RAN Optimization

It is worth saying that as a general consideration in LTE systems design, the difference between FDD and TDD mode are minimized. Besides the differences in physical layer, they have the same design and are transparent to the layers above. However, the tiny differences between TD-LTE system and LTE FDD systems define the different characters, performances and applications. For instance, FDD have identical bandwidth for uplink and downlink, which suits the symmetric application. While TDD does not have to assign identical time-frequency resources to uplink and downlink, making it could facilitate asymmetric applications.

Scheduling and acknowledgement mechanism is more complex for TD-LTE system than LTE FDD system, due to its asymmetry character. For LTE FDD system, each uplink subframe could associate with a downlink subframe, therefore the ACK/NACK information could be rather clear with minimum latency. While for TD-LTE system, most configurations are asymmetric. When the uplink subframe excesses to the number of downlink subframes, not all uplink subframe could find an exclusive downlink subframe to transfer its acknowledgment. The opposite holds as the downlink subframe excesses to the number of uplink subframes. Therefore the solution is that for asymmetric scenarios, an uplink subframe could schedule several downlink subframes, and a downlink subframe could schedule several uplink subframes, if there is not sufficient subframes in one direction.

LTE-TDD (TD-LTE) system uses the same frequency bands for both uplink and downlink transmission, and the transmission slots are adjacent to each other, therefore the radio channel environment are highly correlated. eNodeB could estimate the downlink channel condition, based on the receiving uplink information. FDD based system use two separated bands to transfer uplink and downlink information, the frequency dependent channel fading characters could cause channel reciprocity hard to be exploited, unlike TD-LTE system.

TDD system is continuous in frequency domain, but discontinuous in time domain, so it is important to state that LTE-TDD (TD-LTE) could utilize separated and unpaired spectrum resources which could efficiently mitigate the interference to the primary network against LTE-FDD. LTE-TDD does not require paired spectrum, since it will assign all its bands to both uplink and downlink. To distinguish uplink and downlink transmission, TDD system split its time resources, and switch the system operation frequently between uplink and downlink mode. Due to the uplink and downlink transmission utilize the same frequencies, TDD system will assign a guard period between uplink timeslot and downlink timeslot in case that the synchronization is not accurate enough, which help prevent the interferences between uplink and downlink transmission Moreover LTE-TDD (TD-LTE) channel reciprocity could benefit inter cell interference mitigation and multi antenna beamforming technology, which could efficiently increase cell coverage, capacity and cell edge user performance.

Following 3GPP TR 21.915, the main difference of NSA (Non-Standalone Architecture) and SA (Standalone Architecture) is that NSA anchors the control signaling of 5G Radio Networks to the 4G Core, while the SA scheme connects the 5G Radio directly to the 5G core network, and the control signaling does not depend on the 4G network at all. From that perspective it is clear that NSA is a 5G service that depends on LTE (4G) network deployment, hence any optimization effort will have to be applied to both 5G and 4G technologies simultaneously. SA, on the other hand, allows completely independent operation of a 5G service without any interaction with an existing 4G core, hence optimization is one-dimensional.

Following reference “https://www.qualcomm.com/media/documents/files/fierce-wireless-ebrief-5g-release-16.pdf” it is clearly understood that for Mobile Network Operators (MNOs) targeting to eMBB high-speed connectivity services a non-standalone architecture (NSA) makes the most sense, because it enables them to leverage their existing 4G network investments in transport and mobile core, rather than deploy a completely new end to end 5G network. Combining together virtualization and CUPS (Control and User plane separation) using software-defined networking (SDN) MNOs can with minimum efforts reduce network operating costs.

Following reference “https://www.qualcomm.com/media/documents/files/fierce-wireless-ebrief-5g-release-16.pdf” it is clearly understood that MNOs have also the option to build an entirely new fully virtualized 5G network that includes new radio access network, new transport network, and new 5G mobile core fully functionally separated from their existing 4G and legacy networks. 5G standalone architecture is a fully virtualized, cloud-native architecture (CNA) that introduces new ways to develop, deploy, and manage services. With 5G SA, the distinct advantage here is end-to-end support for 5G speeds and services. And the true promise of 5G is enterprise-driven revenue since it offers changes in the business model from consumer-driven to enterprise-focused opening up entirely new use cases and revenue streams.