Node b and method for prioritization of retransmission of protocol data units to assist radio-link-control retransmission

The present invention relates to the field of wireless communication. More specifically, the present invention prioritizes the retransmission of protocol data units (PDUs) in order to support retransmissions in the radio link control layer. Regarding the system and method for raising.

Third-generation (3G) cellular systems with Frequency Division Duplex (FDD) and Time Division Duplex (TDD) provide reliable end-to-end data transmission. Therefore, there is a retransmission mechanism in the Acknowledgment Mode of the Radio Link Control (RLC) layer. The RLC layer is an isotopic entity that resides in both Radio Network Controller (RNC) and User Equipment (UE).

Figure 1 shows a block diagram of the MAC-hs layer configuration of the UMTS terrestrial radio access network (UTRAN), and Figure 2 shows a block diagram of the MAC-hs configuration of the user equipment (UE). .. The configurations shown in FIGS. 1 and 2 are described in detail in the co-pending Patent Document 1 filed on October 15, 2002, assigned to the assignee of the present invention. The UTRAN MAC-hs30 shown in Figure 1 has a transport format resource indicator selector 31, a scheduling and prioritization entity 32, and multiple hybrid automatic repeats (H-). ARQ Processor (Hybrid Automatic Repeat processor) 33a, 33b, Flow controller 34, and Prioritization Class / Transmission Sequence Number (TSN) Setting Entity (priority class and) transmission sequence number setting entity) 35 is provided.

The UE MAC-hs40 is equipped with an H-ARQ processor 41. As described with reference to FIGS. 1 and 2, the H-ARQ processors 33a, 33b in the UTRAN MAC-hs30 and the H-ARQ processor 41 in the UE MAC-hs40 work together to process the data blocks. ..

The H-ARQ processors 33a, 33b in UTRAN's MAC-hs30 handle all the tasks required for the H-ARQ process to make a transmission and to make a retransmission for an error transmission. .. The H-ARQ processor 41 in the MAC-hs40 of the UE is responsible for transmitting an acknowledgment (ACK) indicating that the transmission was successful and a negative response (NACK) indicating that the transmission was unsuccessful. The H-ARQ processors 33a, 33b, 41 process a continuous data stream for each user data flow.

As will be described in more detail below, the data blocks received for each user data flow are assigned to the H-ARQ processors 33a and 33b. The H-ARQ processor 33a and 33b start transmission, respectively, and in the case of an error, the H-ARQ processor 41 requests retransmission. For subsequent transmissions, the modulation rate and coding rate can be changed to ensure successful transmission. Data blocks and new transmissions resent to the UE are provided by the scheduling / prioritizing entity 32 to the H-ARQ processors 33a, 33b.

The scheduling / prioritizing entity 32 acts as a radio resource manager and determines the transmit latency to maintain the required QoS. The scheduling / prioritizing entity 32 transfers the data blocks to the TFRI selector 31 based on the output from the H-ARQ processors 33a, 33b and the priority of the new data blocks sent.

Combined with the scheduling / prioritizing entity 32, the TFRI selector 31 receives the data block to be transmitted and selects the appropriate dynamic transmission format for the data block to be transmitted. For H-ARQ transmission and retransmission, the TFRI selector 31 determines the modulation and coding.

For several reasons, it is strongly desired that the retransmitted data block arrive at the receiving (ie, UE) RLC entity as soon as possible. First, with lost data blocks, subsequent data blocks cannot be transferred to higher layers due to the requirement for ordered delivery. Second, the UE buffer needs to be large enough to handle retransmission latency while maintaining an efficient data rate. The longer the wait time, the larger the UE buffer size so that the UE can buffer the retained data blocks and the continuously received data until the data blocks in the correct order are transferred to the upper layer. Must. The larger the buffer size, the higher the hardware cost of the UE. This is highly undesirable.

Referencing Figure 3 shows a simplified flow diagram of the data flow between Node B (bottom of Figure 3) and UE (top of Figure 3). PDUs passed from higher level processing are scheduled and can also be multiplexed into a single data block. The data block can contain only higher layer PDUs with the same priority. A unique TSN is assigned to each data block by the scheduler. The upper layer can provide a plurality of streams consisting of PDUs having different priorities, and a continuous TSN is given for each priority. The scheduler is multiple H-ARQ processors P1_B~ P5_BAllocate data blocks to. H-ARQ processor P1_B~ P5_BEach is responsible for processing one block of data. For example, as shown in Figure 3, the priority 1 PDU is B1.₁~ B1_NIncludes the sequence indicated by. Similarly, priority 2 PDUs are B2₁~ B2_NPDUs with priority 3 arranged as B3₁~ B3_NIt is arranged like. These PDUs are scheduled (possibly multiplexed) and given a TSN by a common scheduler. To illustrate the invention, it is assumed that one PDU corresponds to one block of data. Each data block is processor P1_B~ P5_BProcessor P1 that processes the data block after being scheduled to be processed by one of_B~ P5_BAssociated with the processor identifier that identifies.

The data is then scheduled Node B H-ARQ processor P1 to receive and process each data block._B~ P5_BIs entered in. Node B H-ARQ Processor P1_B~ P5_BAre the H-ARQ processors P1 in the UE, respectively._UE~ P5_UECorresponds to. Therefore, the first H-ARQ processor P1 in node B_BIs the first H-ARQ processor P1 in the UE_UECommunicate with. Similarly, the second H-ARQ processor P2 in node B_BIs the second H-ARQ processor P2 in the UE_UECommunicate with. Remaining H-ARQ processor P3 in node B_B~ P5_BAnd the H-ARQ processor P3 in the other UE_UE~ P5_UEThe same applies to. The H-ARQ process is multiplexed on the air interface at the right time, and only one H-ARQ transmission is performed on the air interface at a time.

For example, H-ARQ processor P1_BAnd P1_UELooking at the first pair with which they communicate, the H-ARQ processor P1_BTo multiplex the data block and send it over the air interface, the data block, eg B1₁To process and transfer. This data block B1₁But the first H-ARQ processor P1_UEWhen received by processor P1_UEDetermines if the data block was received without error. Data block B1₁Was received without error, the first H-ARQ processor P1_UEH-ARQ processor P1 on the transmitting side that the data block was successfully received._BInform, send an ACK. Conversely, the received data block B1₁If there is an error in, the receiving H-ARQ processor P1_UEIs the sender's H-ARQ processor P1_BSend NACK to. This process involves the sending processor P1_BBut data block B1₁Continues until an ACK is received for. Upon receiving an ACK, processor P1_BIs "released" to process another block of data. The scheduler, if available, processor P1_BCan be assigned to another block of data and you can choose to resend or start a new transmission at any time.

Receiving H-ARQ processor P1_UE~ P5_UEProcesses each data block and then reorders them based on priority R₁, R₂, R₃Transfer to. Each sort buffer corresponds to each priority level of data. For example, priority 1 data block B1₁~ B1_NIs received and has a priority of 1 sort buffer R₁Sorted within. Priority 2 data block B2₁~ B2_NIs received and has a priority of 2 sort buffer R₂Sorted within. Priority 3 data block B3₁~ B3_NIs received and has a priority of 3 sort buffer R₃Sorted by.

Receiving H-ARQ processor P1_UE~ P5_UEDue to the data block preprocessing by, and the ACK / NACK response procedure, the data blocks are often received in an out-of-order order with respect to the TSN. Sort buffer R₁~ R₃Receives unordered data blocks and attempts to reorder the data blocks prior to transfer to the RLC layer. For example, priority 1 sort buffer R₁Is the first four priority 1 data blocks B1₁~ B1₄Receive and sort them. The data blocks are received, sorted, and then passed to the RLC layer.

On the receiving side, the H-ARQ processor ID may be transmitted on a control channel such as HS-SCCH, or it may be tagged with a data block, but as a UE MAC-hs (MAC-hs control). (Shown) reads the H-ARQ processor ID and which H-ARQ processor P1_UE~ P5_UEDetermine if was used. Same H-ARQ processor P1_UE~ P5_UEIf the UE receives another block of data processed by the H-ARQ processor P1_UE~ P5_UEThe H-ARQ processor P1 regardless of whether the preceding data block processed by_UE~ P5_UEThe UE knows that has been released.

Figure 4 shows an example of a prior art system that includes an RNC, node B, UE, and their associated buffers. In this example, it is assumed that UE is the receiving entity and node B is the sending entity. In this prior art system, the PDU with sequence number (SN) = 3 has not been successfully received by the UE. Therefore, the RLC in the UE requires the peer RLC layer in the RNC to retransmit. Meanwhile, PDUs with SN = 6-9 are buffered in node B, and PDUs with SN = 4 and 5 are buffered in UE. Note that in Figure 4, at most a few PDUs are buffered, but in practice more (eg, more than 100) PDUs and PDUs from other RLC entities can be buffered. I want to be.

As shown in Figure 5, if a PDU with SN = 3 is requested to be retransmitted, this PDU must wait at the end of the queue in Node B's buffer, and the PDU with SN = 6-9 Is sent, and finally it is sent. PDUs in the UE cannot be forwarded to higher layers until all PDUs have been received in order.

In this case, PDUs with SN = 3 stall forwarding to their higher layers, assuming that all subsequent PDUs (ie, SN = 4-9) have been successfully transmitted. In this example, only 11 PDUs are involved, but in normal operation, hundreds of PDUs may be scheduled before the retransmission data PDU, in which case the transmit latency and data Also note that the buffer problem is exacerbated.

<p><patcit num="1"><text>U.S. Patent Application No. 10/270822</text></patcit></p>

<p> It would be desirable to have a system and method that would allow the retransmission data to avoid delays due to transmission buffer congestion.</p>

<p> The present invention relates to a system and a method for transmitting data in a wireless communication system. Multiple data blocks are received and temporarily stored in the first memory. After that, a plurality of data blocks are transmitted. Next, a determination is made as to whether each transmitted data block has been successfully received, or whether the data block has not been successfully received and therefore needs to be retransmitted. Each of the transmitted data blocks that needs to be retransmitted is labeled and stored in a second memory having a higher priority than the first memory. The labeled data block stored in the second memory is transmitted before transmitting the data block stored in the first memory.</p><p> Each labeled data block can include a common channel priority indicator (CmCH-Pi). The labeled data block CmCH-Pi is read and used to determine in which of a plurality of memories the labeled data block should be stored based on the CmCH-Pi.</p><p> According to a preferred embodiment of the present invention, a wireless communication system for transmitting data includes a UE, a node B communicating with the UE, and a wireless network controller (RNC) communicating with the node B and the UE. Including. The RNC sends multiple blocks of data to the UE via node B. The UE sends a status report to the RNC. This report tells you if each data block transmitted was successfully received in the UE or if the data block was not received successfully in the UE and needs to be retransmitted. The RNC labels each data block that needs to be retransmitted and sends the labeled data block to node B. Node B receives the labeled data block, temporarily stores it, and sends the labeled data block with a higher priority than other previously received and stored data blocks. .. Node B sends the labeled data block to the UE before the other data blocks.</p><p> Understanding the following description, presented only as an example, in conjunction with the accompanying drawings will provide a deeper understanding of the present invention.</p>

<figref num="1">It is a figure which showed MAC-hs of UTRAN.</figref><figref num="2">It is the figure which showed the MAC-hs of the UE of the prior art.</figref><figref num="3">It is a block diagram which showed the data flow between node B and UE.</figref><figref num="4">It is a figure which showed the RLC layer which illustrated the disappearance PDU transmission.</figref><figref num="5">It is the figure which showed the retransmission by the RLC layer of the lost PDU transmission.</figref><figref num="6">It is a signal diagram which showed the method of giving priority to retransmission by this invention.</figref><figref num="7">FIG. 6 is a block diagram showing the data flow between node B and UE, where retransmissions are assigned to a higher priority queue.</figref><figref num="8">It is a block diagram which showed the data flow of the DSCH transmission which schedules a PDU which has a CmCH-Pi display.</figref><figref num="9">It is a figure which showed the retransmission by the RLC layer of the disappearance PDU transmission by this invention.</figref><figref num="10">It is a figure which showed the retransmission by the RLC layer of the disappearance PDU transmission by this invention.</figref>

Preferred embodiments of the present invention will be described with reference to the drawings. The same reference number represents the same element in all drawings.

In describing the present invention, the terms "buffer" and "memory" may be used. These terms have the same meaning and are used to indicate multiple data blocks or PDUs in a continuous queue.

In order to reduce the retransmission latency in the RLC layer, the present invention prioritizes retransmission of PDUs over subsequent PDUs in the buffer of an intermediate node, eg, node B.

In the downlink direction (data transmission from the service serving RNC (serving RNC) (SRNC) to the UE), one of the sources of retransmission latency is generated during the buffering application in the UTRAN outside the SRNC. .. For example, buffering for an application may be done at Controlling RNC (CRNC) or Node B. In some applications, the RLC in the RNC sends a PDU to the MAC-d in the RNC, and the MAC-d in the RNC generates a MAC-d PDU and sends it to the CRNC and then to node B. (Note that if the UE does not move out of the SRNC's cell service area, the CRNC is the same RNC and therefore all messages sent are internal. The UE is out of the SRNC's cell service area. When moving, the new CRNC will be the well-known Drift RNC (DRNC). For brevity, the RNC will be referred to as the CRNC in each case).

A MAC-d PDU can be considered equivalent to an RLC PDU because it contains exactly one RLC PDU (and any other MAC information it may contain). Although the description of PDUs in CRNC or node B in this application is for MAC-d PDUs (rather than RLC PDUs), they can be considered equivalent for the purposes of the present invention and are hereinafter described herein. So, when the term PDU is used, it shall refer to both PDUs.

To allow continuous data flow, PDUs from the RLC in the RNC are usually queued in the CRNC or node B buffer for some time and then sent to the UE and therefore to the peer RLC. Will be done. As described in detail below, the method of the invention that resends data with higher priority bypasses buffering / queuing data in UTRAN.

One embodiment of the present invention is RLC retransmission from a wireless network controller (RNC) to a user device (UE) in a system utilizing High Speed Downlink Packet Access (HSDPA). FIG. 6 shows a method 100 for reducing the retransmission waiting time according to the present invention. FIG. 6 shows the communication between RNC102 and node B104 and UE106.

The RLC layer in UE106 generates a status report PDU that indicates the status of PDU reception (ie, transmission success) or PDU loss (ie, transmission failure) (step 108). This status report PDU is sent to RNC102 (step 110). When the RLC layer in RNC102 receives a status report PDU from a peer entity in UE106, RNC102 prepares to resend the lost PDU (step 112).

The present invention implements a method that allows node B to distinguish a retransmitted PDU from other PDUs. In the first embodiment of the invention, the RNC 102 utilizes a field several bits long in the frame protocol (FP) addition region to label the retransmission PDU. The retransmission PDU includes CmCH-Pi, which is updated (or incremented) each time the PDU is sent from RNC102 to node B104 (step 114). This allows node B104 to know how many times the PDU has been sent and therefore can identify the correct queue in which the PDU is stored. Preferably, CmCH-Pi is generally set and updated in RNC102. However, node B104 can also perform this function. Node B104 reads the CmCH-Pi and determines the correct priority queue for the PDU (step 116). The transmit scheduler on node B104 processes the higher priority queue before the lower priority queue. Node B104 puts the resent PDU into a buffer with a higher priority than it had when the PDU was first sent, as a result of the CmCH-Pi configuration by RNC102.

The PDU in the buffer (ie, memory) that has a higher priority than the first send is then retransmitted (step 118). Other transmissions destined for the same UE can be buffered in the lower priority transmit queue of node B104 when retransmitting the PDU. By increasing and setting the CmCH-Pi of the retransmission PDU, scheduling is performed so that the retransmission PDU is transmitted before the other PDUs received and buffered earlier on node B104.

Referring to Figure 7, retransmissions are placed in a higher priority queue so that other data blocks in the same buffer as their "original" transmit buffer are transmitted before they are transmitted. Assigned. Receiving H-ARQ processor P1_UE~ P5_UEProcesses each data block and then sorts them based on priority buffer R₁, R₂, R₃Transfer to. Each sort buffer corresponds to each priority level of data. For example, sort buffer R₂Is the data block B2₁, B2₂, B2₄Sort. Sort buffer R₃Is the data block B3₃, B3₄, B3₆Sort. One data block (X) is data block B2₂And B2₄Has disappeared between. In addition, another data block (X) is data block B3.₄And B3₆Has disappeared between. That is, the expected data block B2₃And B3₅However, it has not been received, for example, because the NACK message was mistakenly interpreted as an ACK message.

The lost data block is then retransmitted. Normally, data block B2₃Is placed in the priority 2 send buffer. But data block B2₃Is lost and must be retransmitted, so data block B2₃Is put into a higher priority send buffer (in this case, a priority 1 send buffer) and is therefore sent faster than if it was put into a priority 2 or 3 send buffer. Similarly, data block B3₅Is normally placed in the priority 3 transmit buffer. But data block B3₅Has disappeared and must be resent, so data block B3₅Is placed in a priority 1 or 2 transmit buffer, resulting in faster transmission than if it were placed in a priority 3 transmit buffer.

When a PDU is received on node B, it uses CmCH-Pi to prioritize queue B1._n~ B3_nIs determined. The scheduler puts the higher priority queue first and the sender H-ARQ processor P1_B~ P5_BAssign send to. If the transmission to the UE is successful, the receiving H-ARQ processor P1_UE~ P5_UEForwards the retransmission PDU to the RLC layer.

This procedure can also be applied to DSCH systems, with the only difference being that the intermediate node is CRNC instead of node B. Referring to FIG. 8, the PDU 805 with the CmCH-Pi display is prioritized by the prioritizing entity 810 and scheduled for transmission by the MAC-sh in the CRNC. MAC-sh maintains multiple priority queues 815A and 815B, and the DSCH transmission scheduler 820 determines which PDU805 should be transmitted based on the data priority. Therefore, by increasing and setting CmCH-Pi of the retransmission DSCH, those transmissions are processed before other data directed to the UE. This is also the case for HS-DSCH, where the MAC-hs entity on node B schedules transmissions.

Reference to FIG. 9 shows a system according to the invention that implements the prioritization method of FIG. After the RLC layer in the UE sends a status report PDU to the RLC layer in the RNC indicating that the PDU with SN = 3 was not successfully received, the RNC retransmits the PDU with SN = 3. The PDU is given a higher priority than the other PDUs in the intermediate node's buffer by being placed in a higher priority buffer. Note that this example only shows at most 11 PDUs, but in reality hundreds of PDUs may be queued.

The advantages of the present invention can be understood by referring to FIG. 10 showing the results of the prioritization function in the receive buffer. Retransmitted PDUs with SN = 3 arrive in the receive buffer and can in turn transfer PDUs with SN = 3 to 5 to higher layers much faster than the prior art scenario shown in FIG.

Although the present invention has been described by its preferred embodiments, other variations within the scope of the invention as defined in the appended claims will be apparent to those skilled in the art.

102 RNC (Radio Network Controller) 104 node B 106 UE (User Equipment) 805 PDU (protocol data unit) 810 Prioritizing entity 815A, 815B queue 820 DSCH transmission scheduler