1、 TECHNICAL REPORT ATIS-1000056 ACCESS NETWORKS ARCHITECTURE TECHNICAL REPORT As a leading technology and solutions development organization, ATIS brings together the top global ICT companies to advance the industrys most-pressing business priorities. Through ATIS committees and forums, nearly 200 co
2、mpanies address cloud services, device solutions, emergency services, M2M communications, cyber security, ehealth, network evolution, quality of service, billing support, operations, and more. These priorities follow a fast-track development lifecycle from design and innovation through solutions tha
3、t include standards, specifications, requirements, business use cases, software toolkits, and interoperability testing. ATIS is accredited by the American National Standards Institute (ANSI). ATIS is the North American Organizational Partner for the 3rd Generation Partnership Project (3GPP), a found
4、ing Partner of oneM2M, a member and major U.S. contributor to the International Telecommunication Union (ITU) Radio and Telecommunications sectors, and a member of the Inter-American Telecommunication Commission (CITEL). For more information, visit . Notice of Disclaimer voice uses the lower frequen
5、cies on the loop and data uses the higher frequencies on the loop. In addition, the CPE Access Gateway includes the filter/splitter installed on the end users premises to prevent interaction between the high-frequency DSL tones and the end users telephone set. 5.1.1.1.3 Network Interface Device (NID
6、) The NID is a CPE device that performs interface functions, such as code conversion, protocol conversion, and buffering, required for communications to and from a network. The NID provides communication between the CPE Access Gateway and the Access Node. The NID may be a standalone device or may be
7、 integrated with the CPE Access Gateway. 5.1.1.1.4 Main Distribution Frame (MDF) The MDF is the first point of connection inside the Access Node, providing physical connectivity between the access loop and the DSLAM. 5.1.1.1.5 Digital Subscriber Line Access Multiplexer (DSLAM) The DSLAM terminates a
8、nd multiplexes end user input using IP and performs the corresponding demultiplexing for downstream traffic. DSL termination and multiplexing functions may be integrated or may be performed in separate equipment. Each subscriber line is connected to a modem in the DSLAM, where the signaling and bear
9、er to and from the corresponding end user undergoes protocol conversion as needed. The DSLAM also separates the end users voice traffic, which is delivered to the PSTN switch, from the data traffic, which is provided as part of an aggregated stream to the Ethernet Aggregation function. Similarly, fo
10、r downstream traffic, the DSLAM performs the necessary protocol conversion and passes the voice traffic to the end user in the low frequency range and the data traffic in the high frequency range. DSLAM management may occur through communication with a Policy Server/Policy Decision Point (PDP) and/o
11、r Policy Enforcement Point (PEP). 5.1.1.1.6 Ethernet Aggregation The Ethernet Aggregation function collects the Ethernet input from multiple DSLAMs for delivery to the Broadband Network Gateway (BNG). 5.1.1.1.7 Broadband Network Gateway (BNG) The BNG is an IP edge router at the border of a Regional
12、Broadband Network. At the BNG, the admission control and the QoS policies of the network are applied. The BNG is also the aggregation point for outbound subscriber traffic. BNG functionality includes subscriber management, advanced IP processing (including IP QoS), and enhanced traffic management ca
13、pabilities. A PEP in the BNG is responsible for traffic policy enforcement of all traffic types in the BNG. Policy enforcement may be applied at an IP session, IP flow, and/or aggregate level. 5.1.1.1.8 Regional Broadband Network The Regional Broadband Network interconnects a Core Network and an Acc
14、ess Node. Typically more than one Access Node is connected to a common Regional Broadband Network. Similarly, multiple Core Networks may be ATIS-1000056 9 connected to a common Regional Broadband Network. Because of the commonality of requirements for a BNG in a Regional Broadband Network and in a C
15、ore Network, this TR does not distinguish between these two network types. 5.1.1.1.9 Policy Server/Policy Decision Point (PDP) The Policy Server/PDP is a functional entity making decisions on subscriber policies on an IP Session, IP Flow, and an aggregate basis. For this TR, there are four possible
16、reference models for the Policy Server/PDP: The International Telecommunication Union Telecommunication Standardization Sector (ITU-T) Resource and Admission Control Function (RACF) architecture Y.2111. The 3GPP Policy and Charging Control (PCC) architecture. The European Telecommunications Standard
17、s Institute (ETSI) Telecommunications and Internet converged Services and Protocols for Advanced Networking (TISPAN) Resource and Admission Control Subsystem (RACS). The Broadband Forum Fixed-Mobile Convergence (FMC) model. Functionally, the ITU-T RACF and 3GPP PCC architectures are similar; the mai
18、n difference is in the consolidation versus distribution of the interfaces and the protocols used. The Broadband Forum FMC interworking architecture uses the 3GPP S9 interface, and is a natural extension of PCC. Many of the existing wireline access components communicate via COPS, not Diameter, maki
19、ng the ITU-T RACF architecture more appropriate for a DSL Access Network. For example, the interface from the Policy Server/PDP to the Application Function for a DSL Service Provider is expected to be the 3GPP Rx interface or the ITU-T Rs interface. In addition, the ITU-T Rw interface from the Polic
20、y Server/PDP to the BNG is assumed, with the modification that an ITU-T Resource and Admission Control Function (RACF) rather than a Radius Authentication, Authorization, and Accounting (AAA) server is used. The Policy Server/PDP is independent of the Core Network and can be used with non-IMS FEs; h
21、owever, this section assumes the Policy Server/PDP is located at the border of the DSL Access Network and the Core Network. A Policy Server/PDP may manage multiple Policy Enforcement Points (PEPs) (in their separate BNGs), coordinating what policies should be enforced at the BNG. The PEP may impleme
22、nt static policies which it obtains from the Policy Server/PDP. These may apply to all IP flows at the BNG or may apply only to a given IP flow. For further details, refer to TR-058 and TR-059 . 5.1.2 Call/Session Flows For call/session flows associated with a DSL Access Network, refer to ATIS-10000
23、49. 5.1.3 Interfaces encryption is used to prevent eavesdropping. Upstream signals are combined using either Time Division Multiple Access (TDMA) or Wavelength Division Multiplexing (WDM). The OLT “ranges” each ONT in order to provide the upstream time slot assignments, since the transmission delay
24、from each ONT can be unique. The OLT measures delay and sets a register in each ONT to equalize delay among the ONTs. The OLT transmits a grant, a permission to use a defined interval of time for upstream transmission, to each ONT. The OLT maintains a grant map, which is dynamically recalculated eve
25、ry few milliseconds. 5.2.1.2 ONT The Optical Network Termination (ONT) is the Customer Premise Equipment (CPE) device used in conjunction with the OLT to provide broadband data transport service over the PON. 5.2.1.3 Access Node The combination of ONT and OLT is sometimes viewed (e.g., for managemen
26、t purposes) as a single system, referred to as an Access Node (AN). The Access Node Management System manages that abstract entity by physically connecting to the OLT. The OLT and ONT communicate, for management purposes, via the ONT Management and Control Interface (OMCI). 5.2.1.4 ODN The Optical D
27、istribution Network (ODN) provides physical connectivity between the ONT and the OLT. It is made up of wavelength division multiplexers8, fiber optic cable, and unpowered optical splitters. An OLT will generally connect many ODNs, the set of which is referred to as the Optical Access Network (OAN).
28、5.2.1.5 Access Node Management System The Access Node Management System manages the Access Nodes (OLT and ONT) via aQ3 interface as defined in Q.812. 5.2.2 Call/Session Flows For call/session flows associated with a Fiber Access Network, refer to ATIS-1000049. 5.2.3 Interfaces & Protocols This secti
29、on is informative and paraphrases information that is described in more detail in the relevant standards. It is provided as background information for readers of this TR. 8The PON uses different wavelengths in the upstream and downstream direction. ATIS-1000056 17 5.2.3.1 Diameter - Based Interfaces
30、 The ITU-T RACF architecture may be more appropriate for the interface from the ANMS to the Application Function. However, the 3GPP Rx interface is assumed to be consistent with IMS Core Network requirements. This Diameter-based interface is described in an earlier section. 5.2.3.2 Q3 Interface The
31、Q3 interface is defined in Q.812. The protocol stack for the Q3 interface for the interactive class of service is shown in Figure 6. In this figure, the Systems Management Application Service Element (SMASE) communicates with lower layer protocols through either the Association Control Service Eleme
32、nt (ACSE) protocol or through the Common Management Information Service Element (CMISE)/Remote Operations Service Element (ROSE) protocols. Figure 6 - Q3 and OMCI Interface Protocol Stack 5.2.3.3 ONT Management & Control Interface (OMCI) In GPON Encapsulation Method (GEM) Mode, each ONT management a
33、nd control protocol packet is encapsulated directly in a GEM packet. The packet format is: GEM header (5 bytes) Contains the PortID of the OMCC for the addressed ONT. Transaction correlation identifier (2 bytes) For request messages, the OLT selects any transaction identifier. A response message car
34、ries the transaction identifier of the message to which it is responding. Message type (1 byte). Device identifier (1 byte). ATIS-1000056 18 Message identifier (4 bytes). Message contents (32 bytes). OMCI trailer (8 bytes). In order to handle messages associated with higher priority requests, a two
35、level priority mechanism exists within the ONT management and control protocol. Figure 7 shows the entities within the ONT to process the two priority levels of the protocol. Figure 7 - Protocol Entities within the ONT 5.2.3.4 ITU-T G.984 Series GPON QoS Mechanisms 5.2.3.4.1 Physical Layer GPON supp
36、orts up to 64 ONTs on the fiber from a single OLT interface. 5.2.3.4.2 MAC Layer Data is mapped onto the GPON physical layer through the GPON Transmission Convergence (GTC) layer. The downstream GTC frame is 125 microseconds long, and consists of the following fields: Physical Control Block (PCB): o
37、 Frame Alignment Pattern. o Identity. o Physical Layer OAM message (PLOAM). o Frame Error Check. o Downstream Payload Length Field. o Upstream Bandwidth Map Field. Native ATM Cell Payload (may be null). GPON Encapsulation Method (GEM) Payload. ATIS-1000056 19 The Upstream Bandwidth Map Field specifi
38、es the start and stop times of the Allocation IDs in the bursts within the 125 microsecond upstream frame window. The start and stop times are typically specified in units of bytes from the beginning of the frame. The upstream GTC frame consists of the following fields: Preamble, Delimiter and Frame
39、 Check transmitted by an ONT at the beginning of its upstream transmission burst. ONT Identifier. Upstream Indicator for ONT status reporting. PLOAM message, if requested by the OLT. Dynamic Bandwidth Report (DBR) for a specific Allocation ID, if requested by the OLT. Payload for that specific Alloc
40、ation ID. DBR and Payload for additional Allocation IDs. Transmission Containers (T-CONTs) are defined in ITU-T G.983.4. There are five T-CONT types defined: T-CONT Type I is characterized by fixed bandwidth only. It has the highest priority in being satisfied when assignable bandwidth becomes avail
41、able and is used for delay sensitive traffic. T-CONT Type II is characterized by assured bandwidth only. Assured bandwidth is defined as a fixed average bandwidth over some specified time interval. Type 2 only guarantees the average transmission rate, while Type 1 guarantees packet transfer delay an
42、d delay variation. Type 2 has the second highest priority in being satisfied. T-CONT Type III has assured bandwidth and non-assured bandwidth. T-CONT Type 3 is allocated bandwidth equivalent to its assured bandwidth only when it has a traffic queue equivalent to or greater than its assured bandwidth
43、. Non-assured bandwidth is allocated across all T-CONTs with assured bandwidth that are requesting additional bandwidth in proportion to the assured bandwidth of the individual T-CONTs e.g., weighted round robin method. The sum of the assured and non-assured bandwidth allocated to a T-CONT will not
44、exceed its maximum bandwidth, which is a provisioned value. Type 3 has the third highest priority in being satisfied. T-CONT Type IV has best-effort bandwidth only. Best effort bandwidth is allocated to each T-CONT Type 4 equally e.g., based on the round robin method, up to the maximum bandwidth. Ty
45、pe 4 has the lowest priority in being satisfied. T-CONT Type V is a superset of all T-CONT types, and can be downgraded to one or more of the other T-CONT types. T-CONTs are identified by an Allocation ID. The OLT makes bandwidth assignments to an ONT on a per Allocation ID basis, where an ONT can h
46、ave one or more Allocation IDs. The OLT controls the bandwidth and QoS of each Allocation ID by the number of upstream time slots assigned to the T-CONT. Fixed and assured bandwidths are known as guaranteed bandwidth. Non-assured and best effort bandwidths are known as additional bandwidth. Surplus
47、bandwidth is defined as the additional bandwidth minus bandwidth allocated for control, such as OAM. The surplus bandwidth is available for dynamic bandwidth assignment (DBA). Two different DBA mechanisms are defined in G.983.4: In the non-status reporting (NSR) strategy, the OLT monitors the amount
48、 of bandwidth each ONT is using, based on the number of idle frames it receives in the upstream GTC frames. More bandwidth is assigned to an ONT if its bandwidth utilization exceeds a predefined threshold. In the status reporting (SR) strategy, a quick indication of a need for more bandwidth in a T-
49、CONT type is communicated in the Upstream Indicator field. A more detailed report of the per T-CONT buffer status is communicated in the upstream dynamic bandwidth reports. The OLT uses the status report information to determine the appropriate bandwidth allocation for each Allocation ID. An OLT can also use a hybrid approach for DBA. ATIS-1000056 20 5.2.4 Policy Control of the GPON There is today no direct policy control of or policy enforcement within the GPON. For this TR, the following policy control process
