In this blog post, the seventh and final part in a series dedicated to 5G NR drive testing, we look into one of the key enablers of cost-efficient 5G NR deployments, the EN-DC evolved-universal terrestrial radio access-new radio (E-UTRA NR) dual connectivity feature.
In this early phase of 5G, with operators continuing not just to support 4G but to rely on it as part of their 5G rollouts, this is a crucial area of expertise. So let’s take a look.
EN-DC functionality
Cost-efficient initial 5G NR deployments are enabled by three factors: spectrum availability; maximized spectrum usage; and minimized infrastructure costs and technical challenges risks.
These are ensured by deploying NR within the sub-6GHz spectrum, preferably with a DSS feature to maximize spectrum usage, as explained in our previous blog on DSS, and in ‘non-standalone’ (NSA) mode. The latter reduces infrastructure costs while minimizing the technical risks which generally come with new technologies deployments by operating anchored in LTE using CUP (Control-User Plane) split functionality, and the LTE core network (EPC). While the control plane runs through LTE, the user plane runs either on LTE or on NR, or on both LTE and NR links, depending on various factors such as:
- The service’s/application’s required data rate;
- The amount of data to be transferred;
- The traffic load; and
- The radio conditions.
The functionality requires that the device exchanges data between itself and an NR base station (gNB), along with simultaneous connection with an LTE base station (eNB). This is possible when there is tight interworking between the eNB, and gNB has been established over an X2 interface, and when the device can be simultaneously connected to LTE and NR or to LTE for control plane and NR for user plane.
Called EN-DC, this dual connectivity feature allows the UE to leverage the benefits of both LTE and 5G connectivity simultaneously. The two nodes, eNB and gNB, are connected to EPC through an S1 interface.
In addition, initial deployments in mmWave spectrum are likely to use LTE not only as an anchor for the control plane but also as an underlayment for handover scenarios for mobilizing mmWave, which refers to connecting high-density traffic areas where mmWave spectrum is likely used.
EN-DC operation mode
EN-DC operation is reflected in both the control and user plane. In the control plane (CP), to set up and modify the EN-DC operation, the UE needs to comprehend both the LTE and NR RRC control signaling, which is performed by transporting RRC messages between the network and the UE using a set of signaling radio bearers (SRBs): master cell group (MCG) SRB; split SRB; and secondary cell group (SCG) SRB.
In the user plane (UP), the user data is transported between the network and the UE over data radio bearers (DRBs). EN-DC supports MCG DRBs, MCG split DRBs and SCG DRBs. An additional data radio bearer SCG split DRB is introduced in EN-DC. The mobility is controlled by the master node in LTE, while either the master node or the secondary node can control mobility in NR, depending on network configuration.
EN-DC testing aspects (NSA mode)
The non-standalone mode of 5G NR implies a smooth LTE-NR synchronization, which requires simultaneous testing and visualization of LTE/5G NR coverage and interworking, as well as overall performance. Some test use cases are as follows:
- Seamless connectivity to 5G NR evaluation performed by a mobility analysis (e.g. gNB addition/removal as secondary node) with full RRC signaling monitoring and measurement reports;
- Analysis of seamless operation of LTE-NR interaction and performance evaluation of handover (on the UP) to gNB (e.g. leg switch time LTE to 5G NR, number of LTE-5G NR and vice-versa transitions, ping-pong LTE-5G NR, and time spent on LTE and/or 5G NR); and
- LTE and/or 5G NR throughput.
Troubleshooting, and the optimization of NSA configuration coverage and RF performance need to be carried out both in idle and connected mode, evaluating device-based measurements of IE/KPIs, such as RSRP/RSRQ/RSSI, SINR and MAC DL BLER, as well as average distance to serving gNB.
Reflecting a network’s availability and accessibility, these measurements need to be correlated with configuration information – such as rank, modulation coding scheme MCS and bandwidth/bandwidth parts (BWD parts) – in order to diagnose possible DL/UL throughput (a.k.a. network integrity) degradations, as well as to evaluate if the expected achievable performance is reached for the given RF condition and configuration.
In recent work, an analysis of data collected and analyzed with Infovista tools showed several interesting EN-DC performance aspects.
For example, we noticed that although both LTE and NR (EN-DC connectivity-mode) connections show similar RF performance (e.g. within statistical significance similar SINR values), the UP data is not necessarily immediately (or even at all) transferred from LTE to (and/or shared with) NR.
In that specific test case, very few EN-DC connections were made – about half of the total number of apparently feasible opportunities. This result can be explained by analyzing the distribution of the UE Tx power. For the NR connection, the distribution was shifted towards the upper limit, while for the LTE-only connection, the distribution was centered around the average expected values. This indicated and then verified that the events were occurring close to the fringe NR coverage area. This example shows that an in-depth analysis is required to understand the EN-DC functionality and performance.
You can find out more about EN-DC feature testing in our white paper, Initial 5G NR drive testing with Infovista.
Key takeaways
Throughout our blog series on initial 5G NR drive testing we’ve talked about the disrupting technologies that are enabling NR, the challenges that arise from them, and how all these result in significant changes needed for testing methodologies and use cases when compared to LTE.
We’ve discussed how the importance of device-based measurements (a.k.a drive testing) has to grow due to user-centric design characteristic of 5G, but it also has to significantly evolve towards automation and prediction.
And we’ve walked through some initial test use cases, concluding the blog series with this piece, which discusses an important enabler for a cost-effective initial NR deployment, the EN-DC feature. Without EN-DC optimally operating, there cannot possibly be any fast and smooth NR deployment, which is very much desired by all telcos. This can be achieved by only using the right testing tools.
To read more on the broader topic of NR drive testing, download our white paper, Initial 5G NR drive testing with Infovista.
