Broadband Subscriber Access Network Overview
Subscriber Access Network Overview
A subscriber access environment can include various components, including subscriber access technologies and authentication protocols.
The subscriber access technologies include:
Dynamic Host Configuration Protocol (DHCP) server
Local DHCP server
External DHCP server
Point-to-Point Protocol (PPP)
The subscriber authentication protocols include the RADIUS server.
Figure 1 shows an example of a basic subscriber access network.
This feature requires a license. To understand more about Subscriber Access Licensing, see, Subscriber Access Licensing Overview. Please refer to the Juniper Licensing Guide for general information about License Management. Please refer to the product Data Sheets at MX Series Routers for details, or contact your Juniper Account Team or Juniper Partner.
Multiservice Access Node Overview
A multiservice access node is a broader term that refers to a group of commonly used aggregation devices. These devices include digital subscriber line access multiplexers (DSLAMs) used in xDSL networks, optical line termination (OLT) for PON/FTTx networks, and Ethernet switches for Active Ethernet connections. Modern MSANs often support all of these connections, as well as providing connections for additional circuits such as plain old telephone service (referred to as POTS) or Digital Signal 1 (DS1 or T1).
The defining function of a multiservice access node is to aggregate traffic from multiple subscribers. At the physical level, the MSAN also converts traffic from the last mile technology (for example, ADSL) to Ethernet for delivery to subscribers.
You can broadly categorize MSANs into three types based on how they forward traffic in the network:
Layer–2 MSAN—This type of MSAN is essentially a Layer 2 switch (though typically not a fully functioning switch) with some relevant enhancements. These MSANs use Ethernet (or ATM) switching to forward traffic. The MSAN forwards all subscriber traffic upstream to an edge router that acts as the centralized control point and prevents direct subscriber-to-subscriber communication. Ethernet Link Aggregation (LAG) provides the resiliency in this type of network.
Layer 2 DSLAMs cannot interpret IGMP, so they cannot selectively replicate IPTV channels.
Layer–3 aware MSAN—This IP-aware MSAN can interpret and respond to IGMP requests by locally replicating a multicast stream and forwarding the stream to any subscriber requesting it. Layer 3 awareness is important when supporting IPTV traffic to perform channel changes (sometimes referred to as channel zaps). Static IP-aware MSANs always receive all multicast television channels. They do not have the ability to request that specific channels be forwarded to the DSLAM. Dynamic IP-aware DSLAMs, however, can inform the network to begin (or discontinue) sending individual channels to the DSLAM. Configuring IGMP proxy or IGMP snooping on the DSLAM accomplishes this function.
Layer–3 MSAN—These MSANs use IP routing functionality rather than Layer 2 technologies to forward traffic. The advantage of this forwarding method is the ability to support multiple upstream links going to different upstream routers and improving network resiliency. However, to accomplish this level of resiliency, you must assign a separate IP subnetwork to each MSAN, adding a level of complexity that can be more difficult to maintain or manage.
In choosing a MSAN type, refer to Figure 2:
Ethernet MSAN Aggregation Options
Each MSAN can connect directly to an edge router (broadband services router or video services router), or an intermediate device (for example, an Ethernet switch) can aggregate MSAN traffic before being sent to the services router. Table 1 lists the possible MSAN aggregation methods and under what conditions they are used.
Method |
When Used |
|---|---|
Direct connection |
Each MSAN connects directly to the broadband services router and optional video services router. |
Ethernet aggregation switch connection |
Each MSAN connects directly to an intermediate Ethernet switch. The switch, in turn, connects to the broadband services router or optional video services router. |
Ethernet ring aggregation connection |
Each MSAN connects to a ring topology of MSANs. The head-end MSAN (the device closest to the upstream edge router) connects to the broadband services router. |
You can use different aggregation methods in different portions of the network. You can also create multiple layers of traffic aggregation within the network. For example, an MSAN can connect to a central office terminal (COT), which, in turn, connects to an Ethernet aggregation switch, or you can create multiple levels of Ethernet aggregation switches prior to connecting to the edge router.
Direct Connection
In the direct connection method, each MSAN has a point-to-point connection to the broadband services router. If an intermediate central office exists, traffic from multiple MSANs can be combined onto a single connection using wave-division multiplexing (WDM). You can also connect the MSAN to a video services router. However, this connection method requires that you use a Layer 3 MSAN that has the ability to determine which link to use when forwarding traffic.
When using the direct connection method, keep the following in mind:
We recommend this approach when possible to simplify network management.
Because multiple MSANs are used to connect to the services router, and Layer 3 MSANs generally require a higher equipment cost, this method is rarely used in a multiedge subscriber management model.
Direct connection is typically used when most MSAN links are utilized less than 33 percent and there is little value in combining traffic from multiple MSANs.
Ethernet Aggregation Switch Connection
An Ethernet aggregation switch aggregates traffic from multiple downstream MSANs into a single connection to the services router (broadband services router or optional video services router).
When using the Ethernet aggregation switch connection method, keep the following in mind:
Ethernet aggregation is typically used when most MSAN links are utilized over 33 percent or to aggregate traffic from lower speed MSANs (for example, 1 Gbps) to a higher speed connection to the services router (for example, 10 Gbps).
You can use an MX Series router as an Ethernet aggregation switch. For information about configuring the MX Series router in Layer 2 scenarios, see the Ethernet Networking User Guide for MX Series Routers.
Ring Aggregation Connection
In a ring topology, the remote MSAN that connects to subscribers is called the remote terminal (RT). This device can be located in the outside plant (OSP) or in a remote central office (CO). Traffic traverses the ring until it reaches the central office terminal (COT) at the head-end of the ring. The COT then connects directly to the services router (broadband services router or video services router).
The RT and COT must support the same ring resiliency protocol.
You can use an MX Series router in an Ethernet ring aggregation topology. For information about configuring the MX Series router in Layer 2 scenarios, see the Ethernet Networking User Guide for MX Series Routers.
LDP Pseudowire Autosensing Overview
A pseudowire is a virtual link that is used to transport a Layer 2 service across an MPLS edge or access network. In a typical broadband edge or business edge network, one end of a pseudowire is terminated as a Layer 2 circuit on an access node, and the other end is terminated as a Layer 2 circuit on a service node that serves as either an aggregation node or an MPLS core network. Traditionally, both endpoints are provisioned manually through configuration. LDP pseudowire autosensing introduces a new provisioning model that allows pseudowire endpoints to be automatically provisioned and deprovisioned on service nodes based on LDP signaling messages. This model can facilitate the provisioning of pseudowires on a large scale. An access node uses LDP to signals both pseudowire identity and attributes to a service node. The identity is authenticated by a RADIUS server, and then used together with the attributes signaled by LDP and the attributes passed down by the RADIUS server to create the pseudowire endpoint configuration, including the Layer 2 circuit.
Pseudowire Ingress Termination Background
In a seamless MPLS-enabled broadband access or business edge network, Ethernet pseudowires are commonly used as virtual interfaces to connect access nodes to service nodes. Each pseudowire carries the bidirectional traffic of one or multiple broadband subscribers or business edge customers between an access node and a service node pair. The establishment of the pseudowire is usually initiated by the access node, based on either static configuration or dynamic detection of a new broadband subscriber or business edge customer arriving on a client-facing port on the access node.
Ideally, the access node should create one pseudowire per client port, where all subscribers or customers hosted by the port are mapped to the pseudowire. The alternative is where there is one pseudowire per client port (S-VLAN), and all subscribers or customers sharing a common S-VLAN on the port are mapped to the pseudowire. In either case, the pseudowire is signaled in the raw mode.
The S-VLAN, if not used to delimit service on the service node or combined with C-VLAN to distinguish subscribers or customers, will be stripped off before the traffic is encapsulated in pseudowire payload and transported to the service node. Individual subscribers or customers may be distinguished by C-VLAN, or a Layer 2 header such as DHCP and PPP, which will be carried in pseudowire payload to the service node. On the service node, the pseudowire is terminated. Individual subscribers or customers are then demultiplexed and modeled as broadband subscriber interfaces, business edge interfaces (for example, PPPoE), Ethernet interfaces, or IP interfaces. Ethernet and IP interfaces may be further attached to service instances, such as VPLS and Layer 3 VPN instances.
In Junos OS, pseudowire ingress termination on service nodes is supported through the use of pseudowire service physical and logical interfaces. This approach is considered as superior in scalability to the old logical tunnel interface based approach, due to its capability of multiplexing and demultiplexing subscribers or customers over a single pseudowire. For each pseudowire, a pseudowire service physical interface is created on a selected Packet Forwarding Engine, which is called an anchor Packet Forwarding Engine. On top of this pseudowire service physical interface, a ps.0 logical interface (transport logical interface) is created, and a Layer 2 circuit or Layer 2 VPN is created to host the ps.0 logical interface as an attachment interface.
The Layer 2 circuit or Layer 2 VPN enables pseudowire signaling towards the access node, and the ps.0 logical interface serves the role of customer edge facing interface for the pseudowire. Further, one or multiple ps.n logical interfaces (also known as service logical interfaces, where n>0) may be created on the pseudowire service physical interface to model individual subscriber/customer flows as logical interfaces. These interfaces can then be attached to desired broadband and business edge services or Layer 2 or Layer 3 VPN instances.
Note that the purpose of the anchor Packet Forwarding Engine is to designate the Packet Forwarding Engine to process the bidirectional traffic of the pseudowire, including encapsulation, decapsulation, VLAN mux or demux, QoS, policing, shaping, and many more.
For Junos OS Release 16.2 and earlier, the creation and deletion of the pseudowire service physical interfaces, pseudowire service logical interfaces, Layer 2 circuits, and Layer 2 VPNs for pseudowire ingress termination rely on static configuration. This is not considered as the best option from the perspective of scalability, efficiency, and flexibility, especially in a network where each service node may potentially host a large number of pseudowires. The objective is to help service providers come out of static configuration in provisioning and deprovisioning pseudowire ingress termination on service nodes.
Pseudowire Autosensing Approach
In the pseudowire autosensing approach, a service node uses the LDP label mapping message received from an access node as a trigger to dynamically generate configuration for a pseudowire service physical interface, a pseudowire service logical interface, a Layer 2 circuit. Likewise, it uses the LDP label withdraw message received from the access node and LDP session down event as triggers to remove the generated configuration. In pseudowire autosensing, it is assumed that access nodes are the initiators of pseudowire signaling, and service nodes are the targets. In a network where a service may be hosted by multiple service nodes for redundancy or load balancing, this also provides access nodes with a select-and-connect model for service establishment. The basic control flow of pseudowire autosensing is shown in Figure 3
The basic control flow procedure of pseudowire autosensing is as follows:
Customer premises equipment (CPE) comes online and sends an Ethernet frame with C-VLAN to the optical line terminator (OLT). OLT adds S-VLAN to the frame and sends the frame to the access node. The access node checks with the RADIUS server to authorize the VLANs.
The RADIUS server sends an access accept to the access node. The access node creates a Layer 2 circuit and signals a pseudowire to the service node through an LDP label mapping message.
The service node accepts the label mapping message, and sends an access request with pseudowire information to the RADIUS server for authorization and for selection of a pseudowire service physical interface or a logical interface.
The RADIUS server sends an access accept to the service node with a service string specifying the selected pseudowire service physical interface or logical interface. The service node creates a Layer 2 circuit configuration, the pseudowire information, and the pseudowire service physical interface or logical interface. The service node signals the pseudowire towards the access node through an LDP label mapping message. The pseudowire comes up bidirectionally.
Sample Configuration
The following configuration explicitly marks the Layer 2 circuit as generated by autosensing. The pseudowire service physical interface and pseudowire service logical interface configuration are optional, depending on whether they preexist.
Router 0
[edit]
protocols {
Layer 2 circuit {
neighbor 192.0.2.2 {
interface ps0.0 {
virtual-circuit-id 100;
control-word;
mtu 9100;
auto-sensed;
}
}
}
}
Layer 2 Services on Pseudowire Service Interface Overview
The pseudowire service logical interface supports the transport logical interface (psn.0) on the MPLS access side and service logical interfaces (psn.1 to psn.n) on the MPLS core side of the subscriber management network.
The pseudowire service on service logical interfaces psn.1 to psn.n are configured as Layer 2 interfaces in the bridge domain or in a virtual private LAN service (VPLS) instance. There is Layer 2 circuit or the Layer 2 VPN across MLPS access between an Ethernet aggregation device and a service edge device with the pseudowire service on transport logical interface psn.0 as the terminating interface of the Layer 2 circuit or the Layer 2 VPN at the service edge device.
Junos OS supports the pseudowire service on service logical interfaces psn.1 to psn.n in the bridge domain or VPLS instance, which receives traffic egressing from the pseudowire service on the transport logical interface at the service edge device. It also enables Layer 2 ingress features such as MAC learning, VLAN manipulations, and destination MAC look up on the pseudowire service on service logical interfaces.
When the traffic is in reverse direction, the destination MAC enters the Layer 2 domain at the service edge device, which is learned as the source MAC on the pseudowire service on service logical interfaces. Starting in Junos OS Release 17.1R1, the pseudowire logical tunnel interfaces support Ethernet VPLS, Ethernet bridge, VLAN VPLS, and VLAN bridge encapsulation next hops to exit Layer 2 traffic. Starting in Junos OS Release 18.4R1, the Layer 2 service support with the pseudowire service logical interfaces is extended to pseudowire service interfaces anchored over redundant logical tunnel interfaces as well. These Layer 2 services are supported only on pseudowire service on service logical interfaces (psn.1 to psn.n) and not on transport logical interface (psn.0). The Layer 2 output features such as VLAN manipulations and others are enabled on the pseudowire service interfaces. The traffic sent out of the interfaces enter the pseudowire service on transport logical interfaces which is the Layer 2 circuit interface between Ethernet aggregation and service edge devices across the MPLS access domain.
For Junos OS Release 16.2 and earlier, Layer 2 encapsulations or features could not be configured on pseudowire service on service logical interfaces.
- Traffic from Customer LAN to MPLS
- Traffic from Service Edge to Customer LAN
- Pseudowire Service Interfaces
- Sample Configuration
Traffic from Customer LAN to MPLS
VPLS-x and VPLS-y instances are configured on the MPLS core side of the service edge device (PE A). A Layer 2 circuit or Layer 2 VPN is configured between the Ethernet aggregation device (EAD 1) and the service edge device. ps0.0 (transport logical interface) is the local interface in the Layer 2 circuit or the Layer 2 VPN at PE A. Junos OS supports pseudowire service on service logical interface ps0.x (x>0) in VPLS instance VPLS-x (VLAN ID in VPLS-x = m) and pseudowire service on service logical interface ps0.y(y>0) in VPLS instance VPLS-y (VLAN ID in VPLS-y = n).
In Figure 4, when the traffic comes from EAD 1 to PE A (on either Layer 2 circuit or Layer 2 VPN) with any VLAN ID, the traffic will exit through ps0.0. Based on the VLAN ID in the traffic the pseudowire service on service logical interface is selected. For example, if VLAN ID is m, then the traffic will enter ps0.x and if VLAN ID is n, then the traffic will enter ps0.y.
When traffic enters pseudowire service on the service logical interface ps0.n, where n>0, the following steps are performed.
The source MAC learning should occur on the Layer 2 pseudowire service on the service logical interface. The source Packet Forwarding Engine for this MAC is the Packet Forwarding Engine of the logical tunnel interface on which the pseudowire service is anchored in a VPLS instance or bridge domain in the PE A device.
The destination MAC lookup is done at the entry side as an input bridge family feature list of pseudowire services on service logical interfaces.
If destination MAC lookup is successful, then the traffic is sent as unicast; otherwise, the destination MAC, broadcast MAC, and multicast MAC are flooded.
If destination MAC lookup fails for the traffic coming on a pseudowire service on a service logical interface, the
mlp querycommand is sent to the Routing Engine and the other Packet Forwarding Engine in bridge domain or VPLS instance.
If a new MAC is learned on a pseudowire service on a service logical interface, then the
mlp addcommand is sent to the Routing Engine and the other Packet Forwarding Engine in bridge domain or VPLS instance.
Traffic from Service Edge to Customer LAN
When traffic enters the VPLS instance or bridge domain at the service edge device and if the destination MAC in the traffic is learned on a pseudowire service on a service logical interface, then the token associated with that pseudowire service logical interface is set at the entry side. The traffic is then sent to the Packet Forwarding Engine on which the logical tunnel interface of the pseudowire service physical interface is anchored through a fabric. When this token is launched, it supports VLAN VPLS, VLAN bridge, Ethernet VPLS, and Ethernet bridge encapsulations. The encapsulation next hop points to the egress logical interface feature list of the pseudowire service on the service logical interface to execute all the Layer 2 output features and send the packet to the entry side of the pseudowire service on transport logical interface ps0.0.
If the MAC query reaches the Packet Forwarding Engine on which the pseudowire service is anchored, then the Packet Forwarding Engine sends the response only when the MAC learned on the pseudowire service on the service logical interface is present. The Layer 2 token associated with the pseudowire service on the service logical interface seen after destination MAC lookup for the MAC learned on the pseudowire service on service logical interface should point to the next hop associated with the access side of the pseudowire service on service the logical interface.
The pseudowire service on the transport logical interface is the local interface ps0.0 of the Layer 2 circuit or Layer 2 VPN between the service edge and the Ethernet aggregation devices. Traffic is sent to the Ethernet aggregation device though the Layer 2 circuit or Layer 2 VPN across the MPLS access domain.
If the destination MAC traffic coming from the entry and exit side of the service edge device is unknown or multicast or broadcast, the traffic needs to be flooded. This requires an customer edge device flood next hop to include the pseudowire service on service logical interface, which acts as an access logical interface for the VPLS instance or bridge domain.
Pseudowire Service Interfaces
The following features are supported on pseudowire service interfaces:
A pseudowire service interface is hosted over a logical tunnel interface (lt-x/y/z). The traffic from a transport pseudowire service on a logical interface to a subscriber pseudowire service on a logical interface is based on the available VLAN ID.
Transfer of traffic from a subscriber pseudowire service on a logical interface to a transport pseudowire service on a logical interface is based on the channelID through an available loopback IP address.
Pseudowire service on service logical interfaces are supported on the virtual routing and forwarding (VRF) routing instance.
-
Pseudowire subscriber (ps) service on a trunk interface to terminate Layer 2 circuit instance in a VPLS-enabled virtual switch. The same Layer 2 circuit can also be terminated in the VPLS instance-type routing instance with different service logical interfaces and Layer 3 VPN VRF instance-type routing instance using another service logical interface as well.
Sample Configuration
The following sample configurations show a pseudowire service on a transport logical interface on a Layer 2 circuit, a pseudowire service on service logical interfaces in a bridge domain and a VPLS instance in a service edge device, and a pseudowire service on a trunk service interface in a VPLS instance:
Pseudowire service on a service logical interface in bridge domain on router 0
[edit]
interfaces {
ps0 {
unit 0 {
encapsulation ethernet-ccc;
}
unit 1 {
encapsulation vlan-bridge;
vlan-id 1;
}
unit 2 {
encapsulation vlan-bridge;
vlan-id 2;
}
}
ge-0/0/0 {
unit 1 {
encapsulation vlan-bridge;
vlan-id 1;
}
unit 2 {
encapsulation vlan-bridge;
vlan-id 2;
}
}
ge-2/0/6 {
unit 0 {
family inet {
address 10.11.2.1/24;
}
family mpls;
}
}
}
protocols {
mpls {
label-switched-path to_192.0.2.2 {
to 192.0.2.2;
}
}
bgp {
group RR {
type internal;
local-address 192.0.3.3;
}
}
l2-circuit {
neighbor 192.0.2.2 {
interface ps0.0 {
virtual-circuit-id 100;
}
}
}
}
bridge-domains {
bd1 {
domain-type bridge;
vlan-id 1;
interface ps0.1;
interface ge-0/0/0.1;
}
bd2 {
domain-type bridge;
vlan-id 2;
interface ps0.2;
interface ge-0/0/0.2;
}
}
Pseudowire service on a service logical interface in a VPLS instance on router 0
[edit]
interfaces {
ps0 {
unit 0 {
encapsulation ethernet-ccc;
}
unit 1 {
encapsulation vlan-vpls;
vlan-id 1;
family vpls;
}
unit 2 {
encapsulation vlan-vpls;
vlan-id 2;
family vpls;
}
}
ge-0/0/0 {
unit 1 {
encapsulation vlan-vpls;
vlan-id 1;
family vpls;
}
unit 2 {
encapsulation vlan-vpls;
vlan-id 2;
family vpls;
}
}
ge-2/0/6 {
unit 0 {
family inet {
address 10.11.2.1/24;
}
family mpls;
}
}
}
protocols {
mpls {
label-switched-path to_192.0.2.2 {
to 192.0.2.2;
}
}
bgp {
group RR {
type internal;
local-address 192.0.3.3;
}
}
l2-circuit {
neighbor 192.0.2.2 {
interface ps0.0 {
virtual-circuit-id 100;
}
}
}
}
routing-instances {
vpls-1 {
instance-type vpls;
vlan-id 1;
interface ps0.1;
interface ge-0/0/0.1;
}
vpls-2 {
instance-type vpls;
vlan-id 2;
interface ps0.2;
interface ge-0/0/0.2;
}
}
Pseudowire service on a trunk service interface in a VPLS instance on router 0
[edit]
interfaces {
ps0 {
flexible-vlan-tagging;
encapsulation flexible-ethernet-services;
unit 0 {
encapsulation ethernet-ccc;
}
unit 1 {
family bridge {
interface-mode trunk;
vlan-id 1;
}
}
ge-0/0/0 {
unit 1 {
encapsulation vlan-bridge;
vlan-id 1;
family bridge;
}
}
}
routing-instances {
vpls-1 {
instance-type virtual-switch;
protocols {
vpls {
site PE3 {
interface ps0.1;
site-identifier 1;
}
}
}
bridge-domains {
bd1 {
vlan-id 1;
}
}
interface ps0.1;
route-distinguisher 65001:1;
vrf-target target:1:1;
}
}
Pseudowire service on a service logical interface in a Layer 2 circuit on router 0
[edit]
interfaces {
ps0 {
unit 0 {
encapsulation ethernet-ccc;
}
unit 1 {
encapsulation vlan-ccc;
vlan-id 1;
}
unit 2 {
encapsulation vlan-ccc;
vlan-id 2;
}
}
ge-0/0/0 {
unit 1 {
encapsulation vlan-vpls;
vlan-id 1;
family vpls;
}
unit 2 {
encapsulation vlan-vpls;
vlan-id 2;
family vpls;
}
}
ge-2/0/6 {
unit 0 {
family inet {
address 10.11.2.1/24;
}
family mpls;
}
}
}
protocols {
mpls {
label-switched-path to_192.0.2.2 {
to 192.0.2.2;
}
}
bgp {
group RR {
type internal;
local-address 192.0.3.3;
}
}
l2-circuit {
neighbor 192.0.2.2 {
interface ps0.0 {
virtual-circuit-id 100;
}
}
neighbor 10.10.10.10 {
interface ps0.1 {
virtual-circuit-id 1;
}
}
neighbor 10.11.11.11 {
interface ps0.2 {
virtual-circuit-id 2;
}
}
}
}
Broadband Access Service Delivery Options
Four primary delivery options exist today for delivering broadband network service. These options include the following:
Digital Subscriber Line
Digital subscriber line (DSL) is the most widely deployed broadband technology worldwide. This delivery option uses existing telephone lines to send broadband information on a different frequency than is used for the existing voice service. Many generations of DSL are used for residential service, including Very High Speed Digital Subscriber Line 2 (VDSL2) and versions of Asymmetric Digital Subscriber Line (ADSL, ADSL2, and ADSL2+). These variations of DSL primarily offer asymmetric residential broadband service where different upstream and downstream speeds are implemented. (VDSL2 also supports symmetric operation.) Other DSL variations, like High bit rate Digital Subscriber Line (HDSL) and Symmetric Digital Subscriber Line (SDSL), provide symmetric speeds and are typically used in business applications.
The head-end to a DSL system is the Digital Subscriber Line Access Multiplexer (DSLAM). The demarcation device at the customer premise is a DSL modem. DSL service models are defined by the Broadband Forum (formerly called the DSL Forum).
Active Ethernet
Active Ethernet uses traditional Ethernet technology to deliver broadband service across a fiber-optic network. Active Ethernet does not provide a separate channel for existing voice service, so VoIP (or TDM-to-VoIP) equipment is required. In addition, sending full-speed (10 or 100 Mbps) Ethernet requires significant power, necessitating distribution to Ethernet switches and optical repeaters located in cabinets outside of the central office. Due to these restrictions, early Active Ethernet deployments typically appear in densely populated areas.
Passive Optical Networking
Passive Optical Networking (PON), like Active Ethernet, uses fiber-optic cable to deliver services to the premises. This delivery option provides higher speeds than DSL but lower speeds than Active Ethernet. Though PON provides higher speed to each subscriber, it requires a higher investment in cable and connectivity.
A key advantage of PON is that it does not require any powered equipment outside of the central office. Each fiber leaving the central office is split using a non-powered optical splitter. The split fiber then follows a point-to-point connection to each subscriber.
PON technologies fall into three general categories:
ATM PON (APON), Broadband PON (BPON), and Gigabit-capable PON (GPON)—PON standards that use the following different delivery options:
APON—The first passive optical network standard is primarily used for business applications.
BPON—Based on APON, BPON adds wave division multiplexing (WDM), dynamic and higher upstream bandwidth allocation, and a standard management interface to enable mixed-vendor networks.
GPON—GPON is based on BPON but supports higher rates, enhanced security, and a choice of which Layer 2 protocol to use (ATM, Generic Equipment Model [GEM], or Ethernet).
Ethernet PON (EPON)—Provides capabilities similar to GPON, BPON, and APON, but uses Ethernet standards. These standards are defined by the IEEE. Gigabit Ethernet PON (GEPON) is the highest speed version.
Wave Division Multiplexing PON (WDM-PON)—A nonstandard PON which, as the name implies, provides a separate wavelength to each subscriber.
The head-end to a PON system is an Optical Line Terminator (OLT). The demarcation device at the customer premises is an Optical Network Terminator (ONT). The ONT provides subscriber-side ports for connecting Ethernet (RJ-45), telephone wires (RJ-11) or coaxial cable (F-connector).
Hybrid Fiber Coaxial
Multi-System Operators (MSOs; also known as cable TV operators) offer broadband service through their hybrid fiber-coaxial (HFC) network. The HFC network combines optical fiber and coaxial cable to deliver service directly to the customer. Services leave the central office (CO) using a fiber-optic cable. The service is then converted outside of the CO to a coaxial cable tree using a series of optical nodes and, where necessary, through a trunk radio frequency (RF) amplifier. The coaxial cables then connect to multiple subscribers. The demarcation device is a cable modem or set-top box, which talks to a Cable Modem Termination System (CMTS) at the MSO head-end or primary facility that receives television signals for processing and distribution. Broadband traffic is carried using the Data Over Cable Service Interface Specification (DOCSIS) standard defined by CableLabs and many contributing companies.
Broadband Delivery and FTTx
Many implementations use existing copper cabling to deliver signal to the premises, but fiber-optic cable connectivity is making its way closer to the subscriber. Most networks use a combination of both copper and fiber-optic cabling. The term fiber to the x (FTTx) describes how far into the network fiber-optic cabling runs before a switch to copper cabling takes place. Both PON and Active Ethernet can use fiber-optic portion of the network, while xDSL is typically used on the copper portion. This means that a single fiber-optic strand may support multiple copper-based subscribers.
Increasing the use of fiber in the network increases cost but it also increases network access speed to each subscriber.
The following terms are used to describe the termination point of fiber-optic cable in a network:
Fiber to the Premises (FTTP), Fiber to the Home (FTTH), Fiber to the Business (FTTB)—Fiber extends all the way to the subscriber. PON is most common for residential access, although Active Ethernet can be efficiently used in dense areas such as apartment complexes. Active Ethernet is more common for delivering services to businesses.
Fiber to the Curb (FTTC)—Fiber extends most of the way (typically, 500 feet/150 meters or less) to the subscriber. Existing copper is used for the remaining distance to the subscriber.
Fiber to the Node/Neighborhood (FTTN)—Fiber extends to within a few thousand feet of the subscriber and converted to xDSL for the remaining distance to the subscriber.
Fiber to the Exchange (FTTE)—A typical central office-based xDSL implementation in which fiber is used to deliver traffic to the central office and xDSL is used on the existing local loop.
Understanding BNG Support for Cascading DSLAM Deployments Over Bonded DSL Channels
Junos OS supports configuring and maintaining the access lines between access nodes and their ANCP subscribers using DSL access multiplexer as the broadband access technology for Copper-to-the-Building (CuTTB) and Fiber-to-the-Building (FTTB). When multiple subscribers share the same access line, the access line could be one of the following types:
PON, Fiber-to-the-Building (FTTB)
Bonded DSL Copper-To-The-Building (CTTB)
Starting in Junos OS Release 18.2R1, Passive Optical Network (PON) access technologies are supported with four levels of quality-of-service (QoS) scheduler hierarchy for residential subscribers in a BBE deployment. This feature extends the Access Node Control Protocol (ANCP) implementation to handle network configuration for residential customers that use PON as the broadband access technology for both CuTTB and FTTB. ANCP uses a statically controlled traffic-control profile on the interface-set for shaping at the subscriber level at the intermediate node to which the subscribers are connected. New DSL types are provided to support access line rate adjustment for the new access technologies.
A new RADIUS VSA, Inner-Tag-Protocol-Id 26-211 is
introduced to fetch the inner VLAN Tag Protocol Identifier value for
L2BSA subscribers to enable maintaining one dynamic profile instead
of two separate dynamic profiles. A new Junos OS dynamic profile variable $junos-inner-vlan-tag-protocol-id allows a VLAN map’s inner-tag-protocol-id to be set by RADIUS or a predefined default
value provided in the configuration.
- Benefits of Cascading DSLAM Deployments Over Bonded DSL Channels
- 4-Level Scheduler Hierarchy
- Use Cases of Cascading DSLAM Deployments Over Bonded DSL Channels
- Bonded DSL for Copper-To-The-Building (CuTTB)
- Hybrid PON + G.fast
- Supported Features
Benefits of Cascading DSLAM Deployments Over Bonded DSL Channels
This feature is useful to support access network deployments where multiple subscribers share the same access line aggegrated by an intermediate node between the access node and the home routing gateways. Another benefit is to conserve Layer 2 CoS nodes. Typically a dummy Layer 2 node is created for each residential household, which could exhaust Layer 2 CoS resources. Therefore, network models using bonded DSL, G.Fast, and PON access models can conserve Layer 2 CoS nodes.
4-Level Scheduler Hierarchy
Junos OS supports 4-Level QoS scheduler hierarchy minimally supporting residential and L2BSA access over Copper-to-the-Building (CTTB) or Fiber-to-the-Building access network deployments. The following QoS scheduler hierarchy levels are supported:
Level 1 Port (Physical interface or AE)
Level 2 Access Line (Logical interface set, represents a collection of subscribers sharing a given access line aggregated by an intermediate node)
Level 3 Subscriber sessions
Level 4 Queues (services)
In Figure 5, residential and L2BSA access require only 4-level scheduler hierarchy. Business subscriber access is currently not supported and hence 4-level scheduler hierarchy is sufficient for CuTTB and PON services targeted to an apartment building.
Use Cases of Cascading DSLAM Deployments Over Bonded DSL Channels
Bonded DSL for copper to the building (CuTTB) introduces an intermediate node Distribution Point Unit-Copper (DPU-C) between the DSL access multiplexer (DSLAM) and a cluster of subscribers at the customer location. Shared access line deployment models may be of type Passive-Optical-Network (PON) or bonded DSL copper lines. Example intermediate nodes are listed below:
DPU-C - bonded DSL for Copper-To-The-Building (CTTB)
ONU - PON (Fiber-to-the-Building (FTTB)
Hybrid PON and G.Fast
Bonded DSL for Copper-To-The-Building (CuTTB)
In Figure 6, each DPU-C has an ANCP session to report access line parameters of individual subscribers connected to the node. The MSAN also has an ANCP session to report access line parameters of the bonded DSL access line to the DPU-C. All subscribers connected to the DPU-C are thus subject to the DSL access-line downstream rate, the DPU-C subscribers are grouped together in an interface set. You can adjust the speeds reported in this Port-Up and apply to the CoS node for the corresponding interface ste maintaining the semantics of the CoS adjustment control profile that is used for individual subscriber lines. The access model consists of a hybrid of bonded DSL access and conventional unbonded access. The DPU-C and the Multi Service Access Node (MSAN) ANCP sessions are completely independent and the PPPoE-IA tags only reflect the attributes reported in the dPU-C ANCP session
Hybrid PON + G.fast
In Figure 7, the OLT has an ANCP session with the BNG and proxies for all downstream native PON nodes. G.fast DSL subscribers are connected to an intermediate node, which has a PON connection to the intermediate ONU in front of the OLT.
A hybrid access network connects DSL based subscriber lines using both PON access and G.fast nodes with an intermediate node between the OLT and home gateways (HGs). Both businesses and residences are connected to the intermediate node, which is the PON leaf. Shaping is required both at the subscriber level and at the PON leaf level. Ths G.fast subscribers are associated with the intermediate ONU like a native PON subscriber. New DSL type TLVs are supported by the AN and their values are reported in the ANCP Port-Up for the correstpnding subscriber access line. However, it is still not possible to distinguish between an intermediate node and a conventional connection for a given PPPoE session.
Supported Features
Support ANCP-based traffic shaping on dynamic iflsets.
Preservation of PPP0E-IA and ANCP independence by CLI configuration for residential subscribers.
New Juniper VSA, ERX-Inner-Vlan-Tag-Protocol-Id (4874-26-211) is supported to source the inner VLAN Tag Protocol Identifier value for L2BSA subscribers as an optimization to maintain two, separate dynamic profiles, one for TPID - 0x88a8 and one for 0x8100, and sourcing the desired value by returning 26-4874-174 (Client-rofile-Name) in the Access-Accept.
The following additional type values for the DSL type TLV are supported. All subscribers include these DSL type TLVs in the PPPoE PADR messages’s PPPoE IA tags.
(8) G.fast
(9) VDSL2 Annex Q
(10) SDSL bonded
(11) VDSL2 bonded
(12) G,fast bonded
(13) VDSL2 Annex Q bonded
Detection of Backhaul Line Identifiers and Autogeneration of Intermediate Node Interface Sets
Before you begin, you must confirm that your
existing access nodes or IAs are not already inserting strings that
begin with the # character. Because this is a system-level
configuration, parsing applies to all ANCP access nodes and PPPoE
IAs globally. The leading # character is not configurable.
Parsing is disabled by default in case some providers use that character
for some other purpose.
Starting in Junos OS
Release 18.4R1, you can configure the router to detect a logical intermediate
node in an access network. The node identifies
subscribers that are connected to the same shared media, such as a
PON tree or a bonded copper line that connects to a DPU-C for CuTTB.
When you configure this detection, the router parses the ANCP Access-Aggregation-Circuit-ID-ASCII
attribute (TLV 0x03) that is received either in the ANCP Port Up message
or PPPoE PADR IA tags. If the TLV string begins with the # character, the string is a backhaul line identifier that is unique
across the network to identify the bonded DSL line or the PON tree.
The same string is reported in the TLV or IA for all subscribers connected
to that DPU-C or PON.
The portion of the string after the # character represents
the logical intermediate node. It is used as the name of the dynamic
interface set for the CoS Level 2 node that groups the subscribers
using that intermediate node. This interface set is known as the parent
interface set. Every PPPoE or VLAN (L2BSA) logical interface with
the same value for TLV 0x03 is a member of that interface set.
The TLV value must match the requirements for interface set naming; it can include alphanumeric characters and the following special characters:
# % / = + - : ; @ . _
This portion of the string also sets the value of the $junos-aggregation-interface-set-name predefined variable in the dynamic profile. This value is used as the name of a CoS Level 2 interface set that groups the subscribers sharing that string. It overrides the predefined variable default, which uses the value of $junos-phy-ifd-interface-set-name as the name of the interface set.
For example, if the value of the TLV string is #TEST-DPU-C-100, the value of the predefined variable—and consequently the name of the interface set—becomes TEST-DPU-C-100.
The Access-Loop-Remote-ID (TLV (0x02) is similarly parsed
for the # character, but the resulting string is not used
in the current release.
Intermediate node detection is supported only for 4-level scheduler hierarchies, so business access is limited to conventional DSL access MPCs.
To enable parsing of the Access-Aggregation-Circuit-ID-ASCII TLV and setting the interface set name:
The following sample configuration shows a dynamic profile for L2BSA subscribers. Three things to note here are the following:
A default value of $junos-phy-ifd-interface-set-name is defined for the $junos-aggregation-interface-set-name predefined variable.
The name of the interface set is configured to be the value of $junos-aggregation-interface-set-name.
The CoS scheduler configuration specifies an interface named with the value of $junos-aggregation-interface-set-name.
When hierarchical-access-network-detection is configured
for the access lines, then the name of the Level 2 scheduler interface
set is determined as follows:
When TLV 0x03 begins with
#, then $junos-aggregation-interface-set-name is the remainder of the string, excluding the initial#.When TLV 0x03 begins with any other character, then $junos-aggregation-interface-set-name is the value of $junos-phy-ifd-interface-set-name.
[edit dynamic-profiles L2BSA-subscriber]
predefined-variable-defaults {
aggregation-interface-set-name phy-ifd-interface-set-name;
cos-shaping-rate 1g;
cos-scheduler-map schedmap_L2BSA;
inner-vlan-tag-protocol-id 0x88a8;
}
routing-instances {
"$junos-routing-instance" {
interface "$junos-interface-name";
}
}
interfaces {
interface-set $junos-aggregation-interface-set-name {
interface "$junos-interface-ifd-name" {
unit "$junos-interface-unit";
}
}
"$junos-interface-ifd-name" {
unit "$junos-interface-unit" {
encapsulation vlan-vpls;
no-traps;
vlan-id "$junos-vlan-id";
input-vlan-map {
swap-push;
inner-tag-protocol-id "$junos-inner-vlan-tag-protocol-id"
vlan-id "$junos-vlan-map-id";
inner-vlan-id "$junos-inner-vlan-map-id";
}
output-vlan-map {
pop-swap;
inner-tag-protocol-id 0x8100;
}
family vpls;
}
}
}
class-of-service {
traffic-control-profiles {
L2BSAShaper {
scheduler-map "$junos-cos-scheduler-map";
shaping-rate "$junos-cos-shaping-rate" burst-size 17k;
overhead-accounting frame-mode cell-mode-bytes 6;
}
L2iflsetShaper {
shaping-rate 1G burst-size 17k;
}
}
interfaces {
"$junos-interface-ifd-name" {
unit "$junos-interface-unit" {
output-traffic-control-profile L2BSAShaper;
classifiers {
ieee-802.1 L2BSA vlan-tag outer;
}
rewrite-rules {
ieee-802.1 L2BSA vlan-tag outer;
}
}
}
interface-set "$junos-aggregation-interface-set-name" {
output-traffic-control-profile L2iflsetShaper;
}
}
}