As operators around the world tap into the huge benefit that IP and Ethernet offer the Radio Access Network (RAN), a question has come up - what designates an IP RAN?
I use the IP designation for any RAN or RAN Backhaul network that leverages Ethernet and/or IP equipment anywhere inbetween a base station and the BSC or RNC
My argument is that once IP and Ethernet enter any part of the RAN (or RAN Backhaul), the operator must have access to IP savvy staff, procedures, etc. to provision, monitor, operate, upgrade RAN, and that steps beyond tradition of TDM circuit based procedures and knowledge.
Others argue that the IP designation comes only for an All IP RAN - i.e., 4G and LTE.
So when to call a RAN an IP RAN?
This is interesting discussion.
It would help to differentiate IP RAN vs. IP (or Packet) based Backhaul. These two terms are sometimes used interchangeably, are sometimes interpreted based on the context or are sometimes misunderstood etc.
In strict sense:
IP RAN means when the end-to-end system is IP based, including radio, RNC (in 3G) or the packet core (in LTE). Those are the end systems and they need to be able to speak IP and support TCP/IP OSI layer stack. For example - IP RAN in 3G/UMTS, the NodeB and Radio will have to be upgraded to support ethernet and IP directly to be called IP RAN.
IP Backhaul are scenarios where the backhaul is IP Transport irrespective of what the end-points are using. Some of the scenarios are described in this post and the earlier one. I also want to use another term for this - Packet Based Backhaul.The fundamental to IP Backhaul is that this uses IP and packet technology and is not so much dependant on Layer-1 and Layer-2. And the RAN end-points like radio and RNC/Core may or may not be IP enabled.
So the examples of psuedowires for carrying 2G and 3G voice/data or MLPPP based on 3GPP2 or pure IP routing or MPLS L2/L3 VPNs - these all are IP Backhaul. The nice thing is that the packet/IP backhaul is not dependant on Layer-1 [wired (copper, optic), uWaves] and Layer-2 [Ethernet, Sonet/SDH, Serial/PPP/MLPPP].
Yes. I believe that this question is more relative to the marketing name given by different vendors that sometimes generate confusions.
Besides GSM originally is based on TDM, CDMA in Frame-Relay, 3G W-CDMA in ATM and only WiMAX started from the ground in IP, most of the 3GPP different generations of technologies started to be moved to the IP direction as the LTE is coming soon.
So, some GSM/3G vendors started to call IP-RAN to the RAN products (BTS / Node_B) that are ready to deliver the traffic and signaling directly to Ethernet/IP backhaul without the addition of any device to convert from E1/TDM or ATM to Eth/IP. This represents a interesting reduction to the CAPEX per site.
The pressure on the market for that move started around 2007 and depending on the roadmap of the vendors the products and upgrade to legacy BTS/Node_B arrived between 2008-2009 for most vendors.
As you point out, IP/Ethernet based networks offer significant savings for RAN Backhaul costs. They're also proven to offer significantly higher capacity. I'll post another discusion to highlight samples of that.
I like your point about the early segmentation of radio choice with network foundation
Cisco's "Unified" RAN Backhaul solution is deployed to support Any Radio, over Any Media, for Many Generations. We use IP/Ethernet extensively in the solution, even for RAN networks that still have lots of TDM and ATM, supporting backhaul between "2G" BTS and BSC. Officially, the solution name is "Unified RAN Backhaul" ... but you'll see people refer to it as "IP RAN", since it is leveraging IP technology, and is built to optimize the transformation and adoption of IP across the entire network, eventually.
Naming nuances aside, the industry agrees, the RAN is destined for IP.
Thanks for your feedback