As if there weren’t already enough deployment options for the upcoming 5G network standard, it turns out that for 5G Non-Standalone Option 3, there are actually 3 different variants: 3, 3A and 3X. So what exactly is the difference between them?
The GTI Whitepaper on 5G has a good description and graphics on page 14 so I’ll just summarize the affair here:
As I’ve described in previous posts, 5G Option 3 is the Non-Standalone architecture that still uses the 4G core network (the EPC). In addition, all different flavors of option 3 use the LTE base station as the signaling anchor, i.e. user connection management and radio link control for the combined 4G/5G base station is always managed by the LTE part. This is the same for all options!
At the edge of the EPC, the MME/S-GW are connected over the IP based S1 interface to the combined 4G/5G base station, i.e. the eNB (4G) and the gNB (5G). As the eNB and gNB are logically different components, there is a standardized IP interface between them, the X2 interface. If the eNB and gNB are combined into a single hardware node, the X2 interface is internal, logical, virtual, or whatever you want to call it. So in other words there is the triangle of MME/S-GW at the border of the core network connected to both the eNB and gNB of a combined 4G/5G base station site (over a single physical fiber backhaul connection) and a logical or physical IP interface between the 4G and 5G part of the base station. In other words, there are different options of how data can flow between the core network and the base station and between the two base station parts:
Option 3: In the plain option 3, all data to and from the network flows to and from the 4G part of the 4G/5G base station, i.e. to and from the eNB. The eNB can then decide to forward part of the data to the 5G gNB part of the base station over the X2 interface. This means that the 5G gNB does not communicate with the 4G core network directly!
Option 3A: In this option, both the 4G gNB and the 5G gNB directly talk to the 4G core network but they do not directly talk with each other over the X2 interface. This means that there can be no load sharing of data over a single bearer over 4G and 5G. That means, for example, that the 4G part only handles the VoLTE voice traffic for a user while his Internet traffic is only handled by the 5G part of the base station. I’d say that in most deployment scenarios, this is not really an option if mobile devices move in and out of 5G network coverage continuously.
Option 3X: In this configuration, user data traffic directly flows to the 5G gNB part of the base station. From there, it is delivered over the air interface to the mobile device. It is also possible to forward a part of the data over the X2 interface to the 4G gNB part of the base station and from there to the UE. Slow data streams, e.g. VoLTE bearers with a different IP address than that used for Internet access can be directly delivered from the core network to the 4G gNB part of the 4G/5G base station. The advantage mentioned in the paper linked to above is that the 5G upgrade of the base station is likely to have the much better performing IP interface so it is better suited to handle the higher data rates that can be reached with a 4G/5G Non-Standalone network deployment.
So many options to choose from. That makes me wonder if 5G will become ‘the standard of many options’?