1、 ATIS-0300104 ATIS Standard on - NEXT GENERATION INTERCONNECTION INTEROPERABILITY FORUM (NGIIF) NGN REFERENCE DOCUMENT NGN BASICS As a leading technology and solutions development organization, the Alliance for Telecommunications Industry Solutions (ATIS) brings together the top global ICT companies
2、 to advance the industrys most pressing business priorities. ATIS nearly 200 member companies are currently working to address the All-IP transition, network functions virtualization, big data analytics, cloud services, device solutions, emergency services, M2M, cyber security, network evolution, qu
3、ality of service, billing support, operations, and much more. These priorities follow a fast-track development lifecycle from design and innovation through standards, specifications, requirements, business use cases, software toolkits, open source solutions, and interoperability testing. ATIS is acc
4、redited by the American National Standards Institute (ANSI). The organization is the North American Organizational Partner for the 3rd Generation Partnership Project (3GPP), a founding Partner of the oneM2M global initiative, a member of and major U.S. contributor to the International Telecommunicat
5、ion Union (ITU), as well as a member of the Inter-American Telecommunication Commission (CITEL). For more information, visit www.atis.org. Notice of Disclaimer May 17, 20133J-STD-110, Joint ATIS/TIA Native SMS to 9-1-1 Requirements and Architecture Specification, March 2013.4J-STD-110.a, Supplement
6、A to J-STD-110, Joint ATIS/TIA Native SMS to 9-1-1 Requirements and Architecture Specification, November 201351This document is available from the Alliance for Telecommunications Industry Solutions (ATIS), 1200 G Street N.W., Suite 500, Washington, DC 20005. 2This document is available from the Inte
7、rnational Telecommunications Union. 3This document is available from the Federal Communications Commission at 4This document is available from the Alliance for Telecommunications Industry Solutions (ATIS) at . ATIS-0300104 6 J-STD-110.1, Joint ATIS/TIA Implementation Guidelines for J-STD-110, J-STD-
8、110, Joint ATIS/TIA Native SMS to 9-1-1 Requirements and Architecture Specification, November 20136Ad Hoc National SMS Text-to-9-1-1 Service Coordination Group (SCG), Interim SMS Text-to-9-1-1 Information Planning Guide, Version 2, May 201473. Definitions, Acronyms, the nature of the specific mechan
9、isms used depends on the networks underlying technology and transmission protocols. For communications paths that span multiple networks, the interoperation of one networks QoS control mechanisms with the control mechanisms of the other network is facilitated via standardized mappings. Some common Q
10、oS mechanisms include: Differentiated Services (DiffServ) Code Point (DSCP) QoS Class Identifier (QCI) Allocation and Retention Priority (ARP) Guaranteed Bit-Rate (GBR) Multi-Protocol Label Switching (MPLS) And others. Quality of Experience (QoE) The Quality of Experience (QoE) is the collective eff
11、ect of a services performance. It determines the degree of satisfaction of a user of the service. End-to-End End-to-End refers to the flow of communications between two end points (e.g., end-user devices) and specifically implies taking into account the end points and all intervening elements. End-t
12、o-Middle End-to-Middle, also known as end-to-edge, refers to connections that take into account a single end-point, an element at the edge of a providers network (often the far edge), and all intervening elements. Middle-to-Middle Middle-to-Middle refers to portions of communications networks that a
13、re either contained within one providers transit network, or span transit networks. Single providers middle-to-middle service is confined to the boundaries of a single providers network, edge to edge. Multiple providers middle-to-middle service spans multiple providers networks, usually from one pro
14、viders near edge to another providers far edge, and includes the effects of interoperability between the service providers networks. Throughput Throughput can be characterized in a variety of ways depending on context. Normally Throughput is defined in the contexts of Internet Protocol (IP) or User
15、Datagram Protocol (UDP) throughput, and Transmission Control Protocol (TCP) throughput. ATIS-0300104 14 IP Throughput focuses on the transmission of packets between two adjacent IP nodes (one or more lower layer devices may exist between the two IP nodes but they are transparent in the determination
16、 of throughput between the two IP nodes). Factors that determine IP Throughput include transmission delays between the nodes and processing delay within the nodes. TCP Throughput is defined as the average rate of successful delivery of data from a source IP node to a destination IP node over a TCP n
17、etwork. Factors that influence IP Throughput (described above) contribute to TCP Throughput. In addition, TCP Throughput is also affected by the constraints of this protocol8. As a result, throughput of TCP traffic will always be lower than the available raw (IP, or UDP) bandwidth. 6.3.2 Performance
18、 Measures Performance measures may apply to a network, or to a specific service provided by a network. 6.3.2.1 Network Performance Measures Network performance measures apply to the performance of a network without regard to any specific service. These may include the following: 1. Throughput 2. Dat
19、a errors (packet loss) 3. Latency (one-way or round-trip delay) 4. Jitter (variability of latency across multiple packets) 5. And others. These network performance measures are described in more detail below: Throughput (“throughput” here refers specifically to IP Throughput) Short-term (burst) thro
20、ughput: Short-term (burst) throughput is the average throughput that a user will receive during the initial, transient period of a transfer. One of the benefits of short-term (burst) throughput is its ability to characterize the throughput of a link on small data transfers. Sustainable throughput: L
21、ong-term (Sustainable) throughput is the steady-state expected throughput that a user should receive after an initial, transient period of higher throughput. Sustainable throughput needs to be considered for both upload and download, to capture asymmetry in data transfer rates. Minimum throughput: M
22、inimum throughput captures degradation in throughput that might result from congestion, throttling, or high network utilization. Bulk Transfer Capacity (BTC): Bulk Transfer Capacity (BTC) represents the achievable throughput by a single connection. BTC applies only to TCP connections and depends on
23、how TCP shares bandwidth among individual TCP flows. BTC is unique among throughput measurements in that it does not directly measure actual throughput, but rather the maximum throughput obtainable. Goodput: Goodput is the effective transmission rate of the payload data while discounting overhead bi
24、ts, control messages, re-transmissions, etc. 8As an illustration, the maximum size of the receive buffer of the destination IP node is limited by TCP to 65,536 bytes. Using a slow satellite path of 0.5 seconds as an example, a single TCP connection can then use only a maximum of 1.05 Mbit per second
25、 (65,536 Bytes / 0.5 seconds), regardless of the links capacity. ATIS-0300104 15 TCP Goodput measures the number of TCP payload bytes per second that the system can transfer. Throughput measurements often suffer from high variability due to differences in access technologies, providers traffic shapi
26、ng policies, and congestion during peak hours. Additionally, the actual throughput of a TCP connection is very difficult to measure due to complicating factors such as transfer size, types of cross traffic, and the number of competing connections. As a result, BTC may be the most useful of the throu
27、ghput metrics. Packet Loss Loss Rate: Loss Rate is the average packet loss over a period of time. Loss Burst Length: Loss Burst Length is the average duration of a packet loss episode. Packet loss measurement is valuable in that it has implications for TCP throughput and TCP timeouts. For instance,
28、studies have shown that the loss of even a single packet can severely degrade streaming video quality, and that bursts of loss have similarly deleterious effects on performance. Latency Round Trip Time (RTT): Round Trip Time (RTT) is the time delay between sending a packet and receiving a response.
29、RTT can be measured by sending small UDP packets to a test node and timing the responses (while treating any response receipt delay longer than a specified threshold as packet loss). Last-Mile Latency: Last-Mile Latency is the latency between an end-user device and the first device inside a service
30、providers network. Last-Mile Latency characterizes the access link and is therefore more useful for describing short data transfers. Last-Mile Latency is a strong metric because service providers can guarantee a maximum Last-Mile Latency that users should expect even during congestion. The standard
31、deviation of Last-Mile Latency can also be used to estimate jitter in a given segment. Latency Under Load: Latency Under Load is the actual latency that a user experiences during an upload or download. Latency Under Load reflects issues with buffer bloat, active queuing, and traffic shaping. Latency
32、, much like throughput, can be difficult to measure due to complicating factors: The quality of access links, modem buffering, and cross-traffic in customer premise equipment all interfere with latency measurements. Jitter Maximum Jitter (i.e., maximum delay variability): Maximum Jitter is the large
33、st value of jitter that users should experience, and indicates whether the data transmission characteristics are satisfactory for a given application. Jitter and latency are closely related, as are packet loss and throughput. 6.3.2.2 Service Performance Measures Service Performance Measures are gene
34、rally specific to each service: The performance measures that apply to voice over IP (VoIP) service, for instance, are different from the performance measures that are applicable to electronic mail (e-mail). Some performance measures, however, are universal (or nearly so) across multiple services. T
35、hese generic service performance measures may include the following: 1. Uplink and downlink goodput (i.e., net throughput, discounting overhead transmissions) 2. Round trip latency per transaction 3. Data session setup time ATIS-0300104 16 4. Data session setup success rate 5. Percentage of failed t
36、ransaction 6. Data session drop rate 7. Mean Opinion Score (MOS)98. And others. The purposes of Service Performance Measures are to allow the service provider to measure the users quality of experience by monitoring their services in near real time, and for identifying and responding to degradations
37、 in performance. From a service-users perspective, Service Performance Measures answer questions such as What is the likelihood that a data session can be completed? Responsiveness, i.e., o How long does it take to connect to the server? o How long does it take to download/upload the desired informa
38、tion? o What is the latency or delay during each transaction or interchange? How often is the connection lost? Is the transmission consistent enough to support the desired usage? While it might be possible to answer such questions directly by using custom software on the users devices, such an appro
39、ach may be cumbersome to develop, deploy and maintain. Alternatively, service providers can collect network performance measurements within their networks and use these to assess the factors impacting user experience. It should be emphasized that the service performance as experienced by the user is
40、 an end-to-end measure, regardless of the fact that there may be multiple intervening network technologies and providers. Thus the ability is required to collect these measures across technologies and providers in a manner that will enable an end-to-end view of the service performance. Conversely, f
41、rom the service providers perspective, Service Performance Measures, i.e., the collection and analysis of service and network measurements enable the providers to answer questions such as: What level of service quality is the network delivering to users? Is the service functioning normally now? Are
42、there any geographic regions where the service is not functioning or functioning less than optimally? What percentage of the time has the service been functioning normally for the past hour? Day? Month? If/when the service was not functioning normally everywhere, which network element(s) or feature(
43、s) were experiencing and/or causing the problem? How often and for how long has the service been down over the last day? Week? Month? What fraction of valid service requests has been honored/denied over the last day? Week? Month? Is the service affected by the presence or operation of other services
44、 on the providers network? Note that Service Performance Measures can be used to provide quantitative measures of user Quality of Experience. Also note that they can be measured objectively, either directly or via lower (network) level measures. 6.4 NGN Functional Entities A functional entity compri
45、ses a specific set of functions at a given location. Groupings of functional entities are used to describe practical physical implementations. In the NGN, functional entities controlling policy, sessions, media, resources, service delivery, security, etc., may be distributed over the infrastructure,
46、 including both existing and new networks. When these functional entities are distributed, they communicate over open interfaces. As such, the identification of reference points is an important aspect of NGN. Gateways provide a 9MOS, a numerical value between 0 and 5, started out as a subjective ass
47、essment of a telephone connections voice quality based on human perception. The International Telecommunications Union Telecommunication Standardization Sector (ITU-T) has redefined MOS ITU-T G.107 so that it now refers to an objective quantitative score computed from measured data of packet loss-ra
48、te, delay, and jitter. ATIS-0300104 17 means of interworking between different networks of operators, both existing networks and NGN (NGN, PSTN, ISDN, Global System for Mobile Communications (GSM), and Code Division Multiple Access (CDMA). 6.5 NGN Functional Architecture The functional architecture
49、is a set of functional entities and the reference points between them used to describe the structure of an NGN, each providing a unique function. The relationships and connections between functions are identified in terms of reference points. ATIS-0300104 18 7. Abstracts of NGN Documentation The abstracts of each of the individual NGIIF NGN Reference Documents will be provided in this section upon publication of each document. The abstracts will also be found in the NGN Master Table of Contents. Note: This section will be p
