TDD mode, which shares the same frequency band for both uplink and downlink, hasn't been widely adopted in 3G cellular systems to date. The new 4G technology LTE actively embraces and adopts this mode, which can make use of a lot of otherwise wasted spectrum, especially at frequencies suitable for small cells.
We spoke with Renuka Bhalerao, Senior Product Line Manager at Radisys, who supply software solutions for small cells, to understand the latest market developments.
Access to lots of lower cost spectrum
Spectrum is one of a mobile operator's most precious resources, often expensive to acquire and strictly rationed by regulators. Yet a surprising amount of it lies fallow and unused because of a lack of device and equipment compatibility. Examples include Clearwire which has been said to have as much as 80MHz unused worth over $6 Billion, and the many 5MHz allocations to European operators for 3G TDD mode which never went live.
TD-LTE is a different story from that 3G TD-SCDMA saga for several reasons:
- The specification is very much closer to FDD mode than was the case for 3G, and contains the differences almost entirely within the MAC-PHY (Physical Layer)
- There is already a much wider support for TDD mode in end user devices and chipsets. According to the GSMA, there are already over 200 TD-LTE devices available.
- TD-LTE is fully compatible with all FDD-LTE services, supporting the full range of services, protocols, interfaces and core network.
These factors help make TD-LTE an easier technology to adopt than 3G TDD, because of the reduced level of investment required.
Frequency bands commercially in use today and which could be very popular for TD-LTE include 2.3GHz (China, Japan, India), 2.6GHz (China plus a 50MHz wide band across Europe) and 3.5GHz (UK, possibly USA in future). Few countries have TDD spectrum allocations below about 2GHz, and so these higher frequencies are most likely to be used with small cells because of their short range.
Which countries are seeing the most action?
Today, we are primarily seeing activity in the Far East, specifically China, India and Japan. At this early stage, live deployments are using macrocells for coverage first. China Mobile is setting the pace, investing $15 billion in TD-LTE during 2013 alone and asking suppliers to tender for 207,000 extra macrocells. Supplementing that, their "Nanocell" initiative is on track for a major deployment from early 2014, and will be a great proving ground for the technology.
Softbank in Japan has been a keen supporter of TDD mode, specifically at 2.3GHz, and have probably the largest commercial TD-LTE networks worldwide, with some 500,000 subscribers. Now that they have acquired Sprint, we may see that influence their investment decisions in the US.
According to the GSMA, some 54 mobile operators worldwide have announced plans for a TD-LTE launch (out of the 200+ for LTE overall), of which 16 are already in service. It could be argued that 150 are potentially missing out; many may have simply prioritised FDD mode with a view to using TDD mode in the longer term. Ovum predicts that FDD mode will dominate, forecasting that by 2016 we may see around 520 million LTE-FDD users vs 180 million LTE-TDD. They foresee that few will be operating in networks with both FDD and TDD mode at that stage.
What are the primary benefits of TD-LTE over FDD mode?
TDD mode is asymmetric and can adapt the proportion of resources used for the downlink vs uplink, maximising the efficiency.
The single band of spectrum required for TDD opens up additional frequencies that otherwise wouldn't be accessible. While this additional supply can drive the cost of spectrum down, some of the more unusual frequencies may not be supported by mainstream smartphones and end user devices.
TDD mode is also good for MBMS (Multimedia Broadcast Multicast Service), where the same data stream can be broadcast to multiple devices simultaneously.
How is voice handled in TD-LTE?
Exactly the same VoLTE (Voice over LTE) service is available using TDD mode as for FDD, and provides the same very fast call setup (under 1 second), optional video calling and other supplementary services.
Operators fall into two camps in their approach to TD-LTE voice:
- CSFB: Those with existing mature 3G networks seem to be using LTE primarily for fast data, retaining 3G for voice by using the CSFB (Circuit Switched FallBack) feature.
- VoLTE only: Those with no 3G network, for example having migrated directly from WiMAX, adopt VoLTE natively.
The packet scheduler needs to take account of specific requirements for voice services such as smaller packet size, low latency and prioritise it appropriately to ensure consistent service delivery.
Can TDD mode achieve high data rates?
Just as for FDD mode, the peak data rates achievable depend on the bandwidth allocated. The downlink speed on TDD often matches or surpasses that of FDD. Carrier aggregation can be used to bond two channels and double overall throughput. China Mobile has demonstrated over 100Mbps in its large scale TD-LTE trials.
At the same time, adoption of small cells can also provide higher speeds and better performance. Being closer to the end-user, the short range radio link can take advantage of higher modulation schemes shared by fewer users to deliver high quality service. Some operators are considering this approach in preference to carrier aggregation.
Today, carrier aggregation is used with signals from the same base station, whether a small cell or macrocell. This doesn't change with TDD mode, which would require even closer co-ordination if used between different cells.
Are there any technical constraints specific to TD-LTE to watch out for?
Phase timing is critical for TDD because of how it works. This is more of a basic platform need. Our software tunes and adapts to the PHY/RF which takes care of that aspect. Timing is much more relevant to how the RF and baseband works - we layer our software on top of it. We have to consider some coordination to deal the time domain switching from DL to UL mode.
Is TD-LTE Mature enough to deploy today?
There's no doubt about whether TD-LTE is technically feasible or not. After extensive trials, there are now multiple operators offering live service on a large scale. There's no need for a long proving phase before deployment.
The eco-system is a lot more global than TD-SCDMA, and we've seen much more customer traction. We've also seen a few vendors using TD-LTE small cells as a lynchpin to launch themselves into the marketplace.
Radisys was first to the market with LTE-TDD software solutions, provides an accelerated TDD Roadmap and is actively supporting customers for TDD trials and deployments. Our Trillium software is already field proven for FDD mode, and we've been able to reuse that experience and Intellectual Property as we support customers like Airspan for their LTE-TDD solutions. We are on track to release a fully TDD compliant and validated product on leading SoCs before the end of 2013.
A few operators plan to deploy both FDD and TD-LTE, refarming different bands over time. In addition to TD-LTE small cells, this will need advanced smartphones which are agile enough to adapt between the different modes and frequencies.
Radisys is a sponsor of ThinkSmallCell.