LTE Market, Technology, Products: DR Nick Johnson, CTO, Issue 1.1, Ip - Access, February 2012

LTE Market, Technology, Products
Dr Nick Johnson, CTO, Issue 1.1, ip.access, February 2012
EXECUTIVE SUMMARY Background The Problem The Solution The ip.access Market Vision INTRODUCTION Market Overview Global Observations North America Japan & Korea BRIC Developing World Europe SOLUTION OVERVIEW Summarising the Benefits Solution Architecture PRODUCT FOCUS The E-100 LTE Access Point Gateway Network Orchestration System TECHNOLOGY TOPICS The Multi-Standard Basestation Inter-RAT Mobility WiFi integration and IP Flow Mobility Content Aware Scheduling Network Offload (LIPA, SIPTO and LIMONET) The Importance of FAPI High Performance, Low Cost, Low Power Consumption Baseband Silicon Wide Band, Technology Agnostic RF Silicon Self-Organisation, Self-Optimisation IMS access SUMMARY
3 3 3 3 3 5 5 5 6 7 8 8 8 9 9 10 18 18 20 20 21 21 21 22 23 25 26 26 26 27 27 29
-2-
ip.access 2012
Executive Summary
Background ip.access is the world leader in W-CDMA femtocell technology, and has been serving the indoor small cell market since 2000. Its GSM picocell technology (trademarked nanoGSM) and corresponding 3G technology (nano3G) are accepted and operating in over 60 mobile network operators worldwide. The Problem As the world enters the 4G age, a series of new challenges and opportunities present themselves. The demand for mobile data is doubling year-on-year is projected to continue As QoS demands increase with more streaming and high bandwidth interactive content, networks are being dimensioned more for peak data throughput than average With 70% of calls originating or terminating indoors, serving those users with sufficient radio and core network capacity from the macro network will get harder and harder Spectrum is being freed for licensing, but in more and more fragmented bands The migration towards 4G cannot rely solely on new spectrum as in previous generations, but must re-farm and coexist in spectrum already in use for older technologies Operators continue to struggle to contain operational costs in a climate of flattening revenue growth and steepling customer expectations.
The Solution In recognizing the key pain points affecting the industry today, ip.access offers a new generation of radio access network, nanoConverge. The key features of nanoConverge address those pain points head on: LTE, 3G and WiFi capable simultaneous multi-standard small cell basestations (nanoCells) deliver high rate data to customers over the most economical, highest performing radio interface with a roadmap to IP Flow Mobility. These same basestations are offered with high-order multi-band capability typically four band coverage allowing operators to serve regions covered by different licensed frequencies with a single small cell access point for maximum logistical efficiency. Innovative architectural solutions allow data delivery with minimum core network loading, while satisfying lawful intercept regulations. High functional Operations and Maintenance solutions, coupled with distributed Self Organising Network (SON) features in the basestation, delivered via a Network Orchestration System (NOS) help operators minimize their operational expenses while delivering a highly optimized performance solution.
The ip.access Market Vision ip.access specialises in providing small cell solutions for indoor enterprise, public coverage and residential applications. We also recognize the commonality between public indoor and public outdoor coverage, and enable our partners to address this important outdoor segment (for both developed and emerging markets) with small cell modules to be incorporated within their existing strand-mount, satellite, marine, aero or security products.
-3-
ip.access 2012
70% of mobile voice and data traffic is generated from inside buildings, divided roughly 50:50 between enterprise and home. By focusing on the in-building space, we are following the pareto rule to make maximum contribution to network performance from our particular market position. Our small cell technology is used by Cisco to deliver AT&T's 3G residential femtocell solution (branded 3G MicroCell) the worlds largest small cell deployment by volume and by value. Our LTE-A technology, including Gateway and Management System, coupled with our existing nano3G HSPA+ technology and enhanced by 802.11n/ac WiFi provides the perfect solution for our customers serving the indoor space.
-4-
ip.access 2012
Introduction
Market Overview The world is moving towards 4G-LTE-A as its cellular radio technology for the foreseeable future. However, different parts of the world are moving at different rates and with different priorities, and there is the important distinction everywhere between the outdoor macro and the indoor small cell in which ip.access has built its business. Global Observations The Cisco Visual Networking Index (VNI) shows the demand for mobile data doubling year-onyear. Importantly, it has shown this trend consistently for the last several years.
Figure 1 - The Cisco VNI, Fixed and Mobile
In terms of trend information, this graph shows two further key pieces of information In 2015, mobile (cellular + WiFi) will overtake fixed data delivery Cellular data is growing at 92% CAGR, with WiFi at 39% CAGR o cellular data will draw level with WiFi by the end of the decade
Additionally, there are multiple press reports, and our first-hand experience in some markets, that signalling capacity in mobile networks is being reached already, especially where smartphones are in heavy use. The traditional 3G radio access network (RAN) architecture is particularly vulnerable to this (the RNC is vulnerable to signalling overload from the rapid data start-stop of smartphone app behaviour). It is one of the drivers towards the femto and LTE RAN architecture, which does not use an RNC and distributes the RAN signalling to the Access Point.
Figure 2 - Fierce Wireless reports Credit Suisse, July 2011
-5-
ip.access 2012
While average monthly mobile data usage is, in fact, still fairly low, the peak load is beginning to strain the capacity of the core networks. This peak loading is becoming a greater problem as 1 operators seek to offer services requiring high Quality of Service (QoS) compliance, and their networks need to be dimensioned for peak load, rather than average. This trend is forcing offload strategies in two directions. On the one hand, operators are offering WiFi services as an alternative/adjunct to mobile cellular. Secondly, RAN offload strategies within mobile networks are being standardised to allow mobile data to be carried without loading the core. In terms of radio layer congestion in the face of the yearly data doubling around the globe, we also see a requirement to exploit all of the spectrum assets accessible to an operator. Therefore multistandard operation, including the integration with WiFi is key to the long-term success of the solution. This allows operators to extract value from their spectrum licenses in a way not possible in previous generations of technology. North America Within the North American market, LTE deployments are furthest ahead, but are still relatively limited in scope, with only major cities being covered (and those only partially) at the time of writing. Key trends are clear already from this market. The requirement for multi-standard
Our lead customers are committed now to a multi-standard world. They recognize that there is so much invested in existing GSM and 3G infrastructure that the idea of LTE displacing it is not tenable.
The requirement for multi-band (the Single SKU requirement)
Within North America, spectrum is licensed on a Major Trading Area (MTA) basis, so that depending on exactly which MTA a device is located in it may or may not be licensed to use particular spectrum. To try and manage the logistics of multiple single-band devices being deployed into particular locations, given the expected scale of small cell rollouts, is a frightening prospect. The only feasible solution is to make the basestations capable of operating in all the licensed bands, so that the device can select at run-time which band to operate on it can never be in the wrong MTA for its capabilities. This so-called Single SKU requirement is a key product and technology driver, and has its equivalents in other parts of the world.
The requirement to support spectrum re-farming In the US, spectrum is not licensed on a specific technology basis, and it is up to the carrier and their suppliers to agree the usage. The idea of re-farming is actually a misnomer in this context, since the operators are free to operate any technology in any band without any special regulatory provision. For instance, we are operating 3G femto in spectrum on which 2G basestations are
We see the Quality of Service (QoS) discussion is being modified by the introduction of the concept of Quality of Experience (QoE), where the end-users experience of the content is measured, rather than the performance of any particular data service. This development is key in ip.access response to the data offload and network performance requirements as well see below.
-6-
ip.access 2012
already deployed co-existing quite happily. It is only in the two 700MHz bands that we expect LTE to be deployed exclusively. In all other bands (PCS1900, Cellular 850 and AWS 1700/2100) we see standards co-existing within the same bands, and even, as described above, in the same channels. We expect this example to become the norm, not just in North America but around the world, where traditional spectrum allocations to specific technologies are becoming obsolete. For example, Europe has already moved to allow re-farming of GSM 900MHz spectrum for 3G. The focus on Enterprise and metro over residential While the business case for 3G residential femto has been proven, and ip.access is supporting the biggest femto deployment in the world, the drive towards LTE residential femto will take some years to bear fruit. The LTE technology base will take several years of Moores law evolution to support the magical $100 price tag. The need for LTE in the home does not appear to be a priority for North American operators. When it does appear, it is likely that it shares the same requirements in terms of multi-standard and multi-band capability as the enterprise and metro equivalents placing further pressure on its pricing, and pushing its earliest arrival date further out. LTE as the Value Play LTE has been traditionally featured as a performance solution, but at least one operator (metroPCS) is deploying it as the lowest cost way of deploying a mobile data network. metroPCS sells exclusively contract-free, pre-pay subscriptions. Its subscribers deliver the lowest ARPU in its market, yet by a selective deployment of LTE alongside its CDMA network, it is offering a compelling combination of voice and data services to deliver a successful business. Its voice service is based on CDMA 1x, but its data service is built around LTE 1.4MHz in the AWS band. This configuration is possible because of the unique heritage of the 3GPP2 operators (Verizon included) in that the voice and packet data services are on wholly different radios within the network and handset, with no integration at the physical layer. For the 3GPP2 operator, the introduction of LTE as an adjunct to or substitute for EvDO is entirely natural, and as far as we can see, entirely successful. The requirement for location verification As remarked above, spectrum in North America is licensed on an MTA basis. It is a license compliance requirement that basestations operate in the spectrum that is appropriate for the MTA in which they are located. While MTA boundaries originally ran through sparsely populated areas, the pace of development has now placed some MTA boundaries within urban or suburban centres. For small cells, which may be deployed by enterprise IT staff or even end-users, rather than carrier-trained experts, it is an essential requirement to verify where the basestation is to high degree of accuracy. This is in addition to the requirement to accurately locate the originator of an emergency call (the so-called E911 requirement). While the E911 requirement for residential femto may be satisfied by simply knowing the location of the basestation (essentially the same requirement as the MTA verification requirement), as small cells get larger for enterprise or public indoor use, its expected that the E911 requirement will rely on existing E-TDOA equipment or direct GPS interrogation of the handset, rather than location verification of the basestation. Japan & Korea The initial LTE focus in many countries will be on the macro network rollout, with small cells being deployed later as an overlay for hot-spot capacity and coverage. However, Japan and Korea are
-7-
ip.access 2012
significant exceptions, where expectations for indoor coverage and performance require that small cell deployments are planned in parallel with the LTE macro network rollouts. For example, SKT plans to deploy LTE in 84 cities across Korea by April 2012, after which it will utilize Wi-Fi integrated femtocells to boost service quality in densely populated regions. Their initial focus for small cells will be the central commercial areas of major cities where LTE traffic and users are concentrated (ie metrocells), followed by expansion to indoor areas with high demand (shopping centres, offices, schools and homes). BRIC The defining characteristic of the BRIC nations (Brazil, Russia, India and China) is speed of deployment. Most of these states have fast-moving network roll-outs in progress, based originally on 2G but increasingly on 3G. Their 4G proposition is still evolving. In markets without a high degree of 3G incumbency, the introduction of 4G alongside GSM is technically highly advantageous. The TDMA nature of both interfaces, with the fine control of spectrum utilisation possible in LTE makes a co-existence of LTE with GSM a highly spectrally efficient and low opex voice and data solution. The compatibility of the air-interfaces (both use relatively narrow RF channels to manage their spectrum) makes the refarming of spectrum from 2G to 4G an important proposition. Developing World This segment epitomises the green-field operator, who is developing a cellular network where there is none in existence today. In this situation, as we saw with metroPCS above, LTE is a compelling technology on which to base a new network for the longer term. The rational strategy here is often to roll out a GSM network, to take advantage of the low-cost technology and high level of deployment experience, and then to position it alongside LTE for the longer term and to serve the growing number of mobile internet users. Europe LTE roll out in Europe is considerably behind that in the US, Japan or Korea. The first introductions are almost exclusively at 2.6GHz, which, owing to the relatively high propagation loss at this frequency, creates an immediate use case for in-building LTE. As further, lower frequency spectrum is auctioned, then the macro network deployments may gather pace. The driving force for European deployments, though, is the classic performance enhancer. LTE adds spectrum, offloads traffic, and therefore improves network performance as perceived by all users.
-8-
ip.access 2012
Solution Overview
Summarising the Benefits Product/Feature A new generation of multi-band, multi-standard basestations, aggregating LTE, HSPA+ and WiFi 802.11n Benefit Multiplies rather than sums spectrum and technology assets. Multiple technology support gives longevity in the face of spectrum refarming. Removes logistical deployment issues any basestation can go anywhere Global coverage with minimal logistical complexity Further increases spectrum options, especially in China Minimises cost-per-bit of data delivery, while maximising end-user performance
Core design capable of two channel (+WiFi) operation on any of four bands The four bands can be selected for US, Europe or other specific territory requirements Roadmap to LTE-TDD A gateway and management architecture that allows self-organisation and global optimisation of performance Supporting the 3GPP self-organisation features for LTE and 3G 2 in the basestation, In the NOS (NOS assisted SON) and In the OSS (via SONiX) rd Open interfaces to support integration with 3 party macro RAN and existing operator OSS solutions NOS northbound interfacing Innovative IMS support options
Minimises network operational expense on a network wide basis (not just the small-cell layer)
Allows interconnect with existing NMS solution to minimise integration costs and maximise ongoing opex savings
Network offload architectures ( LIPA, LIMONET and SIPTO and Content Aware Scheduling) to offload the mobile core network.
Targeted at real-time streaming, interactive, secure local data access and other metro/enterprise services
Enables a migration to a single, IMS based service environment, with long term savings in multiple core network maintenance. Minimises cost-per-bit of data delivery, while maximising end-user performance Only a fraction of mobile data needs to transit the core Reduces operational cost associated with the dependency on QoS management across multiple domains Maximises service revenue from the valuable enterprise segment. Provides a platform for future residential offerings
NOS = the ip.access Network Orchestration System, SONiX = SON Information Exchange LIPA = Local IP Access, LIMONET = LIPA with mobility, SIPTO = Selective IP Traffic Offload.
-9-
ip.access 2012
Unified architecture for 3G, 4G and WiFi
Inter-Radio-Technology (IRAT) mobility to ensure dynamic, balanced utilisation of the multiple radios
Minimum gateway count reduces operational expense, real-estate and power requirements in the data centre. real-time balance is the key to high throughput, high quality radio performance and is the optimal way to multiply three optimal radio technologies. Gives the ultimate flexibility in delivering data to the end user using multiple airinterfaces simultaneously.
Roadmap to IP Flow Mobility (IFOM)
Solution Architecture The basic solution architecture is shown in FIGURE 3Error! Reference source not found.. The day-one architecture shown in Error! Reference source not found. includes the key features for a high-performance Radio Access Network in a modern heterogeneous network (HetNet). All the basestations are hosted in a common framework, with a single gateway and management system operating all devices, 2G, 3G and 4G, residential femto and pico class small cells for enterprise or metro application. The network is flattened, with all RNC functions hosted within the 3G basestations. The architecture includes a HNB-GW function to provide aggregation to the core, as required by the 3GPP standard. The S1 interface from the E-100 LTE device can be aggregated within the gateway or may be directly routed from E-100 to the core via the SecGW. (There are good value-add reasons for using a gateway SON interworking and QoE/QoS management among them which we describe in more detail below.)
Figure 3- The Basic Network Architecture
- 10 -
ip.access 2012
The X2 interface between the E-100 LTE node and the macro network is routed via the gateway, which can provide some aggregation of signalling, and processing of SON measurements as described more fully below. The Network Orchestration System (NOS) provides a layer management function, allowing all the subtended basestations of whatever technology to be managed from a common system. The NOS provides ACS (TR-69 based mass provisioning) for femto style deployments as well as traditional FCAPS for enterprise and metro style deployments. The NOS also manages the other nodes in the layer, including the gateway, the NTP servers and the security gateway. This architecture is enhanced over time to include the architectural and radio-bearer management functions, as follows.
FIGURE 4 - NANOCONVERGE GATEWAY WITH SELECTIVE TRAFFIC OFFLOAD
With the inclusion of core network offload features, the architecture of the gateway is enhanced to include a SIPTO (Selective IP Traffic Offload) port, as shown in Figure 4. Note that there is one SIPTO port for all technologies, including 3G and 4G and can similarly provide a route for WiFi traffic where this option is fitted within the E-100. Such technology-independence provides an important simplification for IP network engineering and deployment management, as well as a single node to connect billing and lawful intercept gateways.
- 11 -
ip.access 2012
Figure 5 - nanoConverge with Enterprise Gateway on t he Customer Premises
And for enterprise deployments with scale, an Enterprise Gateway is added to the architecture as shown in Figure 5. The purpose of the Enterprise Gateway is to provide a single point of management and administration within an enterprise. For small enterprises that only require one or two basestations, the features can be hosted within the E-100 itself. Some key points of the Enterprise Gateway It is a software application that can be hosted on an enterprise router platform it needs no special hardware beyond some IPSec acceleration. The gateway hosts IPSec termination for devices on the enterprise network, including IWLAN compliant WiFi devices. It also originates an IPSec tunnel to the core network gateway. All telco traffic on the enterprise LAN is therefore completely secure. It is only in the clear when intended to be so for direct internet access or other enterprise application use. The embedded LIPA/LIMONET function allows users to access local enterprise data securely and quickly, and with zero tromboning to the mobile core network. The IRAT Mobility function operates on the enterprise cluster to ensure that the basestation load is balanced one to another, and also balanced from one technology to another to ensure global maximum network performance. The IP Flow Mobility (IFOM) function allows data streams to be forked (downlink) and recombined (uplink) within the gateway and routed over the most appropriate air-interfaces to the IFOM handset (note the plural air-interfaces). The architecture allows aggregation of IFOM branches across different basestations so that resources within the network are truly balanced. The V-GW function is a VoIP signalling and codec implementation that allows mobile integration with the enterprise fixed voice network. The V-GW might use native VoIP applications within the handset (where the LIMONET gateway allows VoIP with mobility wholly within the enterprise). It can also provide an interworking function between VoIP on the enterprise LAN and classic Circuit Switched Voice on the mobile segment.
- 12 -
ip.access 2012
How are the benefits summarised above implemented in the ip.access solution? Benefit Multiplies rather than sums spectrum and technology assets. Multiple technology support gives longevity in the face of spectrum re-farming. Network Implementation The E-100 basestations are equipped with Inter-Radio Access Technology (IRAT) mobility functions. Spectrum and throughput resources are balanced between the network technologies within the basestation. In hardware terms, the basestations are built with two independent radio channels that may be tuned anywhere within four selected, geographically optimal bands. The two radio channels may operate in any band, using 3G/HSPA+ or LTE, thus allowing the basestation to continue in use even after the spectrum has been refarmed. The ability to tune within up to four radio bands means that licensing restrictions as to spectrum are removed. The basestation is effectively no longer bandspecific or 3G/4G specific. The underlying technology in the basestation is intrinsically wideband, and supports FDD and TDD operation. The power and filtering requirements for any given geometry are restricted to a subset of the circuit, so delivering a device into a geography with specific band requirements is greatly simplified. The ability of the underlying baseband and RF technology within the basestation to support TDD allows deployment in regions with TDD requirements. The network architecture supports LIPA (local IP access), LIMONET (which is LIPA with enterprise mobility) and SIPTO (for traffic routing optimisation). Such functions may be hosted within the basestation itself, the enterprise gateway or the core network gateway for maximum flexibility. The features allow the data to be routed to its destination without transiting the core network a major cost saving.
Removes logistical deployment issues any basestation can go anywhere
Global coverage with minimal logistical complexity
Further increases spectrum options, especially in China
Minimises cost-per-bit of data delivery, while maximising end-user performance
- 13 -
ip.access 2012
Minimises network operational expense on a network wide basis (not just the small cell layer)
The SON implementation within the network is done at three levels. Within the basestation, specific SON features allow the basestations to configure themselves within constraints set by the network operator Within the NOS, optimisations are performed, based on KPIs passed from the gateway, and configuration passed back to the basestations, specifically for clusters of interacting cells within the management view of the NOS Northbound from the NOS, the NOS itself passes KPI data to the OSS and core network optimisation functions, so that the small-cell layer can be optimised within the context of the whole network Such automation support reduces to a minimum the requirement for manual intervention in the operation of the small cell network. In terms of RF layer optimisation, in larger enterprise clusters, the intertechnology load balancing may be hosted within an Enterprise Gateway to ensure global optimisation of performance and spectrum utilisation within the enterprise. This has the side benefit of minimising RF power settings, and therefore minimising RF interference within any umbrella macro coverage. The NOS provides a rich suite of northbound interfaces for configuration management, fault management, localisation and diagnosis and performance management. 3GPP standard interfaces are used throughout, including support for popular industry standards.
Allows interconnect with existing NMS solution to minimise integration costs and maximise ongoing opex savings
- 14 -
ip.access 2012
Enables a migration to a single, IMS based service environment, with long term savings in multiple core network maintenance.
Minimises cost-per-bit of data delivery, while maximising end-user performance
With the multiple radio interfaces within the basestation all able to pass their data through a single IP gateway (by virtue of the LIPA, LIMONET and SIPTO functions), this allows a very simple integration with the IMS or other unified IP based service delivery platform (such as an enterprise server). The single point of interconnect allows the user to register and be paged through a single port, and the particular radio access technology that is chosen for the session is up to the basestation. This reduces the role of each of the core networks in the service provision, and allows the operator to focus their service innovation resources in a single place (the IMS) rather than having to reduplicate such resources over multiple core networks. The key feature of the LIPA/LIMONET/ SIPTO functions is that only a fraction of mobile data needs to transit the core. Also, that fraction can be dynamically selected, according to service routing requirements and lawful intercept and other regulatory requirements. In this way, the core network can be focused on maximum revenue generation and regulatory compliance function where this is necessary. All other low revenue traffic can be delegated to the shortest, or cheapest route according to policy and current network conditions. The LIPA and related functions within the solution can be configured dynamically with this routing information via the NOS. This is the Content Aware Scheduling feature. This feature enhances and can in some cases offset errors in outerpacket QoS marking by queuing data across both lossy interfaces (the backhaul and the radio) according to its importance to the content in question. It exists within both the gateway and the basestation to ensure that the scheduling algorithm at the gateway is accounting for losses in both backhaul and radio interfaces in calculating its scheduling scheme.
Reduces operational cost associated with the dependency on QoS management across multiple domains
- 15 -
ip.access 2012
Maximises service revenue from the valuable enterprise segment. Provides a platform for future residential offerings
Minimum gateway count reduces operational expense, real-estate and power requirements in the data centre.
Real-time balance is the key to high throughput, high quality radio performance. The optimal way to multiply three optimal radio technologies.
The feature set is specifically designed to offer high performance and secure access in an enterprise environment (see Figure 5 above). While the solution is tuned for performance in the enterprise, the much greater data volumes in the consumer space mean that the offload and RF load balancing techniques pioneered here will be equally valuable to the network operator and consumer alike, when applied in the residential space. The gateway is a single device covering all technologies. The scale of the gateway is determined solely by the scale of the traffic it is having to manage. Coupling the gateway investment to the traffic rather than to the technology is the right way forward. The essential idea here is that any user of the solution, in order to get best service, should be able to choose the lightest loaded network option. The mobility, load-balancing and IP Flow Mobility features within the basestation and Gateway functions mean that the user is never camped on a radio that is more heavily loaded than another. Therefore, when they start a call, the chance of a directed retry to a better cell/better RAT is optimally minimal. Therefore they get best performance, and the network operator gets minimal signalling load.
- 16 -
ip.access 2012
Gives the ultimate flexibility in delivering data to the end user using multiple air-interfaces simultaneously.
This is the IP Flow Mobility feature (IFOM). The IP Flow Mobility gateway may be implemented within the basestation for single cell operation and may be hosted within the Enterprise Gateway for clustered operation. In both cases, the IFOM network entity peers with a corresponding UE entity, and the two signal to each other during the startup phase of the IP Flow, whether further radio interfaces need to be established. For instance, if the flow starts on LTE, the two entities decide together whether the flow should continue on LTE, should transition to WiFi, or should be aggregated with WiFi for super-high bandwidth applications. Other radio combinations are possible, and the policy decisions regarding their relative priorities is configured, but the decision making is done in real time using the policy and also real time measurements of network load and radio link quality.
- 17 -
ip.access 2012
Product Focus
The E-100 LTE Access Point Reproduced below is the datasheet for the E-100 basestation. Note that this is advance information, subject to change without notice.
- 18 -
ip.access 2012
- 19 -
ip.access 2012
Gateway The Gateway for the ip.access LTE solution is identical to the nanoGateway 300 gateway already in use for 3G femto and enterprise systems. The reader is referred to existing documentation for that product. Network Orchestration System The reader is referred to existing product documentation on the Network Orchestration System.
- 20 -
ip.access 2012
TECHNOLOGY TOPICS
The Multi-Standard Basestation The availability of the highly integrated, high performance baseband processing, with RF agnostic modulator chips allows us to introduce multi-standard basestations with very compact form factor, for enterprise as well as metro applications. As the technology develops, the residential form factor and price point will similarly come within reach for these multi-standard devices. The key point here is the ability for a single device to carry user data in a balanced way, over the most appropriate radio, whether that be LTE, HSPA+, 802.11n or a combination of the above. The multi-standard, multi-band basestation allows the operator to trunk their radio interfaces in a way that has not been possible in the past, and get the associated trunking gains of capacity in the radio that they have been used to in the fixed segment for years. Inter-RAT Mobility Key to achieving the radio layer trunking gain, and implementing the Mobility Load Balance SON feature, is the ability to move users from radio layer to radio layer smoothly, with no interruption of data flow. ip.access has a unique framework to achieve this, taking account of multiple factors to achieve the load balance, as shown in the diagram. In previous generations of device, Inter-RAT mobility is implemented, but it is strictly between basestations for instance handover from femto 3G to macro 2G. This is rather simply calculated on the basis of UE measurements and, to some extent, UE service requests, as shown in Figure 6.
Figure 6 - Classical Inter-RAT Mobility
However, the co-existence of multiple RATs serving essentially identical user populations creates new balancing problems/opportunities, as shown in Figure 7. Layer balance (occupancy balancing) o To minimise the chances of blocking, and to maximise the offered throughput to each user, the occupancy in each layer should be balanced according to the technology demands within the population
- 21 -
ip.access 2012
Each new call should experience the same chance of blocking (or throughput availability), regardless of which layer it originates from o This is the function of the Layer Balance module Service/user prioritisation o Decision making is biased by policy factors certain services and users, may be preferentially steered to certain technologies according to QoS CPU load o In the multi-standard basestation, moving devices between layers may have a CPU architecture dependent effect on the total CPU load. Therefore, we provide static limits of layer loads, adapt and predict usage to manage load to a maximum operational level UE measurements and events (RF resource balancing) o The UE population link budget is now managed dynamically, so that layer changes take into account the total link budget quality
Figure 7 - IRAT Mobility with Load Balancing
WiFi integration and IP Flow Mobility In addition to the LTE and HSPA+ air interfaces, adding 802.11n to the mix allows an even greater radio layer trunking gain. 3GPP recognises this, and has an active work item called IP Flow Mobility (IFOM), as shown in the diagram.
- 22 -
ip.access 2012
PDN1
PDN2
PDN1
PGW1
PGW2
PGW1 PDN connection #1 flow #1 flow #2 flow #3 non-3GPP access 3GPP access
DN connection #1 PDN connection #2 non-3GPP access 3GPP access
UE
UE
Figure 8 - IP Flow Mobility (IFOM)
ip.access sees IP Flow Mobility as the logical culmination of efforts to provide technology that allows the operator to use their spectrum assets in the most cost-effective, spectrally efficient and performance enhancing way possible. By trunking all of the air-interfaces, cellular and unlicensed, and providing the technology to move smoothly between them to allow a state of balance to be sustained between them, we are close to the theoretical limit of capacity provided by these three air-interfaces which are, in themselves, also close to their own respective theoretical limits. Content Aware Scheduling LTE offers a much richer mix of QoS categories for its content, but multiple factors affect the ability of a carrier to reliably offer services based on these QoS categories: Appropriate QoS marking of the content may not always happen. In particular, where content is a mash-up from multiple sources, the QoS marking of the stream may in practice be some arbitrary value, unrelated to the QoS requirements of the sub-streams. o Example: a web page that contains YouTube content may be transferred using the QoS appropriate to the static content of the page. The streaming video that the user is interested in may carry the same QoS marking, and the user experience of the streamed content will be poor. The content may be marked appropriately for the LTE air-interface, but, given the desire to offload to or aggregate with other radio technologies, the LTE QoS markings may not be mappable to a WiFi or HSPA+ QoS queue. o HSPA+ defines 4 traffic classes and 13 QoS attributes whose associated complexity leaves them often unused o LTE standardises on 9 QCIs (QCI = QoS Class Identifier an abstraction of a concrete QoS class) and the mapping between the LTE QCI and the 3G Traffic Class + QoS attribute setting is ill defined. Given the imperfections of the backhaul interface, the air-interface QoS marking may be rendered meaningless or impossible to achieve by packet loss or retransmissions in the IP network between the content and the basestation. o Transit networks may or may not respect QoS markings.
- 23 -
ip.access 2012
Effectively, we are asking a single QoS scheme to define behaviour across two lossy networks the backhaul and the radio.
One way of cutting through this complexity is to introduce the concept of Content Aware Scheduling, defined on a stream by stream basis, between the data source and the UE, and implemented jointly on the core network gateway, any intervening gateways (such as LIPA/Enterprise gateways) and the LTE basestation.
Figure 9 - Content Aware Scheduling
The scheme is illustrated in Figure 9, where the top half shows the way things are done today. The UE requests a QoS class at the time of PDP context activation. If the session is invoked from a web-browser, then the QoS class will invariably be Interactive/Background or similar. So, if you browse to YouTube for instance, and expect Streaming QoS then youll be disappointed. Theres nothing in the system to say this content is streaming video Ill change QoS class. This is one of the reasons why such content has its own application in the smartphone world. YouTube and other content providers with specific QoS requirements now give you the option of using their specific phone App, which knows that you are looking for Streaming QoS and a certain bit rate, and can therefore request it when the PDP Context is set up. Even so, for all the reasons described above, this may not be enough. The point of Content Aware Scheduling is to subdivide the QoS requirements of specific content streams, based on their encoding or other information, and can queue them accordingly within QoS queues that the gateway sets up to the basestation on its own assessment of the content, not on the semi-blind request of the PDP Context Activation. Of course, the QoS management within the basestation is somewhat determined by the QoS class request of the PDP Context, but even so, there is enough latitude within the specifications to make a significant improvement in performance of the application as perceived by the user (the Quality of Experience or QoE), to make the exercise worthwhile.
- 24 -
ip.access 2012
Network Offload (LIPA, SIPTO and LIMONET) Network offload architectures, such as LIPA, SIPTO and LIMONET, are key to realising the potential of the radio interface, even when the backhaul to the basestation or enterprise cluster is of limited bandwidth. The LIPA and LIMONET architectures in particular allow enterprise users to access local data and services without having to traverse the mobile core. Figure 10 illustrates the LIPA architecture. The key element is the Local Gateway (L-GW in the figure) whose presence is signalled to the SGSN as a potential GGSN at the time of PDP Context Establishment. The SGSN can then select the L-GW for this PDP Context, according to some policy rules within the core network. This is a relatively simple architecture, effective at offloading the core, but its defining characteristic is that the L-GW is part of the H(e)NB. When the user hands out to another cell, the data path is broken, and so the PDP Context must be re-established on the new cell. Thus, this architecture does not support mobility.
Figure 10 - Local IP Access (LIPA)
An improvement on the LIPA architecture of Figure 10 is shown in Figure 11. This is one of two architectures being considered by 3GPP currently, but is seen as preferred and is therefore likely to become standardised. The essential point here is that it separates the L-GW from the H(e)NB and therefore allows the user plane to be routed to the L-GW from any other basestation also reachable from the serving gateway (S-GW, or SGSN in the 3G case) and from which the user plane is routable.
- 25 -
ip.access 2012
SGi
L-GW
User plane X2
S5
SeGW
S1-U S1-U
HeNB
S1-MME Uu
HeNB GW
SGW
S1-MME
S11
UE
MME
Figure 11 - LIMONET architecture (option 2 preferred)
ip.access implements the LIPA architecture, initially in the basestation, with a roadmap to LIMONET, with the L-GW in the enterprise gateway or the core network gateway. The Importance of FAPI The Femto Application Platform Interface is an innovation within the industry that allows silicon vendors, software vendors and product developers to work together around a common interface that allows the product to make use of the best silicon and software available. It formalises the idea of platform independent software in the small cell. FAPI's main benefit to the small cell provider is that it allows them to use the best silicon for the best application. For instance, one silicon vendor may offer the best performance metro solution, whereas a different silicon vendor may offer the best value residential solution. The product developer can serve both markets, preserving the value in their software, hosting it on the appropriate silicon for the application. ip.access, in chairing the Femto Forum working group responsible for the introduction of FAPI, is a prime supporter and user of the FAPI interface. High Performance, Low Cost, Low Power Consumption Baseband Silicon In the last years, many new silicon vendors have entered the small cell market with FAPI compliant offerings. Such devices form an essential part of the technology base that makes multi-standard, multi-channel small cells feasible. Coupled with a FAPI compliant Physical Layer software component, the range of device families available allows us to target the enterprise and metro indoor application space with the most suitable device in terms of CPU performance and DC power consumption. Wide Band, Technology Agnostic RF Silicon There are two significant sources of wide-band, technology agnostic RFICs on the market, and ip.access is evaluating both of them. The key enabler is the ability to tune the chip to any LTE
- 26 -
ip.access 2012
band in the 600MHz 3.6GHz range and configure to any bandwidth up to 20MHz, with 2x2 MIMO support included. The chips also support 3G modulation and are both ideal candidates for a multistandard product. Self-Organisation, Self-Optimisation Product and system features to allow the RAN to organise and optimise itself are a key technology for LTE and ip.access has an active programme to implement and deliver self-organising features into the first generation product. Key differentiators: Inter-cell Interference Coordination (ICIC) for Physical Resource Block coordination with macro neighbours Classic ICIC is intended for interference coordination at cell edges. We enhance the techniques, noting that small cells actually wont always be at the macrocell edge. In doing so, we use existing ICIC primitives to optimise the technique for small cell deployments. Automatic Neighbour Recognition (ANR) ip.access pioneered the use of an embedded downlink receiver (Network Listen) to detect the local radio environment. Our starting point for ANR is the detection and decode of neighbour cell transmissions to set up the initial neighbour list for reselection and hand-out. Our Network Listen receiver for W-CDMA also decodes GSM, and similarly our LTE NWL receiver decodes W-CDMA and GSM also. For CDMA markets we work with partners to enable ANR to 1x cells. Other SON techniques The other 3GPP SON techniques (including Mobility Load Balancing (MLB), Mobility Robustness Optimisation (MRO), enhanced ICIC (eICIC), Physical Cell-Id (PCI) selection and so on) are subject to active R&D, with patent filings ongoing. Further details on all these SON techniques are available on request. IMS access One consequence of the aggregated RAN vision, coupled with the data offload architectures enabled by LIPA and LIMONET, is an architecture that places the RAN, and in particular the basestation at the centre of the network, in terms of the actual service delivery to the user equipment. The concept is to connect the service client (IMS enabled) in the UE via a generalised PDP context to the service provider in the IMS via the nanoConverge RAN. The IFOM function within the basestation peers with the connection manager in the UE to carry that PDP context over the best combination of radio interfaces LTE, HSPA+ and/or WiFi and deliver it to the IMS gateway via the local LIPA/LIMONET gateway. This architecture enables the acceleration of the transition to IMS hosted services, common to all radio interfaces, since we are simply using the mobile core networks as authentication and mobility servers for the radio bearers. The actual decision as to which bearers to use is negotiated dynamically between the UE and the RAN at the start of the service and adapted continuously for the duration of the service. Additionally, by including IMS to CS voice interworking within the basestation, we can serve R99 handsets on the same infrastructure.
- 27 -
ip.access 2012
With nanoConverge, operators can accelerate their transition to a common IMS service platform, and reduce or even completely avoid continued investment in individual mobile core services. All future service creation investment can be focused on the IMS as a single platform for all the radio technologies in a carriers portfolio. The basic architecture is shown in Figure 12.
Figure 12 - IMS Access using the ip.access E-100
- 28 -
ip.access 2012
Summary
As we move into the era where the consumption of mobile data becomes the norm, rather than a niche, indoor small cell deployments will become key for any operator. The clear preference of subscribers to take their data indoors makes a high performance indoor cellular network layer a necessity. ip.access' nanoConverge small cell Radio Access Network offers a powerful and flexible solution. By a careful analysis of the market trends, and drawing on years of experience serving top tier operators in enterprise markets, ip.access has identified an optimal suite of products, architecture, technology and features. This combination is easy to deploy, maximises the network performance by offloading the macro radio and core, and minimises operational costs by providing abundant self-organisation. In the network operator toolkit of the future, there will be no bigger wrench.
- 29 -
ip.access 2012
ip.access ltd Building 2020, Cambourne Business Park, Cambourne, Cambridge, CB23 6DW, UK T +44(0)1954 713700 F +44(0)1954 713799 info@ipaccess.com www.ipaccess.com

LTE Market, Technology, Products: DR Nick Johnson, CTO, Issue 1.1, Ip - Access, February 2012

Uploaded by

Copyright:

Available Formats

LTE Market, Technology, Products: DR Nick Johnson, CTO, Issue 1.1, Ip - Access, February 2012

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

LTE Market, Technology, Products: DR Nick Johnson, CTO, Issue 1.1, Ip - Access, February 2012

Uploaded by

Copyright:

Available Formats

LTE Market, Technology, Products

Dr Nick Johnson, CTO, Issue 1.1, ip.access, February 2012