Bearers, default Bearers, dedicated Bearers – naughty Bearers

Posted: January 25, 2010 in technical
Tags: , , , , , , , , , , , ,

What about those LTE bearers? What exactly is a bearer?

Well, if we are to believe the 3GPP guys (3GPP TS 23.401 version 8.6.0 Release 8), an EPS bearer is a data structure (that appears on the UE, MME, SGW and PGW), a way of uniquely identifying a traffic flow between the UE and the PGW. We need to _uniquely_ identify these flows because of the QoS we want to use for that UE traffic.

When are these bearers created?

First of all, there are at most 11 bearers that can be created for a specific UE. 11 bearers TOPS – per UE. Why is this so important?


1. the first time an UE connects to an anchor point (PGW) – procedure called Initial Attach, simply by allowing that UE access on the PGW – a new (default) bearer is created – and, yes, those 11 bearers tops decrease once this happens!!!

2. an UE can be “attached” to more than 1 anchor point (PGW) – which means, an UE can have more than 1 “default”/”initial” bearers (of course, created via multiple Initial Attach procedures) – which means those 11 bearers tops decrease again

Leaving us with the rest of the bearers, those NOT created “by default” at the Initial Attach procedure, those which we call dedicated bearers.

***Note: there are not necessarily 11 bearers up and running all the time. The “11” is just the max number that can be active at a certain moment.

How do I use the bearers for QoS?

Each bearer, once created, has assigned a certain TFT set. “TFT” stands for Traffic Flow Template, the set of all packet filters associated with that certain bearer (we’ll look later on soon at the wireshark capture to see exactly how these “bearer” and “tft” look like).

How do I use the TFT for QoS?

TFT, being a set of packet filters, resides as a database tuple in the PCRF – Policy Control and charging Rules Function, a separate cute device that tells the PGW how to route, where to route, and what QoS to use for traffic flowing to and from a certain UE.

! Moment of thinking 1:

HSS – Home Subscriber Server

PCRF – Policy Control and charging Rules Function

The HSS is a database that holds only information regarding the default bearer (which basically identifies the UE as belonging to this network), while the PCRF has the role of “traffic shaping”.

! Moment of thinking 2:

Although the default bearer is more or less automatically created when the UE attaches to this network, as a network confirmation that this UE belongs to it, the dedicated bearer is NEVER initiated by the MME/UE (even if it is, the PGW will gracefully ignore it 😛 ) – the dedicated bearer will ALWAYS be initiated by the PGW, in response to a certain traffic pattern matching a rule in PCRF, though triggering the creation a new and shiny TFT.

  1. vmp says:

    s/rute/route/g ?

  2. @vmp: what would I do without my QAEs ? 😀 Thank you 🙂

  3. Stefan says:

    There can be one default bearer per APN, and IIRC the set of 11 bearers is per APN as well. So the second default bearer is from another set. Thus you have 10 bearers (flows) for dedicated bearers, which at least at this point in time seems enough (just like 64K a while ago 🙂 ).

  4. Santosh says:

    Ok! So a UE can have atmost 11 bearers, no matter to how many APN’s it is connected to. EBI is used to identify the bearer which is a 4 bit field, so there can be 16 values out of which 5 are reserved which leaves us with 11. Now why EBI is 4 bits? This is because of 3G convergence. In 3G we have NSAPI which is of 4 bits and it is mapped to EBI in LTE. Hence EBI is of 4 bits. No idea why NSAPI is set to 4 bits 🙂

    UE identifies the bearers only with EBI, that means UE maps traffic based on the EBI value. SO you just cannot have two bearers with same EBI no matter if they are connected to different APN’s.

    Dedidcated Bearer: Well UE/MME cannot create a dedicated bearer but they can request for it. PGW can accept the request or deny it.

  5. @Santosh: hey, thanks for the correction 🙂

  6. […] I was saying in a previous post, Bearers are structures that define a single QoS traffic between eNB – MME – SGW […]

  7. Rajeshwar says:

    @all : Is PCRF mandatory if I will use default bearer, If no what if P-GW sends a request to PCRF and PCRF didnt reply with a rule , will PDP attach successful ?

    2) How to identify if the request is for Dedicated bearer or default bearer ?

  8. @Rajeshwar: sry for the late reply, I was on vacation, no Internet access 😦
    1. As far as I understand this, no, there is no need for a PCRF if I only do the UE attaching. Still, the PGW would try to find out rules about this UE when it attaches, so that it knows what policies to apply next: like, immediately after attach, the PGW may create dedicated bearers.
    Also, according to TS 23.401, the only case where the PCRF is implicated in the attach procedure is in the case of IP-CAN Session Termination/Establishment. I am not so familiar with this IP-CAN thinggie, but it seems that many applications may require this PCRF verification in order for the PGW to have the OK to send an IP address to the UE.
    So, in case we are talking about this type of implementations (which I believe should be the rule), the PCRF must exist, as it is the one ordering the PGW to send an IP to the UE or not to let it attach to the network.

    2. Well, there are several ways…;the default bearer request is the first to come immediately after attach; there 4 messages exchanged here, from the MME to the SGW (and further on to the PGW): Create Session Request/Response (which attaches partially the network and allows uplink to be sent) and then Modify Bearer Request/Response – which actually creates the full bidirectional default bearer link;
    in case of Multiple APN per UE, this works the same, just that the request is for a different APN (different APN string in the Create Session Request that reaches the PGW);

  9. Vijay says:

    @All: As Santhosh pointed out, ” A UE can have 11 bearers atmost irrespective of the APN’s it is connected to”. So can I assume a UE can have 11 default bearers with 11 different PDN’s ?

  10. Vijay says:

    @ All: Also another thing I assume is the UE can at the max have 10 dedicated bearers as one of the 11 will be default (default gets created the moment UE is switched ON). Correct me if I am wrong.

  11. @Vijay: you are right 🙂 at most 10 dedicated bearers per UE

  12. James says:

    When and how to release default bearer? As it can exist even when UE has no connection to E-nodeB.

    • @James: Firstly, a default bearer is a pair of information saved on the eNB, MME, SGW and PGW. At a certain moment in time, all the bearers, including the default one, may not be active – ECM-IDLE mode. When and how you can release a default bearer?

      Let’s take the case when the UE is in ECM-CONNECTED, so all its bearers are up all the time:
      In this case, “releasing” – if you mean by it “deleting” the default bearer, means detaching from that network-APN. The default bearer can be released when the user loses signal, implicitly; the user can also simple detach from 1 APN, while remaining attached (keeping the default bearers at least) for other APNs. When this happens (via one of the PDN detach procedures, the MME sets a certain flag in the Delete Session message: Operation flag = 1) – this tell the PGW that this UE is detaching from the network.

      ECM-IDLE case:
      When the UE moves to ECM-IDLE, the network receives another type of message: Release Access Bearer Request. This tells the network that the particular eNB where the UE was/is connected is going to delete all the information about that UE from its local resources. The network deletes the USER-PLANE TEIDs (bearer information for user-plane, for data traffic), regarding that UE.
      In this case, if the UE intends to detach itself from the network, it must first exit idle state, return to an active state, and barely then delete the session. You may ask: how does the network know where the UE is at a certain moment in time? – as there is no more data, and the user may travel accross country while in Idle. True, but the UE still signals to the network its position (see Tracking Area Update procedure). This is very important as, if downlink data arrives at the SGW, it must awaken the UE before sending it traffic – the UE is moving from ECM-IDLE to ECM-CONNECTED (re-creating its data traffic bearers). Unless there was the TAU procedure, the SGW won’t know which MME (and further eNB) to signal to get hold of that UE. Even worse, the UE may move out of the coverage area of its current SGW – so the downlink packets may have to reach other SGW – that’s part of the role of the TAU procedure type MME relocation, SGW relocation.

  13. Donna says:

    About your question: When and how you can release a default bearer?
    In 23.401, PDN GW initiated bearer deactivation, it is said: “If the default bearer belonging to a PDN connection is deactivated, the PDN GW deactivates all bearers belonging to the PDN connection.” It does not point out that multiple-PDN is required so we can assume there are cases that the default bearer is deactivated to delete all of the bearers even only with one PDN connection. Am I right? Based on this, I’d like to ask what signalling flow goes for this procedure? Between UE and MME, the NAS is DEACTIVATE EPS BEARER CONTEXT
    ACCEPT; But what is the message between eNB and MME, eNB and UE? In S1 interface, I think UE CONTEXT RELEASE COMMAND or Detach Request may be appropriate to trigger this procedure and RRC_Connection_Recfg in Uu interface. But from in 23.401, seems it is ERAB Release command in S1 and RrcRecfg in Uu, same as dedicated RAB release. What’s your comments? Thank you.

  14. Donna says:

    As a complement, my question can be shorten as:
    Assume that UE is in ECM-CONNECTED with one default bearer and one dedicated bearer established. Then MME sends DEACTIVATE EPS BEARER CONTEXT
    REQUEST with default E-RAB ID to UE. What is the eNB’s incoming message from MME and out-going message to UE which carries NAS-PDU (DEACTIVATE EPS BEARER CONTEXT
    REQUEST with default E-RAB ID )?

  15. qiuhong says:

    hi, I feel confused.36.321 Identity of the logical channel for DRB (DTCH) only 8.That mean max 8 dedicated bearers.But in 36.331 6.4 section ,the value of maxDRB is 11。
    this paper also tell me there are 11 bearers one UE established。So I want to ask you what the number 11 come from。 sorry ,my English is poor。

    • @qiuhong: hmmm, I can’t remember precisely where I took this information from an year ago.

      But, regarding what you say: TS 36.321 (at least Release 8 – which is now being implemented as 4G), section 6.2.1, there are 2 tables: Table 6.2.1-1 Values of LCID for DL-SCH and Table 6.2.1-2 Values of LCID for UL-SCH : both indicate the following range for the “Identity of the logical channel”: 00001-01010.

      While TS 36.331, indeed clarifies: maxDRB INTEGER ::= 11 — Maximum number of Data Radio Bearers

      There are 11 bearers available 😛

  16. Dumitru says:

    I have some clarification to this coment(14):
    “Unless there was the TAU procedure, the SGW won’t know which MME (and further eNB) to signal to get hold of that UE.”

    In ECM-IDLE SGW always has the TEID and MME IP address(for control plane) for each attached subscriber and at the same time it does not have any knowledge of current TAC (it does not need this knowledge because it already has an available tunnel TEID+IP address to MME)

    TAU is needed to be known by MME in order to page the UE. (for example when there’s pending downlink traffic in SGW)

  17. Dumitru says:

    Also one more comment to “Moment of thinking 1”
    “The HSS is a database that holds only information regarding the default bearer (which basically identifies the UE as belonging to this network), while the PCRF has the role of “traffic shaping”.”

    PCRF consist of at least 2 main components(beside PCEF):
    1. SPR
    2. PCRF

    SPR (Subscriber Profile Repository) usually/shall/may be collocated with HLR/HSS and may also share the same back-end (enhanced DB structure) this way SPR functionality is used to enhance HSS functionality with per subscriber and service policy.
    In this context SPR is close to HSS, while PCRF is how you correctly mentioned “PCRF has the role of “traffic shaping””

  18. Nan says:

    Are uplink and downlink (dedicated) bearers independent of each other in terms of establishment, release and QoS parameters?

    • @Nan: hey. First of all, thanks for the questions. I’ll try to answer them punctually, as far as I know 🙂

      1. QoS parameters: are always separate. By separate, I mean there are separate IEs for them: GBRU/GBRD (Guaranteed Bit Rate Uplink/Downlink).
      There can also be separate TFT (Traffic Flow Template) strings (which are basically Layer 3 – 4 Access Lists) for Uplink and for Downlink.

      2. I’ve never seen, nor read any document that described negotiation as separate for Uplink and Downlink traffic. In my opinion, there is no reason this can’t be possible. I believe it is possible to negotiate these bearers separately. You just put, for example, GBRU = 0, and GBRD = 2000 in order to only negotiate values for the Downlink bearer. But then you will have to do more signaling overhead in order to also negotiate the Uplink.
      I believe it’s good to just negotiate them both at the same time. (And don’t worry if you want to have separate TFTs, because you can aggregate and add more TFT strings on the same packet).

      3. I would say the same as 2. above for the release aspect.


  19. Santosh says:

    Hey…there is no QoS negotiation in LTE. PCRF dictates the values. Network/UE either can accept it or reject it but MME/SGW/PGW cannot negotiate.

    In 3G SGSN can negotiate the QoS

  20. @Santosh: fair enough. It was just using the term “negotiation” for…establishment, or transmission.
    Thank you for the attention 🙂

  21. Nan says:

    Cristina-Thanks. I have a followup: I think you are talking about so called “bi-directional” secondary bearers with independent parameters. But uni-directional ones may suffice – for example, if I want to stream down video (via HTTP download), I need the QoS on DL but the UL carries HTTP requests and TCP ACK/NACKs which can be shipped on the default UL bearer, right?

  22. Mahesh says:

    I want to how many default bearer a UE can do at a time?

  23. Josh says:

    Asume a default EPS and say 2 dedicated bearers have been established to the same PGW. What happens when the UE goes to ECM-IDLE? For the the Default bearer it is clear that the tunnel IDs and in general the bearer context will remain active between S-GW and P-GW (s5). However radio resources and S1 bearers will be released. Will the tunnel IDs and QoS contexts preserved in SGW and PGW for the dedicated bearers as well? If yes, for how long?

  24. Alexandra says:

    Hi Josh,

    In TS23.401 section 5.3.5 S1 release procedure, the procedure used for entering IDLE mode, we found the following pieces of information:
    – during this procedure S1-u bearers are released (S1-U is the interface between the eNB and SGW)
    – the PGW is not informed of this release, so the release has no impact on the PGW
    – only eNB related information (address and TEIDs) is released from the SGW. No other elements (including QoS) of the UE’s context are affected.

  25. Ritu says:

    Since the PDN always has the default bearer to reach the UE, establishing a new dedicated bearer becomes simpler. Why ?

    • Hey, Ritu

      What is the question, more exactly? Why is it simpler to create a dedicated bearer in 4G than in 3G?
      Answer may be: because the 4G designers tried to create a simpler design. Simpler is better, more scalable and safer.

  26. Asha says:

    Is it possible to have more than one bearer in INITIAl CONTEXT SETUP message ?

    I think yes because there is E-RAB to be Setup List in this messahe which can contain multiple bearer .

    How can we differntiate that which one is default and which is dedicated bearer here ?

    Many thanks .

  27. qiuhong says:

    hi,everybody.I’m confused a question. TFT is from UE by NAS ,or from MME? Refer to 24.008 ,UE can transfer TFT to CN .

  28. Chou says:

    Hi cristina_crow,
    Could you tell me that how many elements to be include in defualt bearer? and dedicated bearer? i.e. abmr or arp.

  29. Chou says:

    the more detail , more better. thanks a lot.

  30. @Chou:
    A default/dedicated bearer has Mandatory, Conditional and Optional elements. The Specs (ex.: TS 23.401, TS 29.274) (should) show exactly which elements are included, and under which conditions. So please don’t ask me to re-write here an entire Spec 😛 , and read it for yourself: at least TS 29.274, section 7.2 Tunnel Management Messages, where you should see which elements are put in the Create Session Request/Response + Modify Bearer Request/Response messages (responsible for negotiating default bearers) and which elements are in the Create Bearer Request/Response messages, responsible for creating dedicated bearers.
    With regards to the radio bearers, I’m afraid I can’t help you much – but hopefully some other nice blog reader answers that one 🙂

  31. Alexandra says:

    Hi Chou,

    As far as I know the bearers can be categorized by two criteria:
    – Default vs. dedicated bearers
    – GBR vs. non GBR bearers
    GBR = guaranteed bit rate.
    One default bearer is created per APN connection. 0 or more dedicated bearers can be created on each APN connection, there is a limit though of 11 bearers/UE.
    The default bearer is a non-GBR bearer, while the dedicated bearers can be GBR on non-GBR.
    TFT = traffic flow template, is a set of packets filters that describe what traffic will use a certain bearer.
    The default bearer may or may not have an associated TFT. It usually doesn’t or is very general.
    The dedicated bearers, as far as I know, always have assiciated TFTs
    A non-GBR bearer has the following characteristics:
    – QCI (QoS Class Identifier)
    – ARP (Allocation and Retention priority)
    – TFT (optional)
    A GBR bearer has the following characteristics:
    – QCI (QoS Class Identifier)
    – ARP (Allocation and Retention priority)
    – GBR (Guaranteed bit rate) – for up link and down link
    – MBR (Maximum bit rate) – for up link and down link
    – TFT

    The details above can be found in TS 23.401section 4.7.2 The EPS bearer
    The APN-AMBR (APN-Agregated maximum bit rate) is the maximum bit rate and UE is allowed to achieve on all the non-GBR bearers on a certain APN connection.
    The UE-AMBR is the sum the maximum bit rate and UE is allowed to achieve on all the non-GBR bearers and is usually the sum of all the APN-AMBR.
    The APN-AMBR and UE-AMBR are characteristics assigned by subscription to and UE. They are retrieved by the MME from the HSS. The APN-AMBR is set by the MME to the SGW and PGW during attach and can be changed during this process or later.
    What information is kept for each UE and each APN on each network entity is described in TS 23.401 section 5.7 Information storage.
    The entities that do traffic policing/shaping are the eNB and the PGW, the eNB for up link traffic and the PGW for down link traffic. The eNB is not informed of the APN-AMBR, but of the UE-AMBR.
    The UE, for each bearer retains the TFT, the QoS values (ARP, GBR, MBR, QCI) and the core network IPs and TEIDs

    cristina_crow, please correct me where I’m wrong and and what I’ve missed.

    Chou, let me know if this is answering your question.

  32. I guess it’s fine, Alexandra.

  33. Chou says:

    Yes. That’s good! Thanks a lot.

  34. Joseph says:

    Hi – re comment 19 – a UE can have up to 11 EPS Bearers (0101-1111) but there are only 10 MAC LCIDs available (0001-1010) to identify Logical channels on the air interface, so how a UE have all 11 possible EPS Bearers created if only 10 of them can be identified by user channel LCIDs? I’m struggling with the maths a bit!

  35. Adnan says:

    To Comment 41..
    I am trying to find answer to same thing.. please share with us if you come to know

  36. WebHTB – Advanced Linux Traffic Shaping, Traffic Control, QOS…

    […]Windancer – Stairway to …Heaven? » Blog Archive » Bearers, default Bearers, dedicated Bearers – naughty Bearers[…]…

  37. Joseph says:

    OK, I think I found the answer.

    If you look at 36.331:Annex B.1 in the part of the table for ‘bit 20’ it says that the maximum number of bearers a UE is required to support is 2x SRB (SRB1, SRB2) and up to 8 DRBs. Now, 2 SRBs plus 8 DRBs gives us ten logical channels, which exactly matches the number of LCIDs available per UE on the air interface.

    So, the maximum number of active EPS Bearers a UE is able to support is actually limited to 8.

    The 11 EBIs may still all get used up, if new bearers are created as old ones are being torn down, for example.

    8 DRBs/EPS Bearers is the maximum that a UE is required to support, individual UE types might support fewer than that to reduce their cost and complexity. M2M UEs, for example, might only be able to support 2xSRB and 1xDRB.

  38. Kevin Keaveny says:

    Is this blog still active?

  39. Raji says:

    How the direction is known to UE from MME since no parameter related to direction is there in ERAB_SETUP_REQUEST message.

  40. Vladimir says:

    While performing TAU by a UE that comes from UMTS the MME sends TAU_Reject with the reason “No EPS Bearers Activated”.
    This happens when the UE sends EBI(5) inactive in TAU_Request.
    Why do cirtain UEs send EBI(5) inactive and others send EBI(5) not-inactive (comming from UMTS)?
    Does it depend on the NSAPI(5) inactive/not-inactive state they had in UMTS?

  41. Hooman says:

    If a default bearer goes down will all the dedicated to that same go down also?

  42. Yuri says:

    Returning to the Donna ‘s question of December 21, 2010.
    What S1AP message carries NAS “DEACTIVATE EPS BEARER CONTEXT REQUEST” when perfoming an “UE requested PDN disconnect procedure” (TS 24.301 6.5.2) ?

  43. thousid says:

    we have any parameter to set the QCI 9 for all default bearer (even in the case of IRAT HO )?

    • cristinacrow says:

      hi thousid
      it’s been a while for me :), but I *think* that’s the QCI for the default bearer (first default bearer) – afaik this could be changed by the operator (up to them), but the handover itself should not influence its QCI (b/c the bearerId is hooked into the PGW, which doesn’t change during the handover – even if it’s an IRAT HO) …please do double-check this with the 3GPP Spec

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s