1、48b2591 0561415 9 INTERNATIONAL TELECOMMUN CATION UN ION CCITT THE INTERNATIONAL TELEGRAPH AND TELEPHONE CONS U LTATIV E COM M ITTE E INTEGRATED SERVICES DIGITAL NETWORK (ISDN) OVERALL NETVVORK ASPECTS AND FUNCTIONS, IS DN US ER-N ETWORK INTER FACES 1.370 CONGESTION MANAGEMENT FOR THE ISDN FRAME REL
2、AYING BEARER SERVICE Recommendation 1.370 Geneva, 1991 4862593 O563436 O INTERNATIONAL TELECOMMUNICATIO CCITT TH E INTERNATIONAL TELEGRAPH AND TELEPHONE CON SU LTATIVE C OMM ITTE E INTEGRATED SERVICES DIGITAL NETWORK (ISDN) OVERALL NETWORK ASPECTS AND FUNCTIONS, ISDN USER-NETWORK INTERFACES C 1.370
3、CONGESTION MANAGEMENT FOR THE ISDN FRAME RELAYING BEARER SERVICE Recommendation 1.370 I Geneva, 1991 FOREWORD The CCITT (the International Telegraph and Telephone Consultative Committee) is a permanent organ of the International Telecommunication Union (ITU). CCI is responsible for studying technica
4、l, operating and tariff questions and issuing Recommendations on them with a view to standardizing telecommunications on a worldwide basis. The Plenary Assembly of CCITT which meets every four years, establishes the topics for study and approves Recommendations prepared by its Study Groups. The appr
5、oval of Recommendations by the members of CC between Plenary Assemblies is covered by the procedure laid down in CCIT Resolution No. 2 (Melbourne, 1988). Recommendation 1.370 was prepared by Study Group XWI and was approved under the Resolution No. 2 procedure on the 25th of October 1991. CC NOTES 1
6、) telecommunication Administration and a recognized private operating agency. 2) In this Recommendation, the expression “Administration” is used for conciseness to indicate both a A list of abbreviations used in this Recommendation can be found in Annex A. O IT 1991 All rights reserved. No part of t
7、his publication may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying and microfilm, without permission in writing from the ITU. 4862593 0563438 4 R Recommendation 1.370 CONGESTION MANAGEMENT FOR THE ISDN FRAME RELAYING BEARER SERVICE 1 Congestio
8、n management principles 1.1 Scope This Recommendation describes U-plane based congestion management strategy and mechanisms for ISDN frame relay bearer services. It covers both network and end user mechanisms and responsibilities to avoid or recover from periods of congestion. C-plane procedures at
9、the user network interface other than clearing or not accepting calls are not recommended. Procedures, objectives and requirements for C-plane congestion management between networks are for further study. Special provisions for the treatment of continuous bit stream oriented (CBO) traffk is outside
10、the scope of the Recommendation. This strategy is intended to operate for access channel rates up to 2048 kbit/s. 1.2 Definitions access rate The data rate of the user access channel , B or H). The speed of the access channel determines how much data (maximum rate) the end user can inject into the n
11、etwork. committed burst size, Bc The maximum committed amount of data a user may offer to the network during a time interval Tc. Bc is negotiated at call Set-up. excess burst size, Be The maximum ailowed amount of data by which a user can exceed Bc during a time interval Tc. This data (Be) is delive
12、red in general with a lower probability than Bc. Be is negotiated at cal Set-up. committed rate measurement interval, Tc The time interval during which the user is ailowed to send only the committed amount of data (Bc) and the excess amount of data (Be). Tc is computed. committed information rate (C
13、IR) The information transfer rate which the network is committed to transfer under normal conditions. The rate is averaged over a minimum increment of time Tc. CIR is negotiated at call Set-up. forward explicit congestion notification (FECN) See Recommendation Q.922 for the full definition. Recommen
14、dation 1,370 1 4Bb2591 0561419 b = backward explicit congestion notification (BECN) See Recommendation Q.922 for the full definition. consolidated link layer management message (CLLM) See Recommendation Q.922 for the full definition. discard eligibiity indicator This indicates that a frame should be
15、 discarded in preference to other frames in a congestion situation, when frames must be discarded to ensure safe network operation and maintain the committed level of service within the network. fairness An attempt by the network to maintain the committed call parameters which the end user negotiate
16、d at call setup time. An example of this would be first discarding the frames in excess of the committed information rate (CIR) and refusing to allow new call setups to occur prior to discarding committed data traffic. offered load Refers to the frames offered to the network, by an end user, to be d
17、elivered to the selected destination. The information rate offered to the network could exceed the negotiated class of service parameters. congestion management Includes network engineering, OAM procedures to detect the onset of congestion, and real time mechanisms to prevent or recover from congest
18、ion. Congestion management includes but is not limited to congestion control, congestion avoidance and congestion recovery as defined below. congestion control Refers to real-time mechanisms to prevent and recover from congestion during periods of coincidental peak trafk demands or network overload
19、(e.g., resource failures). Congestion control includes both congestion avoidance and congestion recovery mechanisms. congestion avoidance Congestion avoidance procedures refer to procedures initiated at or prior to point A (see Figures 1D.370 and 2D.370) to prevent congestion from progressing to poi
20、nt B. Congestion avoidance procedures operate around point A and within the regions of mild congestion and severe congestion as shown in Figures 1D.370 and 2D.370. congestion recovery Congestion recovery refers to procedures initiated to prevent congestion from severely degrading the end user percei
21、ved quality of service(s) delivered by the network. These procedures are typically initiated when the network has begun to discard frames due to congestion. Congestion recovery procedures operate around point B and within the region of severe congestion as shown in Figures 1D.370 and 2D.370. ingress
22、 node The node that supports the source user-network interface (UNI). 2 Recommendation 1.370 egress node The node that supports the destination user-network interface (UNI). 1.3 Objectives of congestion management The primary objectives of congestion control mechanisms are to maintain, with a very h
23、igh probability, specified quality of service (e.g. throughput, delay, frame loss) for each virtual call or permanent virtual circuit. Congestion in the U-plane of a frame relay bearer service occurs when traffk, arriving at a resource (e.g. memory, bandwidth, processor), exceeds the network enginee
24、red level. It can also occur for other reasons (e.g. equipment failure). The impact of network congestion is performance degradation in terms of throughput and delay. Two levels of congestion in terms of impact on class of service, are defined. Point A is the point beyond which the transit delay in
25、the frame relay network increases at a rate faster than the rate at which offered load is increased. This is due to the network entry into a mild congestion state. This point is the final point on the curve that the network can guarantee the negotiated class of service. A further increase in offered
26、 load may cause a degradation in class of service. Point B is the point where the network begins discarding frames to control the existing level of congestion and prevent additional damage to the network provided services. Point A and Point B are dynamic values determined by the instantaneous condit
27、ion of the network resources. The end user may perceive the movement from point A to point B without increasing the offered load (e.g, due to resource failure or reconfiguration within the network). Threshold values are determined relative to the U-plane Quahty of Service objectives to the end user.
28、 Specific networks may define different values, reflecting different performance objectives (e.g. for the support of different grades of services), even within the same network. Congestion avoidance mechanisms aim to: - minimize frame discard; - maintain, with high probability and minimal variance,
29、an agreed Quality of Service; - minimize the possibility that one end user can monopolize network resources at the expense of other end users; - be simple to implement, and place little overhead on either the end user or the network: - create minimal additional network traffic; - distribute network
30、resources fairly among end users; - limit spread of congestion to other networks and elements within thenetwork; - operate effectively regardless of the traffic flow in either direction between end users; - have minimum interaction or impact on other systems in the frame relaying network; and - mini
31、mize the variance in Quality of Service delivered to individual virtual circuits during congestion. (e.g . individual virtual circuits should not experience sudden degradation when congestion approaches or has occurred). Congestion recovery mechanisms (in addition to the above) aim to ensure recover
32、y of the network from a severely congested state. Recommendation 1.370 3 4862591 056L42L 4 1.4 Requirements of congestion control mechanisms Congestion management mechanisms should have the following characteristics: - Be part of the U-plane. Explicit congestion notification (ECN) shall be provided
33、for in the U-plane. Note this applies to real-time notification aspects of congestion control and assumes that management functions such as gathering of statistics on congestion (i.e. when, where, why) could be accomplished outside the U-plane. - Ensure vansport of explicit congestion notification a
34、cross frame relaying networks. The network(s) shail convey the backward ECN (BECN) towards the source end user and the forward ECN (FECN) towards the destination end user. This requires that these indications (if set) shail not be reset as they traverse the network(s) towards source and destination
35、users. - From a service perspective, call setup negotiations (e.g. throughput) are rate based. This means that from the standpoint of the service provided by the network in a frame relay environment, the rate at which information is offered to the network may be expressed as a number of information
36、units per unit of time and is fundamental to all types of traffic to be carried. - Reaction by the end user to the receipt of explicit congestion notification (FECNBECN) is rate based and may be subject to standardization. It is noted that window mechanisms in terminais approximate rate based mechan
37、isms and may be used to control the rate at which traffic is offered to a network. - Networks should utilize, and users should react to, explicit congestion notification (i.e. not mandatory but highly desirable). - Data sources which are unable to respond to explicit congestion notification (i.e. CL
38、LM), can only be controlled by metering and discard. - The network which perceives congestion should have the option to generate congestion notification using the appropriate congestion control protocols. When ECN is generated, it shall be sent in the appropriate direction(. The policies for sending
39、 ECN will be different for the source control and destination control mechanisms. - The end users (e.g., private networks), may generate ECNs. 1.5 Congestion management strategy Distributed real-time congestion controls are necessary to prevent, and recover from, congestion during infrequent periods
40、 of coincidental peak traffic demands. Congestion avoidance action is the joint responsibility of both the network and the end user and requires coordination between them. Avoidance procedures seek to return network operation to Region I of Figures 1A.370 and 2A.370. Congestion avoidance with explic
41、it signalling and congestion recovery with implicit signalling are considered to be effective and complementary forms of congestion control in the frame relaying bearer service. 4 Recommendation 1.370 4862593 0563422 6 = . Congestion avoidance and recovery schemes are distributed in that traffic mon
42、itoring (e.g. by buffer usage) is most efficient and accurate at congested resources, while traffic rate control is most effective when carried out by an end user. For end users to know when to decreasehmease their traffic rates, there must be a standardized notification mechanism between the networ
43、k and end user. Joint responsibility and procedures between the end user and the network should be verifiable by the network. Congestion recovery initiation is the responsibility of the network. The end user should assist the network by continuing the avoidance procedures. Congestion recovery is use
44、d to help control offered load on severely congested networks and move from region III to region I of Figures 1D.370 and 2D.370. No congestion Mild Severe within network I congestion t congestion Region I Region II Region 111 4 b c 3 Q JZ ul 3 e s s 2 a Z Offered load T1812680-91 Note -The different
45、 lines in the severe congestion region reflect the fact that networks react and degrade differently in the face of severe congestion. FIGURE 1fl.370 Throughput and network congestion Recommendation 1.370 5 4862593 05bL423 8 No congestion Mild Severe within network . congestion congestion Region I Re
46、gion 111 b Offered load T18 12690-9 1 Note -The different lines in the severe congestion region reflect the fact that networks react and degrade differently in the face of severe congestion. FIGURE 2b.370 Delay and network congestion 1.5.1 Congestion control mechanisms 1.5.1.1 Eqlicit congestion not
47、ijcaton Commonly used end-to-end protocols operate with either source controlled or destination controlled transmit mechanisms for which two optional explicit congestion notification mechanisms for the frame relay bearer service are provided. These mechanisms, when implemented, are independent, not
48、mutually exclusive, and may be used concurrently. - Mechanism I: For destination controlled transmitters, the FECN is set in the core aspects protocol. - Mechanism 2: For source controlled transmitters, the BECN is set in the core aspects protocol in frames transported in the reverse direction (Le.
49、toward the transmitter). Alternatively, a consolidated link layer management message may be generated. This provides reverse notification for one or more data link connection identifiers (DLCIs) within a single frame. The CLLM is sent on the layer management DLCI in the U-plane in the backward direction .e. toward the source end user). The CLLM and the BECN may be used together or separately to notify the end user. 6 Recommendation 1.370 4862593 0561424 T 1.5.1.2 Discard eligibility The use of the discard eligibility indicator by the users and the network is optional. Thi
