Table of Contents
An LTE service provider should be able to make services with different QoS requirements available to its users with different subscription levels. So, the service provider needs to be aware of subscription levels of each user (e.g. premium, best effort, etc.) and requested service types (e.g. Internet, voice, etc.) in order to assign radio and network resources to the traffic of each user and manage them properly.
For this reason, the LTE network first classifies user traffic (IP flows) into different SDFs having different QoS classes based on the type of the service that is being provided through the SDFs, and then applies different QoS rules to each SDF. Since SDFs are delivered through EPS bearers in the LTE network, EPS bearer QoS has to be controlled in a way that SDF QoS is maintained.
In this document, QoS mechanisms for SDFs and EPS bearers are explained as QoS mechanisms in LTE, and examples showing how they work are provided.
This document is organized as follows:
Chapter 2 defines an SDF and EPS bearer and
Chapter 3 describes QoS parameters for SDFs and EPS bearers.
Chapter 4 covers QoS provisioning and QoS enforcement, describing by which entity QoS parameters are provisioned and where they are enforced. In Chapter 5, some examples showing how the LTE QoS works are provided and explained based on the concept of QoS discussed in Chapters 2 and 3.
I have a doubt about QoS per SDF. I tried reading 23.203 as mentioned in your reference. However I could not find any information which says there should be a specific MBR value associated with a SDF in case of Non-GBR SDF. If possible kindly guide me to a particular section from where I can draw this conclusion.
A PCC rule defines the policy control information for a service data flow.
Table 6.3 lists PCC rule information (Rule ID, SDF detection, Charging, Policy control, ..). UL/DL-maximum bitrate and UL/DL-guaranteed bitrate are being defined as (part of) Policy control information. For non-GBR SDFs, UL/DL MBR is conditional, but can be used for rate policing.
Thank you. (and sorry for late reply due to business trip..)
First of all Thankyou for such wonderful presentations.
Based on what is written in this document and in the relevant 3GPP specifications, I have developed an understanding why a distinction needs to be maintained between EPS Bearer Flow and SDF. The 3GPP specifications define EPS Bearer as "An EPS bearer uniquely identifies traffic flows that receive a common QoS treatment between a UE and a PDN GW ". All the LTE/EPC nodes work with an EPS Bearer and one EPS Bearer can have multiple services mapped to it. A guideline for mapping the different services to different QoS classes is also provided by the specifications. However the specifications do not talk about the bit rate values for individual services and assignment of bit rate is left to the individual operators. In the absence of SDF, this notion of working only with EPS bearer can lead to 2 problems:
Having the notion of SDF can effectively address both the above problems. If PGW can segregate the IP flows into different Service Data Flows, it can
Request you to kindly confirm if this understanding is correct.
Thanks for sharing your understanding! We are in agreement with your comments.
Thanks NetManias for sharing such a nice information and thanks Vimal for sharing detailed understanding, it helped me to understand the sense of SDFs.
Very nice presentation.
Just a small typo error : Service Date Flow should be Service Data Flow
Thanks for the useful information & nice presentation. Can anyone please tell the exact difference between SDF Template and TFT?
Please refer Table 5. Description for MBR.
It is wrongly mentioned that this QoS parameter applies for non-GBR bearers.
SDF(service data flow) are used to map packets to the TFT(traffic flow template) based on 5 tuples and TFT is responsible for maping packet to the respective bearer which matches the QOS of bearer and the packet.
Is it correct?
SDF template vs. TFT
An SDF template consists of IP packet filters, and each IP packet filter has a set of filter rules (5-tuple consisting of Source IP, Destination IP, Source Port number, Destination Port number, and Protocol ID). IP flows arriving at P-GW are identified by matching these packet filters (SDF template), one by one, and those that match the same filter rule are grouped together and classified as an SDF.
P-GW keeps traffic mapping information (TFT filters) that indicates which EPS bearer can support (QoS of) which SDFs, i.e. QoS the LTE network can support for UE.
Each SDF has a set of QoS parameters (e.g. QCI, ARP, MBR, GBR) and so does each EPS bearer (e.g. QCI, ARP, MBR, GBR, UE-AMBR, APN-AMBR). An EPS bearer has the same variety of QoS parameters that mapped SDF(s) have, except it has access system-dependent parameters as well.
SDFs (or SDF aggregate) with the same QCI and ARP can be delivered through the same EPS bearer. If an SDF cannot be served at any of the currently available EPS bearers (e.g. if a new SDF, SDF 5 (GBR=500Kbps), arrives when SDF 3 (GBR=1Mbps) and SDF 4 (GBR=1Mbps) have already been delivered through EPS bearer 5 (GBR=2Mbps)), another EPS bearer can be created to serve the new SDF.
1>Please explain a bit on Protocol ID in SDF Template.
2>During downlink:how come PGW differentiate whether traffic is voice data (RTP), video streaming, voice signaling (SIP), two-way game, and Internet traffic.
3>Like max number of EPS bearer for one UE can be 11 because of EBI bit value of 4.My question is where in your website i can have understanding of other bit values and there relative constraints.
1. Protocol ID: In the IPv4 there is a field called "Protocol" to identify the next level protocol. This is an 8 bit field (Protocol ID = 6 for TCP and 17 for UDP). In IPv6, this field is called the "Next Header" field.
2. PGW differentiates traffic type based on SDF filters (5-tuple based packet classification rule) and these fileters are part of PCC rule. PCC rule (SDF filters) can be provisoned by PCRF (dynamic PCC) or pre-configured on PGW (predefined PCC). Note that SDF filters are unidirectional, so detection is applied independently for downlink and uplink direction.
- SDF filters for voice data (RTP) are provisioned by PCRF (IMS core -> PCRF -> PGW) whenever voice bearer (dedecated bearer) are established
- SDF filters for voice signalling (SIP) are predefined in PGW (TCP/UDP port = 5060 which means SIP)
- If it is not able to detect video streaming/two-way game traffic based on 5-tuple, DPI (Deep Packet Insepction) feature should be used (DPI supported PGW or external DPI box)
3. I cannot find any EBI value greater than 11 in this document.
First of all Thank you for this wonderful presentation.
I have just 1 query,
In your example (5.1 QoS Operation in Downlink), you have written that when IP packets comes at PGW , for GBR SDF , PGW apply MBR DL policing .
My question is why not GBR DL policing ?
Thanks a lot for wonderful explanation.
However there is one request while downloading it in pdf format, please include option to download comments too as they also contain sometimes valuable knowledge.
I am not able to understand how the SDF aggregation happens and how the sdf's mapping to EPS bearer ? My understanding in SDF aggregation is, For eg - 5.1 QoS Operation in Downlink
in that image SDF 1 and SDF 2 are GBR mapped to dedicated bearer, IS this called SDF aggregation or IP packet SDF 3 and SDF 3 marked in the image are going to SDF 3 this called SDF aggregation ?
Is there any document specifing the UE AMBR QOS values discussed for video calling during VoLTE?
For the downlink icmp packet, the icmp packet contains only the source address, destination address and protocol ID. But durinf SDF template filtering how the source port and destination port were mapped ? Please expain me in details.
Great document, this helps me to catch the idea that EPC/PGW is able to stablish dedicated EPS Bearer without using PCRF in the service set up. I thought before that PCRF is needed for stablishing dedicated bearer on this kind of scenarios. Many thanks
Is the commeny by vplazam is correct ?.....we can understand atleast during Initial attach defualt bearer is setted up based on QOS parameters set by HSS....but how come a dedicated bearer can be setted up by PGW alone without PCRF ? .......For Dynamic PCC rules in the case of realm time traffic we need a PCRF provisioning PCC rules for different service data flows
Atleast i can guess a NON GBR dedicated bearer can be setted up without PCRF
please explain if i'm wrong.....
I had following query - What is the DL/UL Bearer mapping in EPC Network across below interface ( where it is 1:1 and where There is Only one bearer needed for UL and DL Traffic Both )
b -How many maximum number of bearers that gets configured in EPC network for a UE in operator network . Standards says 11 But at certain places I see at max - 3 bearer can be configured per UE ( 1 default + 2 dedicated bearer )
c- what is the service to UL/DL to bearer mapping at each Interface on -