**4. Bandwidth slicing in mobile backhaul networks**

In previous sections, service migration strategy and fog computing resource management scheme have been investigated to support real-time vehicular services. As indicated, mobile backhaul capacity is a main factor that affects the performance of the service migration schemes. Regarding this matter, passive optical network (PON) based mobile backhaul network can be considered to support FeCN due to its high capacity.

In PON-based mobile backhaul network supporting the FeCN, the BS-Fogs are integrated with optical network units (ONUs) through high-speed Ethernet interface, shown in **Figure 10**. The traffic generated by service migration, named migration traffic, is transmitted together with the non-migration traffic. On the one hand, the size of the data generated by service migration can be up to hundreds of MBytes [12]; thus, such migration traffic can be fragmented into Ethernet frames at ONUs and should be carefully handled by the optical line terminal (OLT) that is located at the central office. On the other hand, service migration is usually deadline-driven and has to be handled within a certain time limit. We define migration delay as the time duration from the moment when a migration is initiated until the affected service is successfully transferred to the target fog node. In order to minimize the service interruption, the migration delay should be lower than a pre-defined time threshold, which is usually specified in the QoS requirements and in a magnitude of seconds [13]. The non-migration traffic includes data generated by multi-type applications, which usually have different QoS requirements but less stringent in terms of latency, packet loss ratio, etc. These different types of the nonmigration traffic can be queued independently and scheduled with different priorities according to medium access control (MAC) protocol in Ethernet PON [14].

**Figure 8(a)** shows that one-hop access probability for HP services versus service arrival rate. As expected, due to the fact that the migration delay for both HP and LP services can be reduced by enlarging backhaul capacity, larger transmission capacity (*B*) leads to higher one-hop access probability, which benefits the reduction of migration time. As also shown, when B is large (e.g., *B* = 200Mbps), a higher number of neighbors (*N)* lead to a better one-hop access probability, while when *B* is small, the increase of *N* has little impact on the one-hop access probability for HP services. This is because with a smaller *B*, the backhaul delay between the target and the neighboring BS-Fogs is higher, and the number of neighboring BS-fogs that satisfy the latency requirement decreases, even when N is high. **Figure 8(b)** shows LP service unavailability as a function of service arrival rate. Similarly, increasing backhaul capacity *B* leads to a lower migration delay and thus a reduced service

*Moving Broadband Mobile Communications Forward - Intelligent Technologies for 5G …*

We further compare the performance of the proposed ORM scheme with two benchmarks. The first benchmark is based on the principle of first come first served (FCFS), in which HP and LP services are treated equally. The second benchmark is FRR where a certain amount of resource is reserved for HP. **Figure 9(a)** shows that, in comparison to FCFS and FRR, the one-hop access probability of HP services for ORM is higher when *B* is large (e.g., *B* = 200 Mbps). When *B* decreases to 100 Mps,

unavailability.

**Figure 9.**

**36**

*One-hop access probability and service unavailability versus service arrival rate.*

section. In each polling cycle, request messages are first sent from each ONU to the OLT containing the information about their data size and delay requirements. According to such information, the polling tables (see **Figure 10**) for both nonmigration data and migration data are then updated. Once service migration occurs, the lengths of the slices for both migration and non-migration traffic can be calculated by the resource management controller located at the OLT with the bandwidth

*Low-Latency Strategies for Service Migration in Fog Computing Enabled Cellular Networks*

**Figure 11** illustrates in more detail the proposed bandwidth slicing mechanism. Following similar principles of the considered DBA algorithms, in each polling cycle, the time slots in Slice 1 and Slice 2 are allocated to the non-migration traffic and the migration data, respectively. As mentioned, the lengths of the slices are decided dynamically based on the traffic and resource allocation algorithms. In the case where there is no migration data to be transmitted, the proposed mechanism performs in the same way as the classical DBA mechanism with a FCFS fashion. Note that the same principle as the proposed mechanism applies to both the

Following the proposed bandwidth slicing mechanism, a tailored delay-aware resource allocation algorithm is proposed with the aim to transmit migration traffic within the required deadline by cutting the large-size migration traffic into small

implemented at the end of Slice 1 in each polling cycle. First, the OLT acquires the information of the amount of the non-migration traffic that ONUs need to send in the next polling cycle and their priorities contained in the Report messages. If service migration occurs, such messages also contain the information of the sizes and deadline of the migration data that ONUs need to send. Then, the length of the slice for the migration traffic ( *S<sup>m</sup>* ) can be calculated by OLT, and the migration mechanism will be triggered. In such a case, as illustrated in **Figure 11** the bandwidth slicing mechanism, the polling cycle will be divided into two slices for nonmigration traffic (Slice 1) and migration traffic (Slice 2), respectively. The time slots in Slice 1 are allocated to ONUs for non-migration traffic according to their priority level, while the time slots in Slice 2 are allocated to the migration traffic according to the ascending order of the deadline for finishing migration. Here time slot allocation is purely based on incoming traffic, and with high migration traffic load, it can be expected that the time slots are monopolized. To avoid such situation,

pieces and transmitting them at each polling cycle. Such an algorithm is

allocation algorithms and the information contained in the polling table.

upstream and downstream.

**Figure 11.**

**39**

*Illustration of the proposed bandwidth slicing mechanism.*

**4.2 Delay-aware resource allocation algorithm**

*DOI: http://dx.doi.org/10.5772/intechopen.91439*

**Figure 10.** *PON-based mobile backhaul for the FeCN.*

Likewise, the migration traffic can be also queued based on their deadline requirements.

Time and wavelength division multiplexing (TWDM-PON) has been regarded as a promising candidate for next-generation PON 2 (NG-PON 2), where dynamic bandwidth allocation (DBA) mechanisms are performed on each wavelength for efficient channel sharing [15]. In a classical DBA algorithm, migration traffic and non-migration traffic are scheduled with no distinction. Each ONU reports to the OLT the amount of data that needs to be transmitted in the next cycle and then receives a grant message. According to the information contained in the grant message, including the allocated time slots and polling cycle for transmission, each ONU transmits data based on the principle of FCFS without considering the traffic priorities. Once a service migration occurs, a large volume of migration data arrives, and more than one polling cycle may be needed for the transmission. In such a case, the non-migration data that arrives after the migration data has to wait before being transmitted, thus experiencing a long queuing delay, leading to high latency and jitter. One way to deal with this is to assign higher priority to non-migration traffic that arrives after the migration traffic based upon the existing QoS management mechanism in Ethernet PON. In such a case, the delay for the non-migration traffic can be reduced significantly. However, short delay for transmitting the migration traffic that comes after may not be guaranteed, particularly when the load of non-migration traffic is high.

For balancing the transmission of migration traffic and non-migration traffic, we propose a dynamic bandwidth slicing (DBS) scheme with a bandwidth slicing mechanism and a tailored delay-aware resource allocation algorithm. We present the DBS in the following part together with simulation results.

#### **4.1 Bandwidth slicing mechanism**

A bandwidth slicing scheme for service migration in PON-based mobile backhaul networks is proposed [16]. In the scheme, the cycle time can be cut into several slices dynamically, which are provisioned to different kinds of traffic (i.e., migration traffic and non-migration traffic with different priorities). Such a mechanism is based on the report-grant mechanism, as introduced in the previous

#### *Low-Latency Strategies for Service Migration in Fog Computing Enabled Cellular Networks DOI: http://dx.doi.org/10.5772/intechopen.91439*

section. In each polling cycle, request messages are first sent from each ONU to the OLT containing the information about their data size and delay requirements. According to such information, the polling tables (see **Figure 10**) for both nonmigration data and migration data are then updated. Once service migration occurs, the lengths of the slices for both migration and non-migration traffic can be calculated by the resource management controller located at the OLT with the bandwidth allocation algorithms and the information contained in the polling table.

**Figure 11** illustrates in more detail the proposed bandwidth slicing mechanism. Following similar principles of the considered DBA algorithms, in each polling cycle, the time slots in Slice 1 and Slice 2 are allocated to the non-migration traffic and the migration data, respectively. As mentioned, the lengths of the slices are decided dynamically based on the traffic and resource allocation algorithms. In the case where there is no migration data to be transmitted, the proposed mechanism performs in the same way as the classical DBA mechanism with a FCFS fashion. Note that the same principle as the proposed mechanism applies to both the upstream and downstream.

#### **4.2 Delay-aware resource allocation algorithm**

Following the proposed bandwidth slicing mechanism, a tailored delay-aware resource allocation algorithm is proposed with the aim to transmit migration traffic within the required deadline by cutting the large-size migration traffic into small pieces and transmitting them at each polling cycle. Such an algorithm is implemented at the end of Slice 1 in each polling cycle. First, the OLT acquires the information of the amount of the non-migration traffic that ONUs need to send in the next polling cycle and their priorities contained in the Report messages. If service migration occurs, such messages also contain the information of the sizes and deadline of the migration data that ONUs need to send. Then, the length of the slice for the migration traffic ( *S<sup>m</sup>* ) can be calculated by OLT, and the migration mechanism will be triggered. In such a case, as illustrated in **Figure 11** the bandwidth slicing mechanism, the polling cycle will be divided into two slices for nonmigration traffic (Slice 1) and migration traffic (Slice 2), respectively. The time slots in Slice 1 are allocated to ONUs for non-migration traffic according to their priority level, while the time slots in Slice 2 are allocated to the migration traffic according to the ascending order of the deadline for finishing migration. Here time slot allocation is purely based on incoming traffic, and with high migration traffic load, it can be expected that the time slots are monopolized. To avoid such situation,

#### **Figure 11.**

*Illustration of the proposed bandwidth slicing mechanism.*

Likewise, the migration traffic can be also queued based on their deadline

*Moving Broadband Mobile Communications Forward - Intelligent Technologies for 5G …*

Time and wavelength division multiplexing (TWDM-PON) has been regarded as a promising candidate for next-generation PON 2 (NG-PON 2), where dynamic bandwidth allocation (DBA) mechanisms are performed on each wavelength for efficient channel sharing [15]. In a classical DBA algorithm, migration traffic and non-migration traffic are scheduled with no distinction. Each ONU reports to the OLT the amount of data that needs to be transmitted in the next cycle and then receives a grant message. According to the information contained in the grant message, including the allocated time slots and polling cycle for transmission, each ONU transmits data based on the principle of FCFS without considering the traffic priorities. Once a service migration occurs, a large volume of migration data arrives, and more than one polling cycle may be needed for the transmission. In such a case, the non-migration data that arrives after the migration data has to wait before being transmitted, thus experiencing a long queuing delay, leading to high latency and jitter. One way to deal with this is to assign higher priority to non-migration traffic that arrives after the migration traffic based upon the existing QoS management mechanism in Ethernet PON. In such a case, the delay for the non-migration traffic can be reduced significantly. However, short delay for transmitting the migration traffic that comes after may not be guaranteed, particularly when the load of

For balancing the transmission of migration traffic and non-migration traffic, we propose a dynamic bandwidth slicing (DBS) scheme with a bandwidth slicing mechanism and a tailored delay-aware resource allocation algorithm. We present

A bandwidth slicing scheme for service migration in PON-based mobile backhaul networks is proposed [16]. In the scheme, the cycle time can be cut into several slices dynamically, which are provisioned to different kinds of traffic (i.e., migration traffic and non-migration traffic with different priorities). Such a mechanism is based on the report-grant mechanism, as introduced in the previous

the DBS in the following part together with simulation results.

requirements.

*PON-based mobile backhaul for the FeCN.*

**Figure 10.**

non-migration traffic is high.

**4.1 Bandwidth slicing mechanism**

**38**

