sup> <sup>1</sup> � MTBF

MTBF in Eq. (8) is the mean time between failures, which is used to represent the number of failures per million hours for an optical component [19]. MTTR represents the mean time to repair, which is the time required for reparation or replacement of faulty hardware modules [5, 20, 21]. In order to determine the overall connection availability for each architecture, RBDs with series and parallel combination of functional blocks are extracted as shown in Figure 7.

Ui is the summation of i components unavailability in a protection architecture, then

h � �

� �

h i

þ U<sup>1</sup>:<sup>N</sup> POC � U<sup>1</sup>:<sup>N</sup> POCpt � �

I

i¼1

Ui ð7Þ

þ UDF � UDFpt

ð9Þ

ð10Þ

MTBF <sup>þ</sup> MTTR <sup>ð</sup>8<sup>Þ</sup>

þ UPOC þ UPOC þ UONU

in terms of connection availability of the selected protection architectures through RBDs.

components in PON, then overall connection availability A is given by:

connection availability for each survivable PON can be written as:

A½SR� ¼ 1 � UOLT þ UPOCþUOS þ UFR � UFRpt

þ UFF � UFFpt � �

neously accessing the medium [18].

350 Optical Fiber and Wireless Communications

4. Reliability analysis

where Ui is determined by

A½C� ¼ 1 � UOLT � UOLTpt

þ UONU � UONUpt � �i

� �

If <sup>X</sup><sup>I</sup> i¼1

$$\begin{aligned} A\_{\text{[HR]}} &= \mathbf{1} - \left[ \mathbf{U\_{OL,T}} + \mathbf{U\_{POC}} + \mathbf{U\_{OS}} + \left( \mathbf{U\_{FR}} \times \mathbf{U\_{FR\_{Pl}}} \right) + \mathbf{U\_{POC}} + \mathbf{U\_{2NDPC}} \right. \\ &+ \left( \mathbf{U\_{DF}} \times \mathbf{U\_{DR\_{Pl}}} \right) + \mathbf{U\_{POC}} + \mathbf{U\_{OS}} + \mathbf{U\_{ONU}} \end{aligned} \tag{13}$$

Figure 8 shows the overall connection availability of the selected protection architectures based on Table 1. It is observed that maximum 5 nines connection availability is provided by all architectures, owing to the efficient utilization of redundant components and OF medium throughout the network. Maximum connection availability is provided by type C protection architecture, which duplicates entire PON including OLT and ONU modules. Ring-based architecture at both feeder and ODN, which avoids the duplication of light-wave path, is also observed to provide the required availability. A single failure at the ring-based fiber converts

Figure 7. Reliability block diagrams for (a) type C, (b) single ring (SR), (c) dual ring (DR), (d) tree-based hybrid (HT), and (e) ring-based hybrid (HR) protection architectures.

Figure 8. Overall connection availability of selected protection architectures.

the network into two tree-based architectures with the flow of traffic in both clockwise and counter clockwise directions. Thus, redundancy is achieved without duplicating the entire fiber. Furthermore, it is observed that desirable connection availability can be maintained without duplicating the transceiver modules, which further helps in reducing the overall cost of deployment for such architectures.

## 5. Cost analysis

Cost figures for protection architectures are an important parameter, showing the economic benefits for a common end user at the access domain. This section determines the overall expenditure for the selected architecture while using component costs in Table 1. Cost figures for PON can be categorized as CAPEX and OPEX. CAPEX includes the investment utilized for the deployment of PON at access domain and is calculated by computing the cost expenditure on network devices along with the OF infrastructure. OPEX includes the cost incurred on the network operations from the time of deployment till replacement by a new technology [4, 15, 22, 23].

OPEX primarily includes the cost required for repairing the faulty, failed equipment (including OF medium), service interruption penalty cost that is commonly applicable for business subscribers, and energy consumed by active components at transceiver modules of OLT and ONU. Reparation cost is determined by multiplying the total downtime=year with resources required to remove the fault, which includes the number of technicians along with their wages and miscellaneous charges. Service interruption penalties include the expenditure that is spent on the fine defined in service level agreement (SLA) between network operators and subscribers. Power consumption by each component is determined by multiplying the unit price of electricity with the sum of energy consumption of all active components over the network life span [15, 23].

Analysis is performed for residential customers only, and life span of each network is taken as 20 years. Furthermore, following specifications are considered for fair analysis of the selected protection architectures


the network into two tree-based architectures with the flow of traffic in both clockwise and counter clockwise directions. Thus, redundancy is achieved without duplicating the entire fiber. Furthermore, it is observed that desirable connection availability can be maintained without duplicating the transceiver modules, which further helps in reducing the overall cost

**[C] [DR] [SR] [HT] [HR]**

**Protection Architectures**

Cost figures for protection architectures are an important parameter, showing the economic benefits for a common end user at the access domain. This section determines the overall expenditure for the selected architecture while using component costs in Table 1. Cost figures for PON can be categorized as CAPEX and OPEX. CAPEX includes the investment utilized for the deployment of PON at access domain and is calculated by computing the cost expenditure on network devices along with the OF infrastructure. OPEX includes the cost incurred on the network operations from the time of deployment till replacement by a new technology [4, 15, 22, 23].

of deployment for such architectures.

Figure 8. Overall connection availability of selected protection architectures.

5. Cost analysis

**99.990**

**99.991**

**99.992**

**99.993**

**Connection**

**Availbaility**

 **(%)**

**99.994**

**99.995**

**99.996**

**99.997**

**99.998**

**99.999**

**100.000**

352 Optical Fiber and Wireless Communications

• Per hour cost of electricity is taken to be 0.25 \$=kWh.

If N is the total number of ONUs in each PON and X represents the number of RNs in ringbased topologies at the feeder level, then the CAPEX equations (based on RBDs) for selected protection architectures can be written as:

$$\begin{aligned} \mathbf{C\_{C}} &= (\mathbf{OLT} + \mathbf{OLT}) + (2\mathbf{0} \times \mathbf{FF} \times \mathbf{FF}) + (2 \times 1 : N \, \mathbf{POC}) \, + (5 \times N \times \mathbf{DF} \times \mathbf{DF}) \\ &+ (N \times \mathbf{ONU} \times \mathbf{ONU}) \end{aligned} \tag{14}$$

$$\mathbf{C\_{SR}} = \mathbf{OLT} + \mathbf{POC} + \mathbf{OS} + (25 \times \mathbf{FR}) + (2X \times \mathbf{POC}) + (N \times \mathbf{ONU}) \tag{15}$$

$$\mathbf{C\_{DR}} = \mathbf{OLT} + \mathbf{POC} + \mathbf{OS} + \mathbf{EDFA} + (20 \times \mathbf{FR}) + (2\mathbf{X} \times \mathbf{POC}) + (\mathbf{N} \times \mathbf{DR\_{GW}}) \tag{16}$$

$$\mathbf{A\_{AR}} = \mathbf{ran} \quad \text{(17)} \quad \mathbf{(A1} \quad \mathbf{(B1)}) = (\mathbf{u} \quad \mathbf{Q3}) \quad \text{(18)}$$

$$+\left(\text{N}\times\text{DR}\_{\text{cw}}\right)+\left(2\text{N}\times\text{POC}\right)+\left(\text{N}\times\text{OS}\right)+\left(\text{N}\times\text{ONU}\right)$$

$$\begin{aligned} \mathbf{C\_{HT}} &= \mathbf{OLT} + \mathbf{OS} + \mathbf{EDFA} + (20 \times \mathbf{FF} \times \mathbf{FF}) + (2:N \, \mathbf{POC}) + (5 \times N \times \mathbf{DF}) \\ &+ (N \times \mathbf{DR}) + (N \times \mathbf{POC}) + (N \times \mathbf{OS}) + (N \times \mathbf{ONU}) \end{aligned} \tag{17}$$

$$\begin{aligned} \text{C}\_{\text{HR}} &= \text{OLT} + \text{POC} + \text{OS} + \text{EDFA} + (20 \times \text{FR}) + (\text{X} \times \text{POC}) + (\text{X} \times 2:\text{N} \, \text{POC}) \\ &+ (\text{5} \times \text{N} \times \text{DF}) + (\text{N} \times \text{DR}) + (\text{N} \times \text{POC}) + (\text{N} \times \text{OS}) + (\text{N} \times \text{ONU}) \end{aligned} \tag{18}$$

Figure 9 shows the CAPEX for each protection architecture at different number of subscribers by referring to Table 1. It is observed that ring-based architecture requires minimum CAPEX when compared to conventional tree and hybrid techniques. However, due to the power budget and capacity limitations of single and dual ring-based architectures, their analysis is performed till 16 and 64 number of subscribers, respectively. Overall analysis shows that the deployment cost for each architecture decreases as the number of subscriber increases. It is evident from the fact that total cost incurred is distributed among the number of subscribers, which results in reduced CAPEX per user as the network capacity increases. Furthermore, it is shown that type C architecture requires the highest cost, for all subscribers, in comparison with other schemes due to the extensive duplication of optical components and OF medium throughout the network.

Figure 9 shows that hybrid protection architectures provide nominal performance, for all subscribers, in terms of the deployment cost as they avoid extensive duplication of the OF medium. Figure 10 further elaborates the CAPEX required for the deployment of light wave path in type C, tree- and ring-based hybrid protection architectures for 128 subscribers in the

Figure 9. CAPEX at different number of subscribers.

network. It is observed that the main contributor for CAPEX, besides OLT module, is the DF as each subscriber requires a dedicated path for high speed connectivity. Furthermore, it is shown in Figure 10 that type C requires the highest deployment cost for DFs due to the duplication of each fiber between RN and corresponding ONU module. On the other hand, the deployment of star-ring topology at ODN significantly reduces the CAPEX, on redundancy, for hybrid protection architectures.

Figure 9 shows the CAPEX for each protection architecture at different number of subscribers by referring to Table 1. It is observed that ring-based architecture requires minimum CAPEX when compared to conventional tree and hybrid techniques. However, due to the power budget and capacity limitations of single and dual ring-based architectures, their analysis is performed till 16 and 64 number of subscribers, respectively. Overall analysis shows that the deployment cost for each architecture decreases as the number of subscriber increases. It is evident from the fact that total cost incurred is distributed among the number of subscribers, which results in reduced CAPEX per user as the network capacity increases. Furthermore, it is shown that type C architecture requires the highest cost, for all subscribers, in comparison with other schemes due to the extensive duplication of optical components and OF medium

Figure 9 shows that hybrid protection architectures provide nominal performance, for all subscribers, in terms of the deployment cost as they avoid extensive duplication of the OF medium. Figure 10 further elaborates the CAPEX required for the deployment of light wave path in type C, tree- and ring-based hybrid protection architectures for 128 subscribers in the

**[C] [SR] [DR] [HT] [HR]**

**Protection Architectures**

throughout the network.

354 Optical Fiber and Wireless Communications

**0**

Figure 9. CAPEX at different number of subscribers.

**500**

**1000**

**1500**

**Captial**

**Expenditure**

 **(\$)**

**2000**

**2500**

**3000**

**3500**

**4000**

**4500**

It is also observed that the CAPEX in hybrid protection architecture varies with the type of topology at the feeder level. Since ring-based architecture requires multiple RNs, the cost of deployment increases with the number of subscribers due to increase in RNs. Consequently, tree-based hybrid architecture requires less cost when compared to ring-based scheme for 64 and 128 subscribers. However, the difference is minuscule.

Figure 11 shows the total expenditure for protection architectures with high network capacity in terms of the number of subscribers. It is evident from the fact that each architecture provides the desired connection availability, which significantly reduces their downtime/year. Hence, minuscule amount is spent on reparation cost over the network life span. Type C protection architecture requires highest OPEX due to the duplication of transceiver modules at both OLT

Figure 10. CAPEX for OF infrastructure at 128 subscribers.

Figure 11. Total cost including CAPEX and OPEX at 128 subscribers.

and ONUs, which consumes more energy as compared to hybrid architectures with single OLT and ONU modules. Consequently, hybrid architecture requires minimum cost, collectively, as compared to the conventional protection scheme.

#### 6. Conclusion

This chapter analyzes the effect of fiber duplication, for protection, on feasibility of PON. Increase in the capacity and adaptability of PON demands for an efficient fault detection and restoration mechanism, which is able to provide the required connection availability at minimum cost. Among all components in the PON, OF medium requires significant protection due to its high rate of failure. Furthermore, it is an on-field component and constitutes a major portion of PON; therefore, it is imperative to provide pre-planned redundant paths in order to ensure swift recovery of failures at the OF medium throughout the network. Five protection architectures are considered, which mainly emphasize on the type of topology laid through the OF medium at both feeder level and ODN. ITU-T 983.1 type C and single ring-based architectures are considered to encompass the conventional tree-star and ring topology, respectively. For further analysis, dual ring architecture and tree- and ring-based hybrid protection architectures for PON are analyzed to study the effect of hybrid topologies on PON feasibility in terms of the number of subscribers, reliability, and cost. It is observed that ring-based architectures provide desirable connection availability without duplicating the OF medium; however, the use of multiple POCs to extend the FR significantly elevates the power budget issues. Consequently, the number of subscribers that can access the medium simultaneously is reduced. Hybrid topologies formed by the variation of tree, ring, and star topology provide nominal performance, in terms of all parameters, as compared to the survivable PONs based on conventional topologies. Furthermore, it is shown that duplication of the OF medium at the ODN network is more critical in comparison with the feeder level duplication. Nevertheless, both tree- and ringbased hybrid pre-planned protection architectures provide desirable network capacity and connection availability at minimum cost.
