Guido Priano
| Daniel Alberto Priano | Javier Guevara | Matías Marsicano | Fabio Sergio Bruschetti | Isabella Anguillesi | María Claudia Abeledo*
© 2024 The authors. This article is published by IIETA and is licensed under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).
OPEN ACCESS
This article presents a technical comparison of the three principal interconnection models defined in the Fixed Access Network Sharing (FANS) standard. The three models are Q-in-Q, VXLAN and MPLS. This standard has been developed by the Broadband Forum as a solution to enable multiple operators to share fiber optic infrastructures efficiently, which is a crucial aspect in the deployment of 5G and IoT networks. Each model is evaluated in terms of its scalability, complexity of deployment, capacity for traffic isolation, and suitability for specific use cases. Q-in-Q offers simplicity and compatibility with Ethernet networks, rendering it an optimal choice for them. VXLAN is distinguished by its remarkable scalability in virtualised environments and data centres. MPLS, on the other hand, is particularly suited to networks that necessitate the handling of diverse traffic types with guaranteed Quality of Service (QoS). It can be surmised from this comparison that the decision regarding the most appropriate model will depend on both the specific requirements of the operators in question and the challenges they face in integrating new technologies. The implementation of these solutions has the potential to reduce operational costs and enhance the efficiency with which network resources are utilised.
5G, FANS, GPON, IoT, MPLS, Q-in-Q, shared networks, VXLAN
In recent years, the growing demand for bandwidth, driven by the proliferation of high-definition video and the Internet of Things (IoT), has prompted the telecommunications industry to develop fiber optic infrastructure in closer proximity to the end user (FTTH - Fiber to the Home) with the objective of enhancing the speed of services [1]. This has resulted in significant capital expenditure and a number of challenges for operators, who are seeking solutions to enable the efficient sharing of these infrastructures [2].
In response to this need, the concept of FANS (Fixed Access Network Sharing) was developed, which allows multiple operators to share a single access network [3]. The Broadband Forum standardised FANS through the TR-370 standard, with the objective of offering flexible and automated management for shared networks. This facilitates the interconnection of infrastructure providers (InPs) and wholesale operators, or Virtual Network Operators (VNOs). Consequently, FANS permits the logical partitioning and isolation of shared network resources between operators, and is compatible with virtualisation, whereby control functions are migrated from dedicated network equipment to software running on generic hardware. FANS also provides Network as a Service (NaaS).
This study is specifically focused on fibre optic access networks, with particular emphasis on PON (Passive Optical Networks), and employs GPON (Gigabit PON) as the primary technology for analysis. GPON technology is founded upon the ITU-T G.984.1 standards [4]. The aforementioned standards elucidate the regulatory framework pertaining to the provisioning, protocols, maintenance, fibre installations, privacy and security. In a GPON network, a client concentrator, designated as an OLT (optical line termination), serves to interconnect disparate clients through the use of ONT (optical network terminal) terminals. The OLTs are designed with the capacity to serve up to 128 customers per fibre port. In order to achieve this capacity, each fibre optic strand is split with splitters that separate the signal and send it from one input port to multiple output ports.
In this context, the role of IoT devices is of particular significance, as they represent a key component of the most prominent applications of 5G networks, as defined by ITU-R [5], called mMTC (Massive Machine Type Communication). It is imperative that these networks are capable of supporting the anticipated billions of devices, including smart home appliances, industrial sensors, and emergency systems, all of which will be connected through low latency and high reliability technologies. This article will analyse three main interconnection models within the FANS standard. This article compares and contrasts the characteristics and technical capabilities of three interconnection models: Q-in-Q, VXLAN, and MPLS.
Clearly, these are essential technologies used in modern networks to improve traffic management, scalability and security. Each serves a different purpose, but they can also be integrated to create robust network architectures.
The following is a detailed comparison of these technologies, highlighting their functionalities, benefits and use cases, in the following order: First, an analysis of each of them is provided. Secondly, their advantages and limitations are explained. Thirdly, a comparative synthesis is made, in accordance with the work and experiments carried out in the IPlan company [6]. Finally, conclusions are drawn.
The Q-in-Q model, also referred to as provider bridging or stacked VLANs, has been formalised as IEEE 802.1ad [7], an Ethernet networking standard that was incorporated into the 1998 IEEE 802.1Q standard in 2011. The original specification permitted the insertion of a single VLAN header into an Ethernet frame; however, this technology enables the encapsulation of multiple VLAN tags within a single frame. Together, they form a tag stack, which facilitates the creation of Metro Ethernet networks that can efficiently handle large amounts of traffic and multiple carriers. The term ‘VLAN tag’ is typically employed to denote the 802.1Q VLAN header in a streamlined form. In the context of an Ethernet frame, a Q-in-Q frame is defined as a frame that has two 802.1Q VLAN headers, henceforth referred to as dual-tagged. The Q-in-Q scheme is shown in Figure 1.
Figure 1. Q-in-Q scheme
The 802.1ad standard, published in 2006, delineates the architectural framework and communication protocols for the provision of discrete instances of media access control (MAC) services to multiple independent users on a bridged local area network (LAN). This is achieved in a manner that does not necessitate collaboration between users and with minimal interaction between users and the MAC service provider. The objective is to afford customers the option of operating their own VLANs within the VLAN provided by the service provider. In this manner, the service provider is able to configure a VLAN for the customer, who is then able to treat that VLAN as if it were a trunk.
The necessity for this standard arises from the constraints inherent in its predecessor. The primary limitation of 802.1Q is its 12-bit VLAN ID field, which allows for a maximum of 4,096 tags. However, with the use of double tagging, the number of tags that can be accommodated reaches 16,777,216, which is sufficient for modern networks. The introduction of a second tag enables the execution of operations that would otherwise be inaccessible if the VLAN ID field were to be expanded from 12 bits to 24 bits or more. The presence of multiple tags allows for the modification of frames by switches, including the addition, deletion, or modification of tags.
Furthermore, a frame comprising multiple tags not only possesses multiple VLAN IDs but also multiple VLAN header bit fields. The tag stack provides an effective mechanism for Internet Service Providers (ISPs) to encapsulate customer single-tagged 802.1Q traffic with a single tag, resulting in a Q-in-Q frame at the end. The outer tag is employed for the purpose of identifying and segregating traffic from disparate clients, while the inner tag is retained from the original frame.
The creation of layer 2 (L2) tunnels and the application of Quality of Service (QoS) policies are both enabled by Q-in-Q frames. Furthermore, it is backward compatible with 802.1Q. Although 802.1ad is limited to two tags, there is no maximum limit in the standard that restricts a frame to a maximum of two tags. This allows for future growth of the protocol, as evidenced by the fact that service provider topologies utilise frames with more than two tags.
On the other hand, it is more straightforward for network equipment manufacturers to modify their equipment by creating multiple 802.1Q headers than to modify their equipment to implement a new extended VLAN ID field header that is not 802.1Q-compliant.
The encapsulation is then analysed in this model according to Figure 2. An 802.1Q header, comprising four bytes, is appended to an untagged Ethernet frame as follows: the tag is inserted between the source MAC address (SAMAC) of the untagged frame and its ethertype field. The newly inserted VLAN header's ethertype is set to 0×8100, thereby identifying the following data as a VLAN tag. A total of 12 bits are allocated for the VLAN ID, while the remaining bits within the VLAN fields are populated in accordance with the specific policy (e.g., QoS) associated with the interface on which the tag imposition occurred. Following the insertion of an 802.1Q header into an untagged frame, the original ethertype of the frame appears to have been altered to 0×8100. The original ethertype of the untagged frame (in the single tag frame) is now in close proximity to the data (payload), yet its value remains unaltered.
Figure 2. Q-in-Q encapsulation [3]
Subsequently, a second 802.1Q header is appended to a single tag frame in the following manner: the second tag is inserted in front of the first tag, which is thus situated in closer proximity to the Ethernet header than the original tag. The second tag is then inserted between the MAC SAMAC and the first tag. The second tag is assigned the default ethertype of 0×88A8, which differs from the standard 802.1Q 0×8100. A total of 12 bits are allocated for the VLAN ID, while the remaining bits within the VLAN fields are populated in accordance with the policy of the interface on which the tag imposition occurred.
The insertion of a third tag will result in its placement in front of the previous tags, in closer proximity to the Ethernet header. The original, unencoded Ethernet type of the frame is always located after all tags and adjacent to the data. In the case of an 802.3 frame, the aforementioned ethertype would be a 'length' value, denoting the length of the frame up to the end point. In the case of an 802.3 frame with an LLC header, the LLC header remains after the length field and is adjacent to the data.
Finally, in a tag stack, the push and pop operations are performed at the outer end of the tag. Consequently, the tag added by an insert operation becomes a new outer tag, while the tag to be removed is the current outer tag.
2.1 Interconnection of Q-in-Q model
The Q-in-Q model presents a network topology comprising Service VLANs (S-VLANs) and Client VLANs (C-VLANs) for each provider. The aforementioned VLANs are employed so that the client is able to utilise any VLAN, provided that it is associated with an S-VLAN that is specific to that particular provider. In order to facilitate the implementation of VLANs in the context of FANS, it is proposed that a new Operator VLAN tag, or 'Operator VLAN' (O-VLAN), be introduced as a third VLAN tag, in addition to the existing C-VLAN and S-VLAN. This enables the VNO to oversee two VLAN levels (S + C VLAN) for its service configurations, whereas the InP merely assigns the O-Tag for each VNO.
In this configuration (Figure 3), the ONTs of various ISPs may share the same VLAN on their uplink. This VLAN is then encapsulated within an S-VLAN and transmitted to the provider's network.
Figure 3. Operator, client and service VLAN scheme [3]
As illustrated in the Figure 4, the C-VLAN tag information is transmitted across the network. In the downstream direction, the carrier O-VLAN information is incorporated into the Ethernet frame at the adjacent switch, situated at the reference point (A10). The aforementioned information tags remain intact until they reach the physical access node (pAN). In contrast, in the upstream direction, the S-VLAN and O-VLAN tag information is incorporated into the C-VLAN tag within the Ethernet frame at the pAN. It is crucial to highlight that the O-VLAN data is discarded on the switch situated in proximity to the A10 reference point, whereas the S-VLAN information persists in traversing the VNO network.
Figure 4. Q-in-Q scheme in FANS [3]
2.2 Advantages of Q-in-Q
2.3 Limitations and technical challenges of Q-in-Q
The VXLAN (Virtual Extensible Local Area Network) model represents a technological advancement that enables the establishment of overlay networks over existing physical networks. The MAC Address-in-User Datagram Protocol (MAC-in-UDP) is employed to encapsulate data link layer traffic at the network layer, thereby enabling the transmission of Ethernet frames over an Internet Protocol (IP) network. The original intention was to provide the same services as a traditional VLAN, albeit with limited extensibility and flexibility. The IETF has standardised it in RFC 7348 [13], which is currently in the Informational status category.
The advent of VXLAN can be attributed to the proposals put forth by various manufacturers of networking equipment and other technologies, including Cisco, Arista Networks, Broadcom, Intel, VMware, and others. These proposals were made in an effort to circumvent some of the challenges encountered in large data processing centres (datacentres) when utilizing server virtualization, which often necessitates working in environments comprising thousands of virtual machines [14]. This gives rise to issues pertaining to the scale of MAC address tables. With regard to VLANs, the limitation arises from the 12-bit VLAN ID, which supports a maximum of 4094 distinct networks [15]. Conversely, the deployment of the Spanning Tree Protocol (STP) in data centres to prevent link-layer loops has the consequence of rendering many available links inoperable.
These issues have thus prompted the necessity for the development of a network technology that would facilitate the interconnection of physical servers via an IP network, which in turn requires the utilisation of routing protocols at the layer 3 (L3) level. This would obviate the disuse of STP-generated links to avoid L2 loops, and more complex routing strategies (such as equal-cost multi-path routing, or ECMP) could be employed to assist in the distribution of the network load across all links. Nevertheless, the utilisation of a L2 for direct communication between virtual machines would remain a requisite [13].
In light of these considerations, VXLAN was developed with the objective of addressing the scalability limitations inherent to networks based on VLANs. This is achieved through the use of a 24-bit network identifier (VNI), which allows for the creation of over 16 million virtual networks that can coexist within the same administrative domain [3]. Furthermore, VXLAN employs L3 routing, which eradicates link wastage and facilitates a more optimal utilisation of the available resources. Additionally, it eliminates the issue of the size of the MAC address tables in the switches, which were required to store the MAC addresses of all the server virtual machines that were interconnected by each switch.
VXLAN employs the MAC-in-UDP encapsulation technique (Figure 5). The link layer packet, which contains the MAC addresses of the source and destination host, is augmented with the addition of a VXLAN header. This set is incorporated into the data field of a User Datagram Protocol (UDP) datagram, with IP employed as the network layer protocol. Subsequently, the packet transmitted to the IP network is identified by source and destination MAC addresses corresponding to the (VXLAN Tunnel End Points) VTEPs situated behind the source and destination hosts, respectively. This enables the initial packet to be routed over an IP network, allowing it to reach the VXLAN hosts in a manner analogous to if it were on the same LAN:
Figure 5. VXLAN encapsulation [16]
A VXLAN scheme is comprised of three fundamental elements: the VXLAN header, the VTEP, and the VXLAN Gateways.
In a VXLAN network, each virtual network is identified by a 24-bit VNI, which allows for the creation of up to 16 million virtual networks. This makes it an optimal solution for scenarios where a substantial number of isolated networks are necessary, such as in data centers with virtual machines. It enables the coexistence of distinct logical networks on a single physical infrastructure.
The operation of VXLAN is contingent upon the utilisation of devices designated as VTEP. Such devices are responsible for the encapsulation and subsequent de-encapsulation of Ethernet traffic into UDP datagrams, which are then transmitted over the underlying IP network. VTEPs serve as both entry and exit points for VXLAN tunnels, enabling connected devices to communicate over an IP network as if they were on the same LAN (Figure 6).
Figure 6. VXLAN scheme [16]
3.1 Interconnection of VXLAN model
The VXLAN model describes a methodology whereby data is tunnelled from the OLT to the service provider. In this process of tunnelling, an identifier is employed to delineate the tunnel, and the aggregator equipment is required to merely accommodate this tunnelling. The scheme is presented in the Figure 7.
As illustrated in the schematic Figure 7, the C-VLAN tag information is transmitted through the network. In the downstream direction, the VXLAN tag information is incorporated into the Ethernet frame at the adjacent switch, which serves as a VTEP at reference point A10. The aforementioned tag information is retained at the pAN and functions as a VTEP. In contrast, in the upstream direction, the S-VLAN and O-VLAN tag information is incorporated into the C-VLAN tag within the Ethernet frame at the pAN. It is crucial to highlight that the VXLAN data is discarded on the switch situated in proximity to the A10 reference point, whereas the S-VLAN information persists in traversing the VNO network.
Figure 7. VXLAN scheme in FANS [3]
3.2 Advantages of VXLAN
3.3 Limitations and technical challenges of VXLAN
The Multiprotocol Label Switching (MPLS) model represents a standardised data transport mechanism, originally developed by the Internet Engineering Task Force (IETF) and formally defined in RFC 3031 [19]. It operates between the data link layer and the network layer of the Open Systems Interconnection (OSI) model. The MPLS model was developed with the objective of providing a unified system for data transport over networks that may include both circuit and packet traffic. The principal objective is to enhance the efficiency and velocity of routing decisions, which are based on the utilisation of labels that are assigned to data packets. It is capable of carrying a variety of traffic types, including voice and IP. MPLS was the logical successor to Frame Relay and ATM, becoming the preferred technology for high-speed data and digital voice on a single connection. It offers enhanced reliability and performance, and can potentially reduce transport costs due to its increased network efficiency and prioritisation capabilities.
The deployment of MPLS in IP networks offers a number of advantages in the context of VPN (Virtual Private Network) creation, traffic engineering, fault protection mechanisms, QoS support, multi-protocol support, and Class of Service (CoS) establishment. Its basic features include integration of L2 (data link) and L3 (network) of the OSI model, with optimisation of routing achieved through reduction in algorithmic complexity and maintenance of communication state between two nodes [20].
MPLS is a scalable, protocol-independent technology. In an MPLS network, data packets are assigned labels, which serve to identify and categorise them. The determination of packet forwarding is based exclusively on the information contained within the label, obviating the necessity for an examination of the packet itself. This enables the creation of end-to-end circuits over any transport medium and the utilisation of any protocol. The principal advantage is the elimination of reliance on a specific OSI model data link L2 technology, such as ATM, Frame Relay, SONET or Ethernet, and the avoidance of the necessity for multiple L2 networks to cater to disparate types of traffic. MPLS is a member of the packet-switched family of networks.
MPLS operates at a layer that is generally considered to fall between the traditional OSI L2 and L3 protocol. The objective of the design was to provide a unified data transport service for both circuit-based clients and packet-switched clients, offering a datagram service model. It is capable of carrying a multitude of different types of traffic in a native state.
The similarity of MPLS with ATM and Frame Relay lies in the fact that, at each hop across the network, the label value in the header is changed, in contrast to the forwarding of IP packets. The evolution of MPLS technologies has been informed by a consideration of the relative strengths and weaknesses of ATM which has resulted in MPLS becoming the dominant technology in this field. The objective has been to develop a solution with a lower overhead than ATM, while providing connection-oriented services for variable-length frames. In particular, MPLS eliminates the signalling and switching protocol that is characteristic of ATM. MPLS acknowledges that the use of small ATM cells is unnecessary in the core of modern networks, given the high speeds and lack of queuing delays observed in modern optical networks. This is in contrast to the motivation behind the cellular nature of ATM, which was to reduce delays to support voice traffic [21]. Concurrently, MPLS strives to maintain the traffic engineering and out-of-band control that initially made ATM and Frame Relay appealing for implementation in large-scale networks as can be seen in the Figure 8.
Figure 8. MPLS scheme [22]
As can be seen in the Figure 9, MPLS operates by inserting prefixes into packets with a header of their own, containing one or more labels. These are collectively referred to as a label stack, which comprises four fields for each entry.
Figure 9. MPLS encapsulation [23]
4.1 Interconnection of MPLS model
Another potential technique for managing operator data flows in a FANS scenario is a label-based switching approach, such as MPLS, as illustrated in the Figure 10.
The principal distinction between the MPLS scheme and the VLAN model is the existence of an MPLS tunnel (defined by LSP labels) that encompasses both the C-VLAN label and the S-VLAN label information.
It can be observed that the C-VLAN label information is transmitted throughout the network. In the downstream direction, the MPLS LSP information is incorporated into the Ethernet frame at the PE router situated in proximity to the A10 reference point. The aforementioned information labels are conveyed up to the point at which the pAN assumes the role of the PE router. In contrast, in the upload direction, the S-VLAN information is incorporated into the C-VLAN label within the Ethernet frame at the vAN, while the MPLS LSP information is incorporated into the Ethernet frame at the pAN. It is crucial to highlight that the MPLS label information is discarded at the switch situated adjacent to the A10 reference point, whereas the S-VLAN information persists in traversing the VNO network. Nevertheless, despite the flexibility and scalability of the MPLS network architecture, in a purely L2 evolution context, it may be advantageous to consider utilising the O-VLAN scheme instead of the L2.5 MPLS extension to the access network. Another motivation for this approach is that current access nodes may not support MPLS. However, the addition of MPLS capability may result in increased complexity and cost for the node. The above description is shown in the Figure 10.
Figure 10. MPLS scheme in FANS [3]
4.2 Advantages of MPLS
4.3 Limitations and technical challenges of MPLS
The three FANS models (Q-in-Q, VXLAN and MPLS) present distinct approaches to the shared network management and logical partitioning of resources in a fiber optic infrastructure. Each model has characteristics that render it suitable for specific scenarios, offering a range of advantages and challenges.
5.1 Scalability
In terms of scalability, the VXLAN model offers the highest capacity, allowing the creation of up to 16 million virtual networks. This is made possible by the use of a 24-bit network identifier, which is capable of accommodating a significantly larger number of networks than other models. This figure is considerably greater than the 4096 VLAN limitation of the Q-in-Q model, even when employing multiple labels. While MPLS does not facilitate the management of virtual networks in the same manner, it does provide scalability through the efficient transportation of a multitude of traffic types on a unified infrastructure. Accordingly, VXLAN is the optimal selection for scenarios where the number of virtual networks is a pivotal consideration, such as in expansive data centers.
5.2 Implementation complexity
MPLS is the more complex model in terms of implementation, as it necessitates the presence of specific infrastructure and the configuration of numerous parameters pertaining to forwarding labels and traffic engineering. Although Q-in-Q and VXLAN also present configuration challenges, both models are based on more widely used standards and are therefore more readily integrated into networks that already utilise Ethernet or IP networks. Q-in-Q is particularly straightforward to implement in Metro Ethernet networks, whereas VXLAN benefits from the use of IP routing, which offers greater flexibility in terms of the physical topology of the network.
5.3 Isolation and security
With regard to the issue of traffic isolation, the three models in question achieve this objective in disparate ways.
The Q-in-Q method employs the use of multiple VLAN tags to segregate the data traffic of disparate operators, thereby offering a robust level of isolation in scenarios where the underlying physical network is shared. In contrast, VXLAN offers enhanced isolation due to its overlay methodology and the utilisation of VNI for the identification of virtual networks. The use of labels and the capacity to accommodate VPNs make MPLS an effective solution for networks where security and quality of service are paramount.
Summarising the relevant aspects for the analysis of this point, it can be seen:
5.4 Complexity and costs considerations
The implementations carried out in the IPlan company [6] in terms of complexity and costs allowed the construction of Table 1.
Table 1. Complexity and costs resulting from the implementation carried out
|
Features |
Q-in-Q |
MPLS |
VXLAN |
|
Complexity of Implementation |
Low to moderate |
High |
Moderate to high |
|
Configuration Complexity |
Relatively simple, requires VLAN tagging |
Complex, requires detailed configuration for routing and traffic engineering |
Moderate, requires knowledge of encapsulation and network design |
|
Operational Complexity |
Low, mainly involves VLAN management |
High, ongoing management of SLAs, routing protocols, and performance monitoring |
Moderate, needs management of virtual networks and overlays |
|
Cost of Maintenance |
Low, straightforward maintenance |
High, ongoing costs for maintenance contracts and equipment upgrades |
Moderate, depends on the scale and complexity of the deployment |
|
Scalability |
High (up to 16 million VLANs) |
Moderate, limited by label space |
Very high (up to 16 million VNIs) |
|
Use Case Suitability |
The best for service providers needing VLAN extension without complex routing requirements |
Ideal for enterprises needing reliable, high-performance WAN connectivity with QoS |
Suited for cloud environments requiring flexible, scalable layer 2 networks |
5.5 Technical comparison
The comparative Table 2 is shown below as a summary comparison.
Table 2. Comparison summary
|
Features |
Q-in-Q |
MPLS |
VXLAN |
|
Primary Use Case |
Service provider VLAN extension |
Enterprise WAN connectivity |
Cloud data center virtualization |
|
Scalability |
Up to thousands of VLANs |
Limited by label space |
Up to 16 million VNIs |
|
Traffic Isolation |
Yes |
Yes |
Yes |
|
QoS Support |
Limited |
Extensive |
Limited |
|
Complexity of Implementation |
High |
Low |
Moderate |
|
Cost Consideration |
Generally low |
High, due to specialized hardware |
High, due to complexity and requirements |
5.6 Use cases
Each model has specific use cases where it is particularly effective.
Q-in-Q: Ideal for Metro Ethernet networks, where a simple and efficient way to manage multiple VLANs from different carriers in a shared infrastructure is required. In contexts where Ethernet compatibility is of paramount importance and scalability is not a primary consideration, this model is the optimal choice.
VXLAN: In scenarios that necessitate high scalability and flexibility, it is the optimal model. It is particularly beneficial in data centres and in networks where virtual machines are utilised, necessitating an overlay network to interconnect servers across IP infrastructures.
MPLS: Better suited to networks that require the handling of diverse traffic types with varying priorities. The capacity to implement traffic engineering and QoS policies renders it an optimal choice for networks where the accurate management of voice, video and data traffic is paramount.
A comparative analysis of the three interconnection models of the FANS standard reveals that each possesses distinctive strengths, rendering them suitable for disparate network sharing scenarios. Q-in-Q represents an efficient and straightforward solution for Metro Ethernet networks, offering an appropriate degree of isolation and scalability within existing Ethernet infrastructures. In contrast, VXLAN is the optimal selection for scenarios that necessitate a substantial number of virtual networks, such as data centers and networks supporting virtual machines. This is due to its capacity to construct an overlay on top of IP infrastructures. Finally, MPLS is the optimal solution for networks where quality of service and traffic prioritisation are paramount, offering enhanced flexibility in managing diverse traffic types.
Furthermore, the implementation of these models in shared networks offers not only technological improvements but also economic benefits. The sharing of physical infrastructure allows operators to reduce the costs associated with the construction and maintenance of separate networks. Additionally, FANS enables the development of novel customer services, thereby fostering more dynamic competition at the active layer of networks. In conclusion, the selection of a model is contingent upon the specific requirements of the network environment and the priorities of operators with regard to scalability, cost and complexity of deployment. This emphasises the necessity of adapting the infrastructure to the particular needs of each operator.
|
5G |
Fifth Generation of Wireless Cellular Technology |
|
ATM |
Asynchronous Transfer Mode |
|
BUM |
Broadcast Unknown Unicast and Multicast |
|
C-VLANs |
Client VLANs |
|
CoS |
Class of Service |
|
ECMP |
Equal-Cost Multi-Path |
|
ECN |
Explicit Congestion Notification |
|
FANS |
Fixed Access Network Sharing |
|
FTTH |
Fiber to the Home |
|
GPON |
Gigabit Passive Optical Networks |
|
ID |
Identification |
|
IDs |
Identifications |
|
IEEE |
Institute of Electrical and Electronics Engineers |
|
IETF |
Internet Engineering Task Force |
|
InP |
Infrastructure Provider |
|
InPs |
Infrastructure Providers |
|
IoT |
Internet of Things |
|
IP |
Internet Protocol |
|
ISPs |
Internet Service Providers |
|
ITU-R |
International Telecommunication Union - Radiocommunication Sector |
|
ITU-T |
International Telecommunication Union - Telecommunication Standardization Sector |
|
L2 |
Layer 2 |
|
L3 |
Layer 3 |
|
LAN |
Local Area Network |
|
LLC |
Logical Link Control |
|
LSP |
Label Switched Path |
|
MAC |
Media Access Control |
|
MAC-in-UDP |
Media Access Control in User Datagram Protocol |
|
mMTC |
Massive Machine-Type Communication |
|
MPLS |
Multiprotocol Label Switching |
|
NaaS |
Network as a Service |
|
O-VLAN |
Operator VLAN |
|
OLT |
Optical Line Termination |
|
ONT |
Optical Network Terminal |
|
ONTs |
Optical Network Terminals |
|
OSI |
Open Systems Interconnection |
|
pAN |
physical Access Node |
|
PE |
Provider Edge Routers |
|
PON |
Passive Optical Networks |
|
Q-in-Q |
802.1Q tunnelling |
|
QoS |
Quality of Service |
|
RFC |
Request for Comments |
|
S-VLANs |
Service VLANs |
|
SAMAC |
Source MAC Address |
|
SLAs |
Service Level Agreement |
|
SONET |
Synchronous Optical Network |
|
STP |
Spanning Tree Protocol |
|
TR-370 |
Fixed Access Network Sharing - Architecture and Nodal Requirements |
|
TTL |
Time To Live |
|
UDP |
User Datagram Protocol |
|
vAN |
virtual Access Node |
|
VLAN |
Virtual Local Area Network |
|
VLANs |
Virtual Local Area Networks |
|
VM |
Virtual Machine |
|
VMs |
Virtual Machines |
|
VMware |
Virtual Machine and software |
|
VNI |
VXLAN Network Identifier |
|
VNO |
Virtual Network Operator |
|
VNOs |
Virtual Network Operators |
|
VPN |
Virtual Private Network |
|
VTEP |
VXLAN Tunnel End Point |
|
VTEPs |
VXLAN Tunnel End Points |
|
VXLAN |
Virtual Extensible Local Area Network |
[1] FCGA. (2015). FTTH Council Global Alliance – FCGA, FTTH Council - Definition of terms. https://data.nag.wiki/AlphaMile/Certificates/FTTH_Definition_of_Terms-Revision_2015-Final.pdf.
[2] Fiber Optic Association. (2016). Fiber optic networks. https://www.thefoa.org/Lennie/nets.html.
[3] Broadband Forum. (2020). Fixed access network sharing - Architecture and nodal requirements (FANS). Technical Report TR-370i2. https://www.broadband-forum.org/pdfs/tr-370-2-0-0.pdf.
[4] ITU-T. (2008). Gigabit-capable passive optical networks (GPON): General characteristics. Recommendation ITU-T G.984.1. https://www.itu.int/rec/T-REC-G.984.1-200803-I/en.
[5] ITU-R. (2015). IMT vision – Framework and overall objectives of the future development of IMT for 2020 and beyond. ITU-R Recommendation M.2083-0. Electronic Publication, Geneva. https://www.itu.int/dms_pubrec/itu-r/rec/m/r-rec-m.2083-0-201509-i!!pdf-e.pdf.
[6] IPLAN (Internet Service Provider – Large Enterprise Data Center). https://www.iplan.com.ar/Acerca-de-iplan.
[7] IEEE Standards Association. (2006). IEEE 802.1ad-2005. https://standards.ieee.org/standard/802_1Q-2011.html.
[8] Juniper Networks. (2024). Configuring Q-in-Q tunneling and VLAN Q-in-Q tunneling and VLAN translation. In Ethernet Switching User Guide. Juniper Networks, Inc., Sunnyvale, California, p. 910. https://www.juniper.net/documentation/us/en/software/junos/multicast-l2/topics/topic-map/q-in-q.html.
[9] Cisco. (2022). Configuring Q-in-Q and layer 2 protocol tunneling. In Cisco Catalyst IR8340 Rugged Series Router Software Configuration Guide, Cisco IOS XE Release 17.8.x. Cisco Systems, Inc., San Jose, CA, p. 161. https://www.cisco.com/c/en/us/td/docs/routers/ir8340/software/configuration/b_ir8340_cg_17-8/m-configuring-q-in-q-and-layer-2-protocol-tunneling.html.
[10] Netberg. Q-in-Q configuration tips. https://netbergtw.com/top-support/articles/q-in-q-configuration-tips/.
[11] Huawei Enterprise. (2021). Overview of QinQ. https://support.huawei.com/enterprise/en/doc/EDOC1100137875/ded44dec/overview-of-qinq.
[12] Juniper Networks. (2024). Understanding VXLANs. https://www.juniper.net/documentation/us/en/software/junos/evpn/topics/topic-map/sdn-vxlan.html.
[13] Mahalingam, M., Dutt, D., Duda, K., Agarwal, P., Kreeger, L., Sridhar, T., Bursell, M., Wright, C. (2014). Virtual eXtensible Local Area Network (VXLAN): A framework for overlaying virtualized layer 2 networks over layer 3 networks. RFC 7348. https://doi.org/10.17487/rfc7348
[14] Arista Networks. Data Center Interconnection with VXLAN. http://allvpc.net/Arista_Design_Guide_DCI_with_VXLAN.pdf.
[15] Juniper Networks. (2020). What is VXLAN? https://www.juniper.net/us/en/research-topics/what-is-vxlan.html.
[16] Huawei. (2024). What is overlay network? https://info.support.huawei.com/info-finder/encyclopedia/en/Overlay%20network.html#content1.
[17] Sptel – Reliability with an Edge. (2021). What is MPLS (MultiProtocol Label Switching) for business? https://sptel.com/mpls-advantages-and-disadvantages/.
[18] Dube, C. (2021). Project 01. VXLAN use case study. University of New Hampshire, Connectivity Research Center. https://connectivity.unh.edu/research-projects/next-generation-data-communication-protocols/vxlan-use-case-study.html.
[19] Rosen, E.C., Viswanathan, A., Callon, R. (2001). Multiprotocol label switching architecture. RFC Editor: RFC 3031. https://doi.org/10.17487/RFC3031
[20] De Ghein, L. (2016). MPLS Fundamentals. Cisco Press, Hoboken, NJ.
[21] Goldman, J.E., Rawles, P.T. (2004). Applied Data Communications: A Business-Oriented Approach. Hoboken, NJ: Wiley.
[22] Hunter, P. (2024). MPLS — What is it? In plain English, please…. In SCTE TechExpo 24, Atlanta, Georgia. https://broadbandlibrary.com/mpls-what-is-it-in-plain-english-please/.
[23] GL Communications Inc. (2024). Multi-protocol label switching (MPLS) testing features. https://www.gl.com/telecom-test-solutions/ethernet-networks-devices-switches/multi-protocol-label-switching-mpls-packetexpert.html.
