1、释放流程及异常释放原因分析,2016-05 张恒,概述,主要内容说明,本次主要讨论的内容： LTE业务释放类型及相应信令流程的详解；相关的counter触发定义，以及异常问题分析；通过对这些内容的梳理和讨论，能够了解LTE释放的理论知识，并且在日常KPI和Log分析方面，提供帮助。,释放流程及异常释放原因分析,主要内容,业务承载类型及业务建立过程 业务释放类型及相应流程、原因 E-RAB release UE context release RRC connection release 异常释放的触发情景详解 UE initiated drop eNB initiated drop 异常释放r
2、elated Counter Triggers Network/Field KPI及Drops分析 Summary,业务承载类型,LTE业务承载类型,LTE的业务承载类型分为默认承载、专用承载两种： 默认承载：UE开机时获得承载，一定是MME发起的业务，优先级特别低，Non GBR； 专用承载：一定是由PGW触发建立的； 一个UE可以有不同的专用承载；,GTP variant: EPS bearer termination in PGW,SRB与默认、专用承载关系,业务承载类型,无线承载分类,SRB Signalling Radio Bearers（SRBs）的定义是，仅用于传输RRC和NAS
3、消息的RB（Radio Bearer）；DRB Data Radio Bearer，用于传输业务数据的无线承载；与S1-U承载共同组成E-RAB.分为UM、AM两种；,业务承载类型,UE context及E-RAB,E-RAB UE至SGW之间的承载，E-RAB是指用户平面的承载，由S1-U承载和DRB（Data Radio Bearer，数据无线承载）串联而成，进入LTE系统的业务数据主要通过E-RAB进行传输。因此对于LTE，业务管理主要是在E-RAB层次上进行的。 一个UE可以有多个E-RAB，目前最多支持8个EPS承载（个AM DRB，3个UM DRB）. 在应用上，E-RAB常出现于
4、性能指标的统计中，例如：E-RAB建立成功率。E-RAB建立成功指eNodeB成功为UE分配了通信通道的无线资源和用户面的无线承载，所以这个指标反映了eNodeB或小区接纳业务的能力，可以用来考虑系统负荷情况。,业务承载类型,UE context及E-RAB,UE context Including E-RAB context, security context, roaming restriction, UE S1 signalling connection ID(s), etc. in the eNB.在源eNB的UE上下文中还包含了与漫游和接入限制有关的信息； Initial UE co
5、ntext只能由MME发起，一个UE只有一个UE context； 一个eNB UE context是eNB内和一个处于激活状态的UE相关联的一个信息块。这块信息是为了维持E-UTRAN提供给这个激活态UE的服务所必须的信息； eNB UE context是在UE迁移到激活状态之后开始建立的；或者，eNB UE context是在切换目标eNB完成切换资源分配（切换准备阶段）之后开始建立的。 RRC connection 层三信令连接，UE保持非IDLE的必须连接，一个UE只有一个RRC connection，一旦连接释放，UE所有业务释放，并进入RRC_IDLE态；,LTE完整业务流程,默认
6、承载建立信令流程（Attach）,Steps: Random Access Initial NAS msg NAS security S1 init setup RRC sec + UE cap SRB2/DRB setup NAS reply,LTE完整业务流程,RA及RRC连接建立流程,SRB0： RRC connection request、RRC connection setup（消息中包含SRB1建立的相关信息）； SRB1： RRC connection setup complete；,LTE完整业务流程,专用承载建立流程,Dedicated Bearer Establishmen
7、t： E-RAB建立,释放流程及异常释放原因分析,主要内容,业务承载类型及业务建立过程 业务释放类型及相应流程、原因 E-RAB release UE context release RRC connection release Abnormal release的触发情景详解 UE initiated drop eNB initiated drop 异常释放related Counter Triggers Network/Field KPI及Drops分析 Summary,业务释放类型,release,根据业务承载类型，业务释放过程也分为三类： E-RAB release The purpos
8、e of the E-RAB Release procedure is to enable the release of already established E-RABs for a given UE； E-RAB释放后，UE仍然可以处于ECM-CONNECTED状态； UE context release If the eNB wants to remove all remaining E-RABs, e.g., for user inactivity, the UE Context Release Request procedure shall be used instead. 当UE
9、回到ECM-IDLE状态时，MME向eNodeB发UE CONTEXT RELEASE消息。收到这条消息后，eNodeB删除eNB UE context。MME继续保留MME UE context.,业务释放类型,release,根据业务承载类型，业务释放过程也分为两类： RRC connection release The purpose of this procedure is to release the RRC connection, which includes the release of the established radio bearers as well as all r
10、adio resources； RRC连接释放后，UE进入RRC_IDLE状态。,E-RAB release,释放类型：E-RAB release,根据3GPP定义(3GPP TS 36.413)，E-RAB release过程分为两大类： ENodeB Initiated MME Initiated E-RAB管理的相关信息都embed在context管理信令中。,MME Initiated,ENodeB Initiated,释放流程信令详解,E-RAB release ：MME Initiated,E-RAB release command消息内容： shall contain the i
11、nformation required by the eNB to release at least one E-RAB in the E-RAB To Be Released List IE. If a NAS-PDU IE is contained in the message, the eNB shall pass it to the UE. E-RAB To Be Released List IE. the UE Aggregate Maximum Bit Rate IE. E-RAB release response消息内容： - A list of E-RABs which are
12、 released successfully shall be included in the E-RAB Release List IE. - A list of E-RABs which failed to be released, if any, shall be included in the E-RAB Failed to Release List IE.,释放流程信令详解,E-RAB release ：MME Initiated,eNB收到E-RAB RELEASE COMMAND后，eNB将会释放被请求的ERABs. 对于每一个要释放的E-RAB，eNB会释放其相应的DRB和空口
13、资源。 The eNB shall pass the value contained in the E-RAB ID IE received for the E-RAB to the radio interface protocol for each Data Radio Bearer to be released. The eNB shall release allocated resources on S1 for the E-RABs requested to be released.,Network triggered (from Delete Bearer),释放流程信令详解,E-R
14、AB release ：eNB Initiated,E-RAB RELEASE INDICATION消息内容： The E-RAB RELEASE INDICATION message shall contain at least one E-RAB released at the eNB, in the E-RAB Released List IE.Upon reception of the E-RAB RELEASE INDICATION message the MME shall normally initiate the appropriate release procedure on
15、 the core network side for the E-RABs identified in the E-RAB RELEASE INDICATION message.Interaction with UE Context Release Request procedure: If the eNB wants to remove all remaining E-RABs, e.g., for user inactivity, the UE Context Release Request procedure shall be used instead.,释放流程及异常释放原因分析,主要
16、内容,业务承载类型及业务建立过程 业务释放类型及相应流程、原因 E-RAB release UE context release RRC connection release Abnormal release的触发情景详解 UE initiated drop eNB initiated drop 异常释放related Counter Triggers Network/Field KPI及Drops分析 Summary,UE context release,释放类型：UE context release,根据3GPP定义(3GPP TS 36.413)，UE context过程同样分为两大类：
17、 ENodeB Initiated MME Initiated This procedure is used to release the logical S1-AP signalling connection (over S1-MME) and all S1 bearers (in S1-U) for a UE. The procedure will move the UE from ECM-CONNECTED to ECM-IDLE in both the UE and MME, and all UE related context information is deleted in th
18、e eNode B.,UE context release,释放的触发条件UE Context Release Request,该信令，使基站可以根据E-UTRAN的需要，向MME请求UE-associated S1链接的释放；ENodeB发起的Context release情况： Normal release：“User Inactivity”，CS Fallback triggered，UE Not Available for PS Service，Inter-RAT Redirection，Time Critical Handover，Handover Cancelled，“Detach
19、” Abnormal release：“Radio Connection With UE Lost”, “CSG Subscription Expiry”, “Redirection towards 1xRTT”，Unspecified Failure，O&M Intervention，Repeated RRC signalling Integrity Check Failure，etc.,ENodeB Initiated,UE context release,释放的触发条件UE Context Release Command,MME Initiated,该信令，使MME在多种情形下，可以控制
20、UE-associated 链接的释放；MME发起的Context release情况： Normal release： e.g, completion of a transaction between the UE and the EPC, or completion of successful handover, or completion of handover cancellation, or release of the old UE-associated logical S1-connection when two UE-associated logical S1-connecti
21、ons toward the same UE is detected after the UE has initiated the establishment of a new UE-associated logical S1-connection, or the UE is no longer allowed to access the CSG cell (i.e., the UE becomes a non-member of the currently used CSG cell). Abnormal release： “Load Balancing TAU Required”，auth
22、entication failure, detach, not allowed CSG celetc.,UE context release,正常释放原因信令实例,eNB initiated,MME initiated,MME initiated,eNB initiated,eNB initiated,UE context release,异常释放原因信令实例,MME initiated,eNB initiated,MME initiated,MME initiated,MME initiated,释放流程信令详解,UE context release：ENodeB Initiated,The
23、 UE CONTEXT RELEASE REQUEST message shall indicate the appropriate cause value, e.g., “User Inactivity”,“Radio Connection With UE Lost”, “CSG Subscription Expiry”, “CS Fallback triggered”, “Redirection towards1xRTT”, “Inter-RAT Redirection”, “UE Not Available for PS Service”, for the requested UE-as
24、sociated logical S1-connection release. Interactions with UE Context Release procedure: The UE Context Release procedure should be initiated upon reception of a UE CONTEXT RELEASE REQUEST message.,释放流程信令详解,UE context release：ENodeB Initiated,Transition due to inavtivity. UE consumes less battery pow
25、er in idle mode; Less active Ues which network has to handle; Fast going to inactivity has drawback of longer connection setup time(more transitions idle to connected).,释放流程信令详解,UE context release：ENodeB Initiated,释放流程信令详解,UE context release：MME Initiated,The UE CONTEXT RELEASE COMMAND message shall
26、 contain the UE S1AP ID pair IE if available, otherwise the message shall contain the MME UE S1AP ID IE. The MME provides the cause IE set to “Load Balancing TAU Required” in the UE CONTEXT RELEASE COMMAND message sent to the eNB for all load balancing and offload cases in the MME. MME UE S1AP is re
27、leased when the S1 connection is released (next time a new value is assigned) eNB UE S1AP is released,释放流程信令详解,UE context release：MME Initiated,Upon reception of the UE CONTEXT RELEASE COMMAND message, the eNB shall release all related signalling and user data transport resources and reply with the
28、UE CONTEXT RELEASE COMPLETE message. In case of eNB supporting L-GW function for LIPA operation, the eNB shall also release the node internal tunnel resources.,释放流程信令详解,UE context release：MME Initiated,释放流程信令详解,UE context release：MME Initiated,释放流程及异常释放原因分析,主要内容,业务承载类型及业务建立过程 业务释放类型及相应流程、原因 E-RAB re
29、lease UE context release RRC connection release Abnormal release的触发情景详解 UE initiated drop eNB initiated drop 异常释放related Counter Triggers Network/Field KPI及Drops分析 Summary,RRC connection release,释放类型,根据3GPP定义(3GPP TS 36.331)，E-UTRAN initiates the RRC connection release procedure to a UE in RRC_CONNE
30、CTED： RRC connection release The purpose of this procedure is to release the RRC connection, which includes the release of the established radio bearers as well as all radio resources. RRC connection release requested by upper layers The purpose of this procedure is to release the RRC connection. Ac
31、cess to the current PCell may be barred as a result of this procedure.,RRC connection release,释放原因,The release of the RRC connection is initiated by E-UTRAN. The procedure may be used to re-direct the UE to another frequency or RAT. In exceptional cases the UE may abort the RRC connection, i.e.move
32、to RRC_IDLE without notifying E-UTRAN.,释放流程及异常释放原因分析,主要内容,业务承载类型及业务建立过程 业务释放类型及相应流程、原因 E-RAB release UE context release RRC connection release Abnormal release的触发情景详解 UE initiated drop eNB initiated drop 异常释放related Counter Triggers Network/Field KPI及Drops分析 Summary,Call drop,Who Initiates Drops, UE
33、, MME or eNB?,UE can ”initiate a drop” by starting RRC Connection Re-establishment Note: besides RLF also other causes can trigger RRC Conn Re-EstablishmenteNB can initiate abnormal S1 + RRC release due to Radio network layer problem (TA timer, RLF, PDCCH Order failure) Transport network layer probl
34、em (GTP-U error, Treloc expiry, Path Switch problem) Other abnormal causeMME can initiate abnormal S1 + RRC release due to Radio network layer problem Other abnormal cause,UE initiated drop,RLF,When UE is in RRC_CONNECTED and RRC security is active, it can trigger RRC Connection Re-establishment upo
35、n T310 expiry upon reaching the maximum number of UL RLC retransmissions upon handover failure (T304 expiry) upon non-HO related random access problem If successful, RRC Conn Re-Establishment reconfigures SRB1 to resume data transfer of RRC msgs re-activates RRC security without changing algorithms
36、NOTE: DRB should be re-configured after SRB1 NOTE: if UE is in RRC_CONNECTED while RRC security is not active, UE goes to RRC_IDLE, performs cell reselection and TAU,UE initiated drop,RLF due to T310 Expiry at UE,UE initiated drop,RLF due to Maximum UL RLC Re-Tx Reached,from RRC Connection Reconfigu
37、ration: drb-ToAddModListdrb-ToAddModList value 1drb-Identity : 1rlc-Configamul-AM-RLCt-PollRetransmit : ms40pollPDU : p32pollByte : kB25maxRetxThreshold : t8,UE initiated drop,RLF due to HO Failure,from RRC Connection Reconfiguration:mobilityControlInfotargetPhysCellId : 33t304 : ms1000newUE-Identit
38、yBin : 14 EB (= 5355),UE initiated drop,RLF Due to Non-HO Random Access Failure,“Non-HO random access” means PDCCH order -triggered RA Random Access Scheduling Request,释放流程及异常释放原因分析,主要内容,业务承载类型及业务建立过程 业务释放类型及相应流程、原因 E-RAB release UE context release RRC connection release Abnormal release的触发情景详解 UE i
39、nitiated drop eNB initiated drop 异常释放related Counter Triggers Network/Field KPI及Drops分析 Summary,eNB initiated drop,RLF,eNB can drop the call due to following triggerseNB-detected radio link problems PUSCH RLF CQI RLF Ack/Nack RLF PDCCH Order failure SRS RLF TA timer expiry Maximum RLC retransmission
40、s exceeded GTP-U failure,eNB initiated drop,Radio Link Problem Detection at eNB,eNB radio link problem detection mechanisms are NSN-internally specified Multiple methods (called “link monitors”) are defined to detect a radio link problem in the eNB. When one link monitor detects a problem, it is rea
41、lly a radio link problem even if other link monitors have not yet indicated anything. Each link monitor has its internal criteria to decide when radio link problem is flagged and de-flagged (radio link recovers). If the RLP persists longer than T_RLF, RRC+S1 release is triggered T_RLF = t310 + t311
42、(eNB-internal timer) Link monitors: Uplink PUSCH DTX detection for scheduled uplink data CQI DTX detection for periodic CQI reports in PUCCH and PUSCH Uplink Ack/Nack DTX detection for transmitted downlink data PDCCH Order RLF SRS DTX detection (TD-LTE),eNB initiated drop,1. PUSCH RLF: RlsCause_Pusc
43、hRlf_ON,When UE is scheduled for PUSCH transmission, eNB expects to receive UL transmission on the scheduled PRBs If signal from UE cannot be detected at all, PUSCH DTX is declared NOTE: The case where UL TBS is received but it fails CRC check is not DTX (its a NACK) DTX PUSCH indication is provided
44、 by the UL physical layer. The result is received by LTE MAC in reliableULtransmissionFlag parameter. Both counter-based and timer-based RLF detection is supported Timer-based PUSCH RLF detection: If “DTX“ is received on the PUSCH for a configurable period of time (rlpDetMaxTimeUl), PUSCH RLF is set
45、 on Counter-based PUSCH RLF detection: If “DTX“ is received on the PUSCH for a consecutive number of times (rlpDetMaxNoUl), PUSCH RLF is set on The recovery of the radio link is indicated when for a configurable number of contiguous UL resource assignments data is detected on PUSCH (ACK or NACK rece
46、ived). Defined by parameter rlpDetEndNoUl.,eNB initiated drop,1. PUSCH RLF: RlsCause_PuschRlf, Counter-based RLF Detection Example,T_RLF = T310 + T311,3 2,vendor-file parameters in this example:,eNB initiated drop,2. Periodic CQI RLF: RlsCause_CqiRlf_ON,The eNB supports CQI DTX detection for periodi
47、c CQI reports on PUCCH and PUSCH. If MAC layer receives nCqiDtx consecutive reports from UL PHY, the MAC declares CqiRLF_ON If the MAC has set CqiRLF_ON for a specific UE and nCqiRec consecutive CQI reports are again detected successfully for that UE, the MAC sets CqiRLF_OFF The parameters nCqiDtx a
48、nd nCqiRec are in the vendor-specific parameter file For both PUSCH and PUCCH the periodic CQI is encoded using a Reed Muller block code and comes along without any CRC. Hence, the UL PHY indicates a DTX detection for periodic CQI reports on PUCCH or PUSCH whenever a report is configured but no reli
49、able transmission from the UE could be detected. So the output of the detector shall be either the detected CQI report or a DTX indication. NOTE: CQI_RLF detection does not apply to aperiodic CQI report on PUSCH,eNB initiated drop,2. Periodic CQI RLF: RlsCause_CqiRlf, Example,6 1,vendor-file paramet
50、ers in this example:,T_RLF = T310 + T311,eNB initiated drop,3. Ack/Nack RLF: RlsCause_AckNackRlf_ON,After DL scheduled data, eNB expects HARQ ACK or NACK on PUCCH or PUSCH at known UL TTI Timer-based ACK/NACK RLF detection: If ACK/NACK “DTX“ is received for a configurable period of time (rlpDetMaxTi
51、meDl), ACK/NACK RLF is set on Counter-based ACK/NACK RLF detection: If ACK/NACK “DTX“ is received for a consecutive number of times (rlpDetMaxNoDl), ACK/NACK RLF is set on The recovery of the RLF is indicated when for a configurable number of contiguous ACK/NACK opportunities ACK or NACK is detected on PUSCH or PUCCH (no DTX). Defined by parameter rlpDetEndNoDl. .,
