The Telecoms.com Podcast: OpenRAN, Ericsson and retail tech

Mike Ferris 07/12/2021 @ 1:01 pm

I’m a bit late listening to this, admittedly, but there’s one point Ian made that I have to disagree with. There’s nothing stopping Ericsson from developing a RIC type function in their closed RAN solution. It’s only an extension of existing SON capabilities. Why do they need 3GPP to write a standard for it? Standards are only needed to enable interoperability, and you don’t need interoperability if the RAN is closed anyway. You need it when you talk to the core or the handset. 3GPP focusses on what’s needed to get the ecosystem to work across many vendors. Vendors have historically adopted the standards but still added some of their “special sauce” on top where the vendor lock in allows them to do it. That’s why nobody mixes traditional RAN elements in a real network deployment, not lack of a standard. It’s the same reason CPRI (which is open, by the way) on its own, is not enough to get an RRH from one vendor to work with a BBU from another. The industry is littered with examples of lack of interoperability between vendors where the protocol between devices is completely open. It’s a BAU problem. What’s different about O-RAN is that the whole premise is that components must interoperate fully to create this healthy ecosystem of alternative vendors, each with their own specialty. Having said that, it was a great debate you guys were having about it, and I enjoyed it, as usual.

One comment

Mike Ferris 07/12/2021 @ 1:01 pm


I’m a bit late listening to this, admittedly, but there’s one point Ian made that I have to disagree with. There’s nothing stopping Ericsson from developing a RIC type function in their closed RAN solution. It’s only an extension of existing SON capabilities. Why do they need 3GPP to write a standard for it? Standards are only needed to enable interoperability, and you don’t need interoperability if the RAN is closed anyway. You need it when you talk to the core or the handset. 3GPP focusses on what’s needed to get the ecosystem to work across many vendors. Vendors have historically adopted the standards but still added some of their “special sauce” on top where the vendor lock in allows them to do it. That’s why nobody mixes traditional RAN elements in a real network deployment, not lack of a standard. It’s the same reason CPRI (which is open, by the way) on its own, is not enough to get an RRH from one vendor to work with a BBU from another. The industry is littered with examples of lack of interoperability between vendors where the protocol between devices is completely open. It’s a BAU problem. What’s different about O-RAN is that the whole premise is that components must interoperate fully to create this healthy ecosystem of alternative vendors, each with their own specialty. Having said that, it was a great debate you guys were having about it, and I enjoyed it, as usual.