Evolution of Transport for 5G

4G Transport Options Today

Ethernet Based Backhaul

4G transport options are tied to base station functionality and operator’s deployment scenarios. Traditional base stations are decentralized and have relatively low capacity and latency requirements. Although there are many scenarios were traditional base stations are desirable, operators are looking at ways to reduce costs and optimize performance. As the RAN is evolving to become greener and more profitable in many cases Ethernet alone does not meet RAN evolution.

Fronthaul

The use of C-RAN Fronthaul has allowed operators the centralization of all high layer processing functions. This centralization provides significant energy and labor savings simplifying the deployment of features like Carrier Aggregation (CA) and Coordinated Multi-Point (CoMP). This capability of C-RAN networks is achieved by using high capacity and low latency transport networks. C-RAN fronthaul connects the RF and the remaining L1/L2/L3 functions using CPRI. CPRI transports digitized time domain IQ data with data rates dependent on the number of antenna ports. The use of C-RAN Fronthaul has allowed operators the centralization of all high layer processing functions. This centralization provides significant energy and labor savings simplifying the deployment of features like Carrier Aggregation (CA) and Coordinated Multi-Point (CoMP). This capability of C-RAN networks is achieved by using high capacity and low latency transport networks. C-RAN fronthaul connects the RF and the remaining L1/L2/L3 functions using CPRI. CPRI transports digitized time domain IQ data with data rates dependent on the number of antenna ports. The figure shows deployments for 2 to 8 antenna ports (more common for 4G deployments) for multiple base station channel sizes. It is anticipated that the deployment of massive MIMO for 5G networks will require in many cases up to 256 or more antenna ports using larger channels. A comparison of both scenarios shows that CPRI does not scale well for 5G deployments with multiple antennas and wide channels.

Figure 1 Required Fronthaul (CPRI) capacity for 5G network

5G Transport Options

The solution of this problem requires a rethink of the functional split used for RAN and C-RAN deployments. The main change in 5G is that the original Base Band Unit (BBU) in 4G/LTE is 5G is now split into three parts as defined in TR 38.801:

  • Central Unit (CU)
  • Distributed Unit (DU)
  • Remote Radio Unit (RRU)

This redesign of the BBU functionality is motivated to facilitate RAN virtualization and reduce fronthaul capacity and latency requirements. This redesign allows to categorize transport based on the different capabilities of different base station functional splits (For more on functional split see 3GPP TR 38.801):

  • Backhaul: Connection from gNB (Supports connectivity to new radio and core) to 5G core
  • Midhaul: Connection between CU and DU
  • Fronthaul: Connection from gNB to transmitter (RRU)

Deployment Examples

The main drivers for this new RAN design are to reduce the fronthaul capacity requirements, and meet latency demands. This new architecture also makes virtualization easier and allows for a 1 to N relationship between the CU and DU. This architecture is more flexible for operator’s deployment scenarios some examples of possible deployment scenarios include:

Figure 2 Co-located CU and DU

  • Co-located CU and DU: For this deployment scenario there is only fronthaul and backhaul as CU and DU conform one block. This scenario is like traditional C-RAN.

  • Independent RRU, CU and DU locations: In this scenario the connection between the transmitter (RRU) and CU is high capacity and low latency fronthaul. The linking between the CU and DU is midhaul and the connection between the CU and the core is backhaul.

Figure 3 RRU and DU Integration

  • RRU and DU integration: This scenario is most common for in building deployments were CU can connect using fiber to an integrated RRU and DUU.

Figure 4 RRU, DU and CU integration

  • RRU, DU and CU integration: This scenario is more like the current deployments of macro stations or small cell deployments using an integrated base station.

Evolution of Transport for 5G: Base Station Functional Splitting

Figure 3 shows the 3GPP TR 38.801 functional splits. The lower the split point (towards 8) the higher the requirements to transport network and the more difficult to virtualize the CU. Also, lower splits also have more coordination and centralization capabilities and more cost-effective DU.

Figure 5 3GPP base station functional splits

Figure 4 shows an example of the possible capacity requirements for the assumptions shown in table 1 (from 3GPP TR 38.801 Annex A). Latency requirements are for options 1-5 of about 1~10 milliseconds except for option 4 DL that is 100 µseconds. For options 6 and 7 latency requirements are 250 µseconds (same as CRPI). This scenario assumes a base station channel size of 100 MHz (DL/UL), modulation od 64QAM (DL/UL) and 8 MIMO layers and 32 antenna ports (for more details see table A2 form Annex A 3GPP TR 38.801.

For this example, backhaul and midhaul have similar capacity requirements with greater differences between the fronthaul options 6 and 7-1, 7-2 and 7-3. It is significant that midhaul and backhaul do not support centralization functions making fronthaul options the transport choice for advanced features such as CA and CoMP.

Conclusion

The new 5G RAN provides more transport flexibility and allows different base station functional splits. This flexibility is necessary to support operator’s different deployment scenarios and allows to have more transport options. The definition and scoping of functional splitting is evolving rapidly and we should pay close attention to this as it should form the basis of transport systems for the 5G networks of the future.


Leave a Comment

Subscribe to our newsletter