hash stringlengths 32 32 | doc_id stringlengths 5 12 | section stringlengths 5 1.47k | content stringlengths 0 6.67M |
|---|---|---|---|
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7.2 Impacts to the PCU | PCU needs to be able to handle double amount of bits per radio slot in uplink. Other impacts to the PCU are minimal e.g. related to the RLC/MAC and resource management. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.7.3 Impacts to the BSS radio network planning | Without any DSR specific RRM optimisation the DSR can be used for frequency reuses up to 1/1. IRC receiver can typically cope with increased UL interference and voice capacity is not decreased e.g. assuming existing networks employ MRC or have sufficient unused gain from IRC as shown in subclause 9.5.5.
Indeed it is po... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.8 Impacts to the core network | No impacts. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.9 Impacts to the specification | The impacted 3GPP specifications are listed in table 130.
Table 130: Impacts to the 3GPP specifications
Specification
Description
3GPP TS 43.064
GPRS Stage 2
3GPP TS 44.018
Radio Resource Control (RRC) protocol
3GPP TS 44.060
Radio Link Control / Medium Access Control (RLC/MAC) protocol
3GPP TS 45.001
Physical layer on... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.10 Possible enhancements | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.10.1 Dual Symbol Rate in downlink | The deployment of DSR in DL as well would need either high diversity terminal penetration or dedicated band and radio resources for DSR users. Both of these are pretty unrealistic in release 7 timeline. Indeed dual carrier offers already similar throughput gain in DL. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.11 Compliance to the objectives | Following tables summarise compliancy to the objectives given in subclauses 4.2 and 4.3.
Table 131: Compliance with performance objectives for DSR
Objective
Required value
Evaluated result for DSR
Compliance
Spectrum efficiency/capacity gain
> 50 %
80 % for data (in BCCH scenario)
Compliant
Peak data rate increase
100 ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 9.12 References | [1] AHGEV-010 "Dual Symbol Rate for GERAN evolution", Nokia.
[2] 3GPP TS 25.101 "User Equipment (UE) radio transmission and reception (FDD) (Release 6)".
[3] GP-052657 "Uplink enhancements for GERAN evolution", Ericsson.
[4] GP-060189: "DSR performance in legacy MRC network", Nokia.
[5] GP-060775: "Modified Dual Symbol... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10 Latency enhancements | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1 Introduction | Four different enhancements are studied and evaluated:
- Improved ACK/NACK reporting.
- Reduced Transmission Time Interval (TTI).
- Variable sized Radio Blocks.
- Hybrid ARQ.
The enhancements reduce overall latency and have a second order effect on mean/average and peak bit rates as reduced latency (i.e. lowering the r... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1 Performance gains | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.1 Web-browsing | Tests have been performed using a "frozen" CNN-test page of a size 137 kB built by 37 objects. Results of "time to download" the web page have been obtained as a function of different latency and bit rate combinations. The values are based on measured results in lab with interpolation of values in-between.
Figure 286 s... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.2 Delay estimations | Figure 289 shows a sequence chart containing GERAN interfaces identifying the interfaces for possible delay reductions in GERAN, namely Abis and Um. The Abis interface is typically not a matter of standardisation but have an important role. Delay estimations for a few cases are shown below for different cases including... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.3 Email | There are number of email protocols defined, including X400, IMAP, and POP (ref [1]). Of these, the widest used are POP and IMAP. This document considers some performance issues of the POP3 protocol. First the use case of a user downloading email headers from a server is considered. Then the use case of a user download... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.3.1 Download of Email Headers | Initially the user is able to download the message headers, in order to quickly decide which emails should be downloaded in their entirety.
Assumptions are as follows:
Item
Value
Note
Coding Scheme on downlink
MCS-8 or MCS-9
permits at least 130 octets/ MAC PDU
Timeslot allocation
4 down, 1 up
E-mail download protocol
... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.3.2 Download of Email Content | After the download of the email headers, the client typically downloads the emails. This may take place in the background, allowing the user to examine the headers for priority. Nevertheless, the faster this can be done the better.
Assumptions are as follows:
Item
Value
Note
Coding Scheme on downlink
MCS-8 or MCS-9
per... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.4 Impact to TCP performance | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.4.1 Introduction | Reducing the latency is not only important for VoIP but also for other services, notably those run over TCP/IP. The purpose of this subclause is to show that reduced latency is also important for FTP and HTTP applications that use TCP. To this end, simulations have been made on a model of the BSS and CN to obtain value... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.4.2 System model and TCP parameters | We begin by describing the simulation network topology and method.
RTT simulations using constant bit rates are performed using a simplified model of GERAN and the CN, see figure 290. We only consider RLC retransmissions, the most important loop.
Figure 290: Simulation network topology
The purpose of the simulations is... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.4.3 FTP performance | First we consider an ftp application. An ftp download of 100 kB application data is performed 10 times. Figure 291 displays the end-2-end average bit rates for EDGE MCS-6. Rates above 90 kbps are obtained for RLC below 100 ms if the Server to BTS delay (one way) is low, but the decay of the bit rate with delay is rathe... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.4.4 HTTP download | Next we consider download of a WEB page (http). The format of the web page is shown in table 142.
Table 142: WEB page
WEB page
Number of objects
8
Packet size [bytes]
1 500
Size object 1 [packets]
41
Size object 2 [packets]
45
Size object 3 [packets]
8
Size object 4 [packets]
7
Size object 5 [packets]
4
Size object 6 [... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.4.5 Measured delays in BSS and CN | The above results have shown the importance of reducing the RLC RTT and Server to BTS latency. The BSS to MS delay is indeed the major contributor to the delay, at least in the current network implementations.
To give an idea about the delay values some measurement data is presented below. The measurements were taken u... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.1.1.4.6 The importance of the PING size | The average end-2-end RTT is important for TCP, but the smaller PING sizes do not show the impact of the RLC RTT for TCP. A 64 byte PING will be contained in a single PDU, whereas a 1 500 Byte PING is segmented into many PDUs where the probability that at least one of them is retransmitted is substantially higher.
To s... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2 Improved ACK/NACK reporting | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1 Concept description | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.1 Event based RLC ACK/NACK reports | Currently, the RLC/MAC ACK/NACK reporting is a time consuming procedure. This is especially true for downlink transfers. The procedure for DL transfers is that the BSS periodically (RRBP) polls the mobile for ACK/NACK reports. Considering the periodicity for the polling, it is realistic that it takes in the order of 15... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.2 Event based RLC ACK/NACK reports | A different approach to the periodic polls from the BSS would be that when the receiver (MS) realises that a RLC data block is missing - from BSN sequence out of order or, in the EGPRS case, when the BSN is successfully decoded but the RLC data is not - it could report this to the BSS without waiting for a regular poll... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.3 ACK/NACK in Uplink Data | There are two ways how ACK/NACK information can be included (piggy-backed) into RLC data blocks:
- Include ACK/NACK in the header of an RLC data block.
- Include the ACK/NACK in the payload part of an RLC data block. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.3.1 ACK/NACK in RLC header | There are spare bits in the RLC/MAC header for EGPRS uplink data blocks. These could be used for ACK/NACK blocks for the DL TBF. The method should be event based as described in subclause 10.2.1.2, which means that the bits shall be used only if there are lost blocks. Since there are very little room for sending BSN, t... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.3.2 Fast Ack/Nack reporting sending Ack/Nack in payload of an RLC data block | Due to the very little room in the header of EGPRS uplink RLC data blocks, an alternative to the solution outlined in the sub-clause above (subclause 10.2.1.3.1) is to reuse part of the payload of an EGPRS uplink RLC data block to convey signalling information. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.3.2.1 Short bitmap in an RLC data block | The principle is that, instead of sending back to the transmitter a long bitmap in a dedicated radio block (i.e. the PDAN), the receiver could send back a very short bitmap, embedded in a normal RLC data block, leaving a lot of space to carry user plane data in the reverse direction.
The short bitmap would contain info... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.3.2.2 Short bitmap in a single burst | A further possibility is the alternative submission of the short bitmap in a single burst (e.g. a newly formatted normal burst), if no other data and/or measurement reports have to be submitted.
As soon as the MS has no additional data payload to submit, e.g. during the Extended UL TBF phase, the MS could send a single... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.3.2.3 Reduced MS reaction time | The other proposal is to reduce the minimum reaction time during the polling procedure, bringing it from 13 TDMA frames to 8 (or 9) TDMA frames, thus gaining 20 ms. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.3.2.4 Co-existence with legacy procedures | In any case the new reporting mechanism based on the short bitmap has to co-exist with the old one. This is needed for instance when several subsequent polling requests/short bitmaps are not received (by the MS or the network). To obtain feedback information from the MS the only possibility in this case would be to ask... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.4 Fast Ack/Nack reporting in UL and DL | Piggybacking Ack/Nack reporting in both Uplink and downlink can be done by locating the Ack/Nack data in the payload. An example is shown in subclause 10.2.1.3.2 for piggybacking in UL. This subchapter describes a method to do it in both UL and DL; including a BSN based short Ack/Nack structure. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.4.1 BSN based short Ack/Nack report | The suggested report has the size of n*3 byte, where n is the minimum number of segments that are needed for a report. Each segment of 3 byte consists of 5 fields, as shown in Table 9. One segment of 3 bytes is normally sufficient if errors are close together and belong to only one TBF, but it can be extended if needed... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.4.2 Ack/Nack reporting sent in UL direction | Subclause 10.2.1.3.2 describes a method how to do fast reporting in the UL direction by using piggybacking if data blocks are to be sent and it reuses the polling mechanisms with a slightly modified header such that legacy operation is still possible.
Legacy polling is still needed in order to support legacy MSs and as... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.4.3 Ack/Nack reporting sent in DL direction | In order for the mobile station to determine that a short Ack/Nack report is included in the data block, the receiver need to know determine this, if possible without any double decoding. A bit in the header would indicate this. A redefinition of the RRBP and ES/P fields in the header of the EGPRS DL data blocks can be... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.1.5 Possible usage | A possible use of the improved ACK/NACK procedures is (considering data transmitted in downlink):
UL TBF exist:
- If UL data is sent use the "ACK/NACK in uplink data" / "Fast Ack/Nack reporting" mechanism.
- If UL data is not sent use the Packet Downlink Ack/Nack message as a response to a USF/ "Fast Ack/Nack repor... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.2 Modelling assumptions and requirements | The simulation environment is shown in figure 311. It is a single user RLC protocol simulator where the LLC packet sizes, MCS, multislot and polling could be set. The radio channel is modelled by a BLER setting that also includes incremental redundancy. The feedback channel is ideal. The radio channel introduces the ra... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.3 Performance characterization | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.3.1 Performance gain of "Event based RLC Ack/Nack reports" | The performance results are shown in figures 312 and 313. The reference case is when polling is used (a poll is executed every 12th RLC block). For each case of the cases, i.e. event based and polling, a poll is also done on the last RLC block in the RLC send buffer.
Figure 312: CDF vs. time to correctly receive a LLC ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.3.2 Performance gain of the "Fast Ack/Nack reporting" mechanism | The following assumption are considered:
- The intrinsic RLC RTT (i.e. the time to cross two times the A-bis and Um interfaces plus the processing delay of network nodes) is assumed to be of 100 ms.
- The considered multislot allocation is 4 (DL) + 1 (UL).
- In case of conventional Ack/Nack reporting procedure, the pol... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.4 Impacts to the mobile station | The mobile station needs be able to schedule and send event based ACK/NACK.
The mobile station needs be able to insert a short bitmap or a short Ack/Nack report in an UL RLC data block when polled accordingly.
The mobile station needs to be able to receive a short Ack/Nack report in a DL RLC data block. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.5 Impacts to the BSS | The PCU needs to be updated to handle new ACK/NACK scheme. Reuse of existing signalling messages except for the case using ACK/NACK in uplink data/"Fast Ack/Nack reporting" mechanism. There is an impact on BTS algorithms to exploit Incremental Redundancy in case of the "Fast Ack/Nack reporting" mechanism.. No impacts t... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.6 Impacts to the Core Network | No impacts. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.7 Impacts to the specifications | The impacted 3GPP specifications are listed in table 154.
Table 154: Impacted 3GPP specifications.
Specification
Description
Comments
3GPP TS 43.064
GPRS Stage 2
3GPP TS 44.060
Radio Link Control / Medium Access Control (RLC/MAC) protocol |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.2.8 Open issues | Definition of the spare bits in the UL RLC/MAC header to report erroneous/missing blocks in DL (including aspects of multiple TBF handling). |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3 Reduced transmission time interval | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1 Concept description | A reduced transmission time interval (TTI) will reduce the Round Trip Time. The present situation (assuming ideal radio conditions) is shown in figure 314 where the delay related to the radio block period of 20 ms is shown. Depending on the MS capability and the radio conditions one or more radio blocks are necessary t... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.1 Radio block mapping in time-slot domain | Figure 315 shows two examples of radio block mapping in time-slot domain.
Figure 315: Different mappings in time-slot domain of four bursts onto a radio block
The mapping of the four bursts to a radio block consists of:
i) two consecutive TDMA frames and two consecutive timeslots (left figure).
ii) one TDMA frame and f... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.2 Radio block mapping in frequency domain (inter-carrier interleaving) | By mapping the four bursts of a radio block onto two consecutive TDMA frames on two separate carriers, the TTI can be reduced to 10 ms - without sacrificing frequency diversity. Similarly, quadruple-carrier EGPRS can be used to reduce the TTI to 5 ms. This is illustrated in figure 316. In order to avoid losing frequenc... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.3 USF scheduling of shorter TTI and legacy mobile stations | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.3.1 Basic principle | In order to support legacy mobile stations the USF structure must remain the same. This means that also the new, shorter TTI mobile stations need to adhere to the basic principle, possibly including some backward compatible additions if needed. However, one simple and backward compatible way is to completely rely on th... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.3.2 Decoding USF in downlink when having both 10 ms and 20 ms TTIs | The issue investigated is if USF and Stealing Flags (SF) can still be used for the different multiplexing possibilities deployed in downlink (i.e. multiplexing of 10 ms and 20 ms TTIs). Tables 156 and 157 give an analysis and solutions. Based on the analysis and solutions it can be concluded that there is no real probl... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.4 Introducing 2-burst radio block option | One of the advantages of reducing the TTI by transmitting 4 bursts in parallel onto 2 different timeslots is that legacy MCSs could be probably maintained, although changes would anyway be needed to the header and the interleaving scheme. Note that the assumption is that the header would be distributed across the 2 tim... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.5 Detailed proposal for a 5 ms TTI solution | The problem of a "5 ms TTI" solution is that it is apparently not possible to have a symmetric solution (i.e. a solution which is valid for both the downlink and the uplink) due to impossibility for (single-carrier) mobile stations to receive and transmit on 4 timeslots in the same TDMA frame.
But receiving and transmi... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.6 Coexistence of legacy and RTTI TBFs (including 4-burst and 2-burst options) | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.6.1 Stealing Flags setting and decoding | For RTTI blocks transmitted with GMSK there is the need to specify how the MS could distinguish them from GPRS Coding Schemes (CS1-4) and EGPRS ones (MCS1-4), and in particular how the MS could distinguish between the 2‑burst or 4-burst format of RTTI blocks.
The solution could reuse the same approach adopted when EGPR... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.6.2 USF setting and decoding | NOTE: This subclause addresses the issue handled in subclause above, but when the 2-burst radio blocks option is considered.
Another problem to solve is to allow legacy USF decoding (which needs 20 ms), in conjunction with fast (10ms) USF scheduling for RTTI TBFs.
The solution could be to have both a normal USF (read a... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.7 Coexistence of legacy and RTTI TBFs (simplified RTTI solution) | The introduction of all the options in 10.3.1.6 would increase the complexity of the RTTI proposal:
- The 5 ms TTI solution would imply a 4 DL + 4 UL timeslots capable mobile station (although it would not be required to transmit AND receive on 4 timeslots during the same TDMA frame, see [6]). Furthermore the 5 ms TTI ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.7.1 The Stealing Flags problem | For RTTI blocks transmitted with GMSK there is the need to specify how the MS could distinguish them from GPRS Coding Schemes (CS1-4) and EGPRS ones (MCS1-4).
The solution is simple and reuses the same approach adopted when EGPRS was introduced: for MCS1-4 all the legacy stealing flags are set to indicate CS4, while fo... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.1.7.2 The USF decoding problem | Another problem to solve is to allow legacy USF decoding (which needs 20 ms), in conjunction with the need to schedule RTTI blocks in the UL every 10 ms.
A simple solution - already included in the Feasibility Study on GERAN Evolution - is to specify that MS in RTTI TBF mode will read:
- the USF allowing UL transmissio... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2 Link level performance | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.1 Modelling Assumptions and Requirements | The following EGPRS services have been simulated for the timeslot mapping option; MCS-1, MCS-4, MCS-5 and MCS-9. Each simulation point has been run with 10 000 radio blocks (40 000 bursts).
Simulated radio scenarios:
- Sensitivity: TU50iFH.
- Co-channel interference: TU3iFH and TU3noFH.
Other simulator settings:
- Blin... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.2 Performance Characterization | The link level performance for case i, ii, iii and iv (as described in subclause 10.3.2.1) is summarised in tables 160 and 161. Reference performance with the regular radio block mapping is also shown. Detailed results can be found in Annex A. BLER means the total BLER, i.e. both header BLER and data BLER are considere... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.3 Conversational services with reduced TTI | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.3.1 Introduction | This subclause aims at evaluating some worst case scenarios compared to TU3iFH. From the link level results some conclusions are drawn on the effects on application level.
The losses at more severe radio conditions are relevant regardless of the service. However, for other services than VoIP, the reference is legacy EG... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.3.2 Application Level Effects | In the previous Ericsson contributions a reduced transmission time interval (RTTI) of 10 ms was evaluated where the reduction in transmission time was achieved in the downlink by dual carrier and in the uplink by dual time slot transmission. Therefore, the link level simulations in this contribution have both been eval... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.3.3 Simulator settings | A state-of-the art link level simulator has been used for the evaluation of the different channel models. The parameter settings are shown in table 162. Thus, the frequency band is kept at 900 MHz for all simulations. However, since RA250noFH is simulated, the faster fading experienced at higher frequency bands is assu... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.3.4 Simulation results | Simulations have been performed on different channel profiles with different speeds depending on the channel used. For the evaluation of each channel model, TU3iFH is used as reference.
RA250noFH
In figure 323 the performance of MCS-2 on Rural Area channel at terminal speed 250 km/h is shown when radio block format 1 i... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.4 Reduced TTI and fast ACK/NACK | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.4.1 Definition of the new coding schemes | New RTTI coding schemes are defined as shown in table 165.
Table 165: Definition of new coding schemes
RTTI
2 w/o bitmap
RTTI
2 w/ bitmap
RTTI
3 w/o bitmap
RTTI
3 w/ bitmap
RTTI
5 w/o bitmap
RTTI
5 w/ bitmap
RTTI
6 w/o bitmap
RTTI
6 w/ bitmap
Raw
Header
31
31
31
31
37
37
37
37
Bitmap
0
20
0
20
0
20
0
20
Data
226
194
29... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.4.2 Header coding | The coding of the header is kept unchanged. Only the burst mapping changes in order to allow reduced TTI operation. In other words, the header bits are mapped on to four bursts and the first two bursts are transmitted in the same TDMA frame in a multislot transmission and the other two bursts are transmitted in the nex... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.4.3 Bitmap coding | A constant number of bits (20) per block are assumed as the payload size of the bitmap. The coding of the bitmap is done independently. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.4.4 Block code | A 6 bit CRC is added to the 20 bitmap bits. The method used for generating the CRC bits of TCH/AFS logical channels is reused for this purpose. The result is a block of 26 bits. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.4.5 Convolutional code | Tail biting convolutional code of 1/3 rate is used to encode the resultant 26 bits (from the block coding mentioned above) the code polynomials used for MCS-1 header coding are used for this purpose. The result is a block of 78 bits. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.4.6 Puncturing | For RTTI 2 and RTTI 3 type logical channels, the convolutionally coded bitmap data bits are punctured from 78 bits to 54 bits. A pseudo random uniform puncturing pattern is used for this purpose. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.4.7 Interleaving and burst mapping | The bitmap is interleaved as if it is part of data. The method is described in figure 326.
Figure 326: Interleaving of RTTI type data |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.4.8 Data coding | The size of the payload for each RTTI scheme is as shown in table 165. The basic coding scheme of the data is kept unchanged. (i.e. block coding, convolutional coding etc). However the algorithms operate on blocks of different sizes. The puncturing is done again using a pseudo random uniform puncturing pattern. The int... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.2.4.9 Simulation results | Simulations are done for RTTI 3 and RTTI 6 type coding schemes for TU3 and TU 50 ideal frequency hopping channels for low band (GSM 800) and also for TU 50 in upper band (GSM 1800/1900). The results are shown in this subclause. The RTTI type coding schemes are assumed to always have the bitmap for first and second tran... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3 Application performance | This subclause makes an evaluation of five traffic cases with use of a simulator:
1. End-to-end latency with Ping.
2. Conversational service over packet data (VoIP).
3. E-mail receiving and sending, using POP3 and SMTP respectively.
4. E-mail synchronization, using IMAP.
5. Web page download, using HTTP/1.1 without Pip... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.1 Modelling Assumptions and Requirements | Table 166 gives the simulation assumptions. Any further assumptions are defined in direct relation to the results in next subclause.
Table 166: Simulation parameters used in the simulations
Type
Value
Comment
Radio Conditions
TU3iFH, C/I 9dB and 15dB
-
RLC re-transmission scheme
RLC acknowledged mode
Unlimited number o... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2 Performance Characterization | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.1 Single-user cases | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.1.1 Introduction | Single-user cases have been analysed with simulations for Ping, VoIP and e-mail sending and receiving. Single-user means, in this context, that there is only one mobile reserved per packet channel. Thus, more users can be served by one base station or cell (and even TRX) and the results in this subclause are still appl... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.1.2 Ping | End-to-end latency is typically benchmarked with a Ping traffic case. The default Ping size is 32 bytes which typically ends up with 70 bytes of RLC data: 32 bytes ICMP payload, 8 bytes ICMP header, 20 bytes IP header and 10 bytes LLC/SNDCP header. 70 bytes is valid both for request and response. The simulations perfor... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.1.3 PS Conversation Service, VoIP | A conversational service over GERAN packet data is analysed in this subclause. The scenario analysed is a mobile to mobile conversation, and the results are presented in mouth-to-ear delay for the different scenarios.
The VoIP scenarios assume 2 AMR 7.95 frames per IP packet, which correspond to 40 ms of speech per IP ... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.1.4 E-mail | An e-mail service is analysed in this subclause. The scenarios analysed are sending of an e-mail using the SMTP protocol and reception of an e-mail using the POP3 protocol.
The simulation parameters used in the e-mail simulations are listed in table 166. |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.1.4.1 E-mail sending, SMTP | The scenario analysed is sending of an e-mail using the SMTP protocol. A flow-graph of a typical SMTP scenario is described in annex B.
The results, expressed as session time, from the simulations are summarized in tables 170 and 171.
Table 170: Result for the single-user SMTP simulations (milliseconds)
E-mail of 5 kby... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.1.4.2 E-mail receiving, POP3 | The scenario analysed is a reception of an e-mail using the POP3 protocol. A flow-graph of a typical POP3 scenario is described in annex B.
The results, expressed as session time, from the simulations are summarized in tables 172 and 173.
Table 172: Result for the single-user POP3 simulations (milliseconds) - E-mail o... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.1.4.3 E-mail synchronization using IMAP | The scenario analyzed is a synchronization of an e-mail inbox using the IMAP protocol. The e-mail account consists of a number of folders, where only the Inbox is synchronized. The scenario starts with the e-mail client already running and when establishing a TCP connection. Then the e-mail client performs a login to t... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.1.5 Web service | A large web page, consisting of 53 objects and 289 kbytes in total, was used in this analysis. HTTP/1.1 without Pipelining, as used for example by Microsoft Internet Explorer, was used for the analysis. Two parallel TCP connections are used to download the objects.
The results, expressed as session time, from the simul... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.2 Multiple-user cases | |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.2.1 Introduction | Multiple-user cases have been analysed with simulations for Ping, VoIP and Web download. Multiple-user means, in this context, that there is more than one mobile reserved per packet channel. Thus, the results in this subclause are applicable when packet channel sharing applies. The purposes with the multi-user cases ar... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.2.2 Ping | The same Ping case is used as described in subclause 10.3.3.2.1.2. The reference mobile is performing Pings, whereas the other mobiles perform a constant UDP flow in both UL and DL. A "round-robin" MAC scheduling is applied to all mobiles, which means that all mobiles have the same priority.
The simulation parameters u... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.2.3 PS Conversation Service, VoIP | The same VoIP case is used as described in chapter 10.3.3.2.1.3. There are two reference mobiles performing a mobile‑to-mobile VoIP conversation, whereas the other mobiles perform a constant UDP flow in both UL and DL. A "QoS‑based" MAC scheduling is applied to all mobiles. The reference mobiles have an absolute priori... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.2.4 Web donwload | The same Web download case is used as described in subclause 10.3.3.2.1.5. The reference mobile is performing Web download, whereas the other mobiles perform a constant UDP flow in both UL and DL. A "round-robin" MAC scheduling is applied to all mobiles, which means that all mobiles have the same priority.
The simulati... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.3 Summary of Results | This subclause summarizes the gains achieved with the RTTI cases compared to the legacy cases. In table 181, the gains are shown for the single-user cases and table 182 shows the gains for multiple-user cases.
Table 181: Summary of results, single-user cases
Application
Relative Gain
(legacy case vs RTTI case)
Absolute... |
5692480e365e21a881d790cc4eaea394 | 45.912 | 10.3.3.2.4 Conclusions |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.