NSN Lte-Advanced White Paper
NSN Lte-Advanced White Paper
NSN Lte-Advanced White Paper
CONTENTS 1. Overview 1.1 LTE-Advanced, the evolution of LTE 1.2 Status of LTE-A (as of October 2013) 3 3 3 4 5 5 5 8 11 13 14 16 16 19
2. Drivers 3. The LTE-A toolbox 3.1 Overview 3.2 Carrier Aggregation 3.3 Advanced MIMO schemes 3.4 Coordinated multipoint transmission and reception 3.5 Relay Nodes 3.6 Heterogeneous Networks 3.7 Self Organizing Network and network architecture evolution with LTE-A 3.8 Outlook
4. Summary
Page 2
nsn.com
1.Overview
1.1 LTE-Advanced, the evolution of LTE
The introduction of LTE was driven by the industrys quest for a more efficient technology that could help deliver ever faster mobile broadband services. In comparison with basic HSPA networks, LTE delivered this enhancement by offering the state of the art combination of new air interface base technology (OFDMA/SC-FDMA), greater flexibility for utilizing spectrum like for example support of 20MHz bands and TD-LTE for using unpaired spectrum, as well as a toolbox to support further enhancements like MIMO and Higher Order Modulation. In fact, a similar toolbox has been applied to HSPA for facilitating a seamless evolution to HSPA+. At the same time, we continue to witness exponential growth in mobile broadband traffic; thereby necessitating further enhancement in the overall efficiency, with a view to deliver faster mobile broadband services to a constantly increasing user base. LTE-Advanced abbreviated as LTE-A has primarily been conceptualized to address both the aforementioned demands. LTE-A marks the evolution of LTE; it continues to deploy the air interface base technology of LTE which provides highest efficiency and a smooth evolution in the deployment of the existing LTE ecosystem towards LTE-A. It allows operators to deploy larger bands than 20MHz in particular by carrier aggregation, while also enabling an advanced toolbox with advanced MIMO schemes and totally new features like Relaying. Moreover, it is fully backwards compatible with the earlier LTE releases, implying that legacy devices can operate in LTE-A networks but may not necessarily benefit from all the new features of LTE-A. Thanks to these advanced features, LTE makes its transition to a true 4G technology, in accordance with the requirements of ITU for IMTAdvanced. This paper introduces the advanced toolbox of LTE-A, including information on new technologies, features and enhancements to existing technologies, as well as discusses the benefits that LTE-A provides to operators and end-users.
Page 3
nsn.com
One of the main drivers of the technical enhancements and timetable for LTE-A development has been IMT-Advanced. ITU initiated the IMTAdvanced process to define the requirements for the next generation of Radio Interface Technologies (RIT) that were released in a circular letter in early 2008. Meeting the IMT-Advanced requirements has been the goal that 3GPP has to achieve and standardization in 3GPP has progressed well. The first LTE-Advanced specifications have been frozen in the first half of 2011 while evaluations conducted by 3GPP contributors and external parties have demonstrated that LTE-A meets all the IMT-Advanced requirements. As a consequence, ITU-R has already approved LTE-Advanced as IMT-Advanced RIT or true 4G system in November 2010. The first commercial LTE-A networks have been launched by SK Telecom, LG U+ and KT in Korea during summer 2013. All three operators use carrier aggregation with NSN as a supplier. In summer 2012, the major requirements for the evolution of LTE-A with 3GPP Rel.12 were collected in a 3GPP workshop. As of October 2013, the release content of 3GPP Rel.12 has been refined, but not yet formally frozen. The evolution of LTE-A with Rel.12 and beyond is subject of another NSN whitepaper.
2.Drivers
Looking ahead, the exponential growth in data traffic is expected to continue on the same lines owing to certain key drivers: Increased adoption of mobile broadband Enhanced coverage (spreading across more locations) Increase in usage intensity Greater availability & choices in terms of devices (smart devices, phones, pads, booklets, netbooks...) Machine-to-machine communications stepping alongside human users
VEHECILE
Mobility
WALKED
UMTS A detailed analysis reveals that data traffic is distributed in an uneven way. Eventually mobile broadband networks need to evolve in a 1 10 100 1 manner which goes beyond the conventional approach of applying one Mbps Mbps Mbps Gbps standard remedy to the capacity squeeze. Also, the laws of physics Figure 1: Evolution of data imply that conventional mobile broadband networks are approaching speeds for stationary and mobile the theoretically achievable spectral efficiency, which in turn implies use cases the costs per bit/Hz. Consequently the need for higher bandwidths and higher efficiency can only be answered by combining several tools optimized for specific network scenarios.
This is the prime reason for using the term toolbox in this paper. LTE-A defines a large set of tools focused on enhancing the mobile broadband user experience, as well as reducing the costs per bit.
FIXED
Page 4
nsn.com
Page 5
nsn.com
can be arranged between local area nodes to provide better inter-cell interference coordination.
LTE-Advanced
Enhance macro network performance
Capacity and cell edge performance enhancements by active interference cancelation Peak data rate scaling with antenna paths for urban grid and small cells Peak data rate and throughput scaling with aggregated bandwidth
8x Heterogeneous networks
Relaying
Coordinated Multipoint
MIMO
4x
Carrier Aggregation
up to 100 MHz
Carrier1 Carrier2 Carrier3 Carrier4
Figure 2: LTE-A support both: enhancing the LTE macro network and enabling the efficient introduction of small cells
40 30 20 10 0 0 5 10
Uplink
LTE Rel-8
LTE Rel-10
3GPP macro #1
with 2x20MHz for DL, 1x2 for UL
2x2 SU-MIMO
25
30
Dynamic trac
with Poisson arrival and nite buer
Downlink
Rel-8 UE case
(one CC per UE) and LTE-A UE case (2-CCs per UE) back-o assumed for UL with Tx on two CCs
1 dB power
60
70
Figure 3: Carrier Aggregation improves average cell throughput both in uplink and downlink due to more efficient utilization of radio resources, i.e. by statistical multiplexing
Carrier Aggregation supports cross component carrier scheduling, implying that the control channel at one carrier can be used to allocate resources at another carrier for user data transmission. It can be used to provide both frequency diversity and interference coordination in frequency domain at the same time, underlining its
Page 6
nsn.com
significance as a powerful technology for effective utilization of radio resources. Carrier Aggregations capability to improve single user throughput depends on the number of users in a cell. The number of users is directly proportional to the overall statistical multiplexing gain even on a single carrier, so scheduling high number of users over multiple carriers provides only marginal gain. However, if the number of users is low, scheduling over multiple carriers provides significant throughput gain since all radio resources can be allocated to the user(s) with the most favorable radio conditions. Gain in uplink is lower than in downlink, since the UE can not always utilize multi-carrier transmission due to limited transmit power. If carriers are at different frequency bands they have different propagation losses and different interfering systems which affect achievable data rates, transmit power and usage of resources, e.g. far-off UE could be better served with a low frequency carrier and near cell center UE with a high frequency carrier. Inter-band Carrier Aggregation provides more flexibility to utilize fragmented spectrum allocations but one must take UE capabilities into account. There must be enough (but not all) inter-band capable UE before the feature can improve network performance. Studies have been conducted on the benefits of extension carriers, e.g. without common control channels, to have lower control channel overhead and better efficiency, but the improvements seem quite marginal for the scenarios evaluated in LTE Release 10. Future releases might include extension carrier for specific use cases, e.g. energy efficient machine-to-machine communication.
Page 7
nsn.com
80 70
30
60 50 40 30 20 10 0
Coverage [Mbps]
Marginal Avg. TP loss in Rel8 by assigning cell center UEs on 2.1GHz carrier
25
20
15
10
10
20
30
40
50
10
20
30
40
50
Figure 4: Inter-band Carrier Aggregation enables to benefit from different propagation characteristic of different frequency bands
# of UE antennas 8 Downlink [Mbps] Uplink [Mbps] 1102
2 1
- 2x20 MHz Carrier Aggregation and 64QAM with 9/10 code rate
Figure 5: Carrier Aggregation and MIMO provide high peak data rates bounded by allocated bandwidth and the number of transmit and receiver antennas
nsn.com
transmissions are for a single user, then the technology is called Single-User MIMO (SU-MIMO), for multiple users Multi-User MIMO (MUMIMO). The better the system can utilize these communication channels for multiple transmissions, the higher is the capacity that the system can provide. MIMO performance is subject to a large number of parameters: the number of transmitter and receiver antennas, reference signals and algorithms for channel estimation, feedback of channel estimation data from the receiver to the transmitter and spatial encoding methods. Consequently a comprehensive design is crucial to provide optimum system performance. Transmission peak date rates depend on the number of antennas on the transmitter and the receiver, the used bandwidth and the configuration of radio parameters like the resource allocation for control channels. The maximum peak data rates vs. the number of transmitter and receiver antennas can be seen in Fig. 5 for 40 MHz band allocation for both, the downlink and the uplink. LTE Releases 8 and 9 support multi-antenna (MIMO) technology with up to four transmit and receiver antennas in downlink, but only single antenna transmission in uplink. Release 10 extends the MIMO support for eight transmit and receiver antennas in downlink and introduces uplink MIMO by supporting up to four transmit and eight receiver antennas.
Capacity - Correlated
4.00
Average SE [bps/Hz/cell]
Ideal MMSE/SIC
Realistic MMSE/SIC
3.00
2.44 2.02 2.17 1.57
2.75 2.09
100%
108%
1.77
121%
136%
2.30
100%
113%
133%
147%
1x4 no MIMO
2x4 MU dual
Coverage - Correlated
Cell-edge user SE [bps/Hz/user]
Ideal MMSE/SIC
Realistic MMSE/SIC
0.125 0.100 0.075 0.050 0.025 0.000 1x4 no MIMO 1x4 MU-MIMO Rel-8 1x4 MU-MIMO Rel-10 2x4 MU dual
0.075 0.080 0.049 0.084 0.055 0.088 0.060
100%
107%
112%
117%
0.060
100%
113%
123%
122%
Figure 6: an example how uplink MU-MIMO improves system performance with different TX/ RX antennaconfigurations
Ideal: Perfect knowledge of interference assumed at the receiver Realistic: Only have estimate of interference power available at the receiver.
Page 9
nsn.com
Release 10 has enhanced the reference signal design with user specific reference symbols for signal demodulation and common reference symbols for feedback purposes in downlink and more orthogonal reference signal structure in uplink. The enhanced design enables better performance when the number of antenna branches is high. Uplink MIMO provides significantly higher peak rates and improved spectrum efficiency in uplink direction. SU-MIMO provides mainly increased data rates in lightly loaded networks for high-end multitransmitter UE, whereas MU-MIMO can offer significant improvement of spectrum efficiency even with single transmitter UE. This can boost network capacity at low costs and is depicted in Fig. 6 and 7. The LTE-A system can operate in both SU and MU-MIMO modes at the same time using dynamic user specific MIMO transmission configuration.
Capacity - Correlated
Average SE [bps/Hz/cell] 4.00 3.00 2.00
1.25 2.02 1.45 1.06 100% 1.35 162% 1.57 148% 2.25 180% 1.84 173%
Ideal MMSE/SIC
Realistic MMSE/SIC
1.00 0.00
100%
116% 127%
1x2 no MIMO
2x2 SU-MIMO
1x4 no MIMO
2x4 SU-MIMO
4x4 SU-MIMO
Capacity - Correlated
Cell-edge user SE [bps/Hz/user]
Ideal MMSE/SIC
Realistic MMSE/SIC
0.088 199% 0.064 190%
0.100 0.075
0.055 0.075 169% 0.045 133% 0.049 146% 0.081 183% 0.052 155%
123%
1x2 no MIMO
2x2 SU-MIMO
1x4 no MIMO
2x4 SU-MIMO
4x4 SU-MIMO
Figure 7: an example how uplink SU-MIMO improves system performance with different TX/ RX antennaconfigurations
Ideal: Perfect knowledge of interference assumed at the receiver Realistic: Only have estimate of interference power available at the receiver.
Downlink MIMO has already been included in LTE Release 8. The LTE Release 8 codebook and reference symbol design was found to be quite optimum for two and four transmit antennas (2x2, 2x4 and 4x4 antenna configurations), but the channel state information feedback from UE to eNB could have been more accurate. This limitation is overcome by the new reference symbol design of Release 10, which is also more effective when the number of transmit antennas is higher. Based on the studies and numerous contributions in 3GPP,
Page 10
nsn.com
it can be safely concluded that the higher the number of antennas, the higher is the gain that Release 10 MIMO provides in downlink. With two eNB and two UE antennas, Release 10 downlink MIMO provides no improvements over Release 8 in SU-MIMO mode but small performance improvements have been gained in MU-MIMO mode. In most cases it is best to operate two TX antenna eNBs in Release 8 SU-MIMO mode. When eNB has four transmit antennas, Release 10 downlink MIMO gain is more than 20% over Release 8 and with eight transmit antennas a bit higher. Reference symbol overhead effects on system performance are significant with four and eight transmit antennas. Therefore the selection of MIMO operating modes and system parameters for both Release 8 and 10 UE is a critical network optimization task. An important point worth remembering is that the network should also support Release 8 and 9 UE which does not benefit from the Release 10 enhancements. The capacity gain from Release 10 downlink MIMO enhancements could even be negative since new reference symbols create overhead for all UE. However, these overheads can be decreased by decreasing the Release 8 and 9 specific reference symbols, but this would prevent non-LTE-A UE to operate in MIMO mode and thus lower their data rates. Additionally, there would be negative effects on common control channel performance. Consequently, the timing of the introduction of the new features and the configuration of the system parameters are essential for an optimum performance of the LTE network.
Page 11
nsn.com
Downlink CoMP Category Cell Avg. Cell Edge Intra-site 2TX (4TX) 5% (10%) 12% (20%) Intra & Inter-site 20% 21% Intra-site joint reception 5% 5%
System simulations Downlink with ideal CSI feedback, realistic CQI feedback, realistic reference symbol overhead (10%) and ideal inter-cell communication Uplink with ideal feedback, ideal inter-cell communication, ideal cell selection, realistic MMSE/SIC receiver and realistic closed loop power control 2 RX and 2 TX antennas in eNB 2 RX and 1 TX antennas in UE Gain over Release 8 Single User MIMO Typical Urban Micro, max. 500 m inter-cell distance, 10 users per cell
Figure 8: JP/JT CoMP system performance gain in an urban environment with ideal CSI feedback and realistic system andreceiver implementation The system performance gains of realistic CoMP deployments with an ideal channel state information (CSI) feedback is presented in Fig. 8 and 9. The critical system deployment issue is the communication between the cells. Intra-site CoMP deployment, in which the communication is between the sectors of a single eNB, is likely the most feasible system solution. CoMP studies in 3GPP continue in a Release 11 study item kicked off in December 2010 and will focus on finding practical concepts with real performance benefits, taking into account implementation and interoperability issues of UE, eNBs and transport technologies.
Downlink CoMP Category Cell Avg. Cell Edge Intra-site CS/CB 13% 13% Inter-site CS/CB 14% 13%
System simulations Downlink with ideal CSI feedback, realistic CQI feedback, realistic reference symbol overhead (10%), ideal inter-cell communication and MRC receiver Uplink with ideal feedback, ideal inter-cell communications, ideal cell selection, realistic MMSE/SIC receiver and realistic closed loop power control 4 RX and 4 TX antennas in eNB with /2 antenna spacing 2 RX and 1 TX antennas in UE Gain over Release 8 Beamforming (1 CRS, 1 DRS, single stream) 3GPP Case 1 3D, 500 m inter-cell distance, 10 users per cell
Figure 9: CS/CB CoMP system performance gain in an urban environment with ideal CSI feedback and realistic system and receiver implementation
Page 12
nsn.com
600% 500% 400% 300% 200% 100% 0% ISD 500m 1 relay 4 relays 10 relays
-100%
downlink
ISD 1732m
ISD 500m
ISD 1732m
uplink
120% 100% 80% 60% 40% 20% 0% ISD 500m ISD 1732m ISD 500m ISD 1732m
downlink
uplink
Figure 10: System performance gain of Relay Node deployment of one, four and ten Relay Nodes per a macro-cell
LTE Release 8 supports simple amplify and forward relays (also called repeaters) that can be used for coverage extension. However, those do not use the radio resources efficiently. The enhanced relaying technology in LTE-A is based on self-backhauling base stations sharing features with (pico) base stations. For the user equipment the relay node is just a cell of the Donor eNB. The management of the network is straightforward. LTE Release 10 specifies a new interface Un between Donor eNB and Relay Node (RN), see Fig. 11. The new interface uses MBSFN (Multicast-Broadcast Single Frequency Network) subframes which were introduced in Release 8 already to hide the Un interface from UE operating on the same carrier and thus make it fully backward compatible: UE interprets Un transmission as MBSFN transmission for which they are not subscribed and simply ignore them. The so called Proxy S1/X2 concept forwards both S1 and X2
Page 13
nsn.com
messages towards the RN transparently for the Core Network which sees a Relay Node as a sector of the Donor eNB as well. Thus relaying is also backwards compatible for both the MME and Serving Gateway which serve the UE.
App. TCP/UDP IP GTP-u UDP IP PDCP RLC MAC PHY User Equipment PDCP RLC MAC PHY Relay PDCP RLC MAC PHY GTP-u UDP IP PDCP RLC MAC PHY Donor eNB GTP-u UDP IP L2 L1 IP GTP-u UDP IP L2 L1
(serving the UE)
S-GW/P-GW
Page 14
nsn.com
Share of sites growing 100 300 m 1 5W Share will grow in future 10 100 m, < 500 mW License exempt growing & Secondary services emerging
Local area
WLAN
Local area
WLAN WLAN
WLAN
Local area base stations and access points are deployed and in many cases operated by end users directly without network planning by an operator. These local area nodes create interference with each other and wide area base stations may also translate into degraded system performance like lower throughput and an increase of call drops. As such, automated management methods are required to remove the need for manual maintenance of a large number of local area base stations, as well as to prevent excessive inter-cell interference that could degrade the performance of the wide area base stations and other local area nodes. The evaluation cases for heterogeneous network deployments have been included in LTE Release 10. There are multiple technologies that can be used for the interference coordination based on LTE Release 8 specification, e.g. HeNB power control and escape carrier or using Carrier Aggregation of LTE Release 10. LTE Release 10 includes one new interference coordination technology based on coordinated muting of the Transmission of overlapping cells. This technology is called TDM eICIC (Time Domain enhanced Inter-Cell Interference Coordination) and its basic principle is described in Fig. 13. Part of the transmitted signal is muted by sending Almost Blank Sub-frames, that allows other eNBs to transmit with lower inter-cell interference. TDM eICIC needs time synchronization between the macro and femto layers, a pre-condition that could be difficult to guarantee with respect to HeNBs deployed by the users. Simpler frequency domain methods are then more likely to be used in case the operators frequency and deployment plans allow. Later releases are likely to introduce new cost efficient small cell interference coordination and rejection technologies, since cost effective small cell deployment offers the most promising way to increase the capacity of mobile broadband networks in a focused way.
Page 15
nsn.com
3.7 Self Organizing Network and network architecture evolution with LTE-A
LTE development is not only focusing on air interface performance enhancements. Cost of deployment and operation can be decreased with self organizing and optimization (SON) technologies. Automatic Neighbour Relation (ANR) and Minimization Drive Test (MDT) technologies have been developed to enable automatic configuration, optimization of handovers, as well as other radio resource management parameters. Moreover, other SON technologies are also in the process of being developed, e.g. for automated fault recovery and energy saving for complex deployments.
One sub-frame
Macro-layer
HeNB-layer
Figure 13: Inter-cell interference reduction with Almost blank sub-frames of TDM eICIC Some deployment concepts and network architectures are common for HSPA and LTE: Home base stations are a way to provide reliable and secure mobile broadband services in home and office environments. Local Break Out solutions (LIPA and SIPTO) decrease cost of transport and enable lower end-to-end latency for distributed services. Given the fact that a majority of mobile broadband networks fall under the domain of multi-radio networks, common solutions for HSPA+ and LTE-Advanced translate into lower cost for operators and seamless service experience for end users.
3.8Outlook
Development of LTE-Advanced will continue in future 3GPP releases. Multi-hop and moving relays could increase efficiency in providing broadband services in high-speed trains and interference cancellation receivers will improve air interface capacity. Decreasing power consumption of the network and the user equipment enables the usage of battery powered devices for machine-to-machine applications wide bandwidth demand. LTE-A already has means for flexible spectrum management, self-configuration and multilayer deployments. Once the spectrum regulation defines the framework for usage of cognitive radio resource management methods, adoption of these methods can be easily adopted in LTE-A.
Page 16
nsn.com
In real-world network deployments, the described LTE-A system features are closely related to network element implementation for the complete base station sites, including transport. Without a compact multi-antenna site solution, multiple antenna system technologies cannot be cost effective. Multi-site CoMP technologies need fast connectivity between base stations and remote radio heads which can be provided by modern optical transport solutions and open interface specifications. Carrier aggregation provides higher system bandwidths which need wide bandwidth high efficiency power amplifiers in base stations and terminals. There are various multisystem multi-band combinations which need tight control of spurious emissions and good receiver blocking performance. The new features of LTE-A increase spectrum efficiency and cell edge performance, thus the bits per Hz ratio increases. This means that the probability of multi-stream transmission, higher order modulation and lower coding rates increases, with the consequence that the modulation accuracy of the transmitters also needs to improve in order to have sufficiently low inter-symbol interference. Power amplifiers, duplex filters, transmitters analogue and digital parts etc. have to be in a good balance and tightly integrated. One example of these dependencies is the case where the power amplifier of the eNB gets into saturation with the consequence that the quality of signal deteriorates to an extent that higher order modulation can not be supported anymore. In that case the introduction of LTE-A features would not provide the expected system performance gain. Such balancing considerations are most relevant in the hot zones of the network, where additional bandwidth will be needed first. Many of the new technologies introduced by LTE-A are based on complex algorithms, so more baseband processing capacity is needed in both the base stations and the terminals. Fig. 14 summarizes relations between the evolution of implementation technologies and LTE-A system technologies. In fact, suppliers of LTE-A networks need to develop a core competency in terms of integrating a variety of products to support multiple modes to deploy the network. Network operation should be reliable and cost efficient, while maintaining optimum levels of customer satisfaction. In a majority of cases, there are other wireless and cellular technologies to inter-work and co-exist with. Therefore the supplier needs to have a good understanding of LTE-A, its preceding technologies, devices, services, along with end-user behaviour and expectations.
Page 17
nsn.com
More spectrum
Baseband processing capability Multiple power ampliers in UE Multi-antenna BTS site Low cost small BTS
Relays
LTE-A
LTE-A provides a powerful and versatile toolbox, which helps network operators to differentiate in mobile broadband user experience and to increase network efficiency. As the interdependencies between the tools and the network implementation are complex, an experienced partner with a holistic view is needed to make the most of this toolbox. The text box above shows NSNs long track record in LTE-A research. Combined with its leading role in commercializing LTE, this makes NSN the partner of choice when planning and implementing LTE-A.
Page 18
nsn.com
Spectrum eciency
Multi-Layer
Multi-Layer
SON
Spectrum utilization
4.Summary
LTE-A enables a smooth and backward compatible evolution of LTE and TD-LTE towards true 4G performance LTE-A comprises of various tools to enhance mobile broadband user experience and network efficiency There are serious interdependencies between network implementation and the various tools of LTE-A, which require an experienced partner when planning and implementing LTE-A NSN has always been at the forefront of LTE-A research and development, with a strong focus on real operator opportunities in terms of efficiency and user experience
Peak rate
++ ++ (o) o o o
+ o ++ + ++
= clear gain = moderate gain
* without increasing the number of antennas ** not in LTE Release 10 *** with multiple Relay Nodes per cell
Page 19
nsn.com
Nokia Solutions and Networks P.O. Box 1 FI-02022 Finland Visiting address: Karaportti 3, ESPOO, Finland Switchboard +358 71 400 4000 Product code C401-00833-WP-201310-1-EN 82013 Nokia Solutions and Networks. All rights reserved.
Public NSN is a trademark of Nokia Solutions and Networks. Nokia is a registered trademark of Nokia Corporation. Other product names mentioned in this document may be trademarks of their respective owners, and they are mentioned for identification purposes only.
nsn.com