Firstly, lets remember that the Femto Forum is an industry group lobbying for and promoting femtocell adoption. It consists of both vendors and operators with a remit to present a united front and common story. Up to now, its been quite friendly, but we've heard that as the operators insist on a common standardised solution, the meetings are more lively because each vendor has invested in their own versions and any differences cost time and money. The 3GPP is a standards organisation comprising members from all the main regional standards bodies. It aims to reach a consensus position with well argued and reasoned debate, and it issues and approves the specifications adopted by the regulatory authorities around the globe. Therefore, it is really what 3GPP specifies that determines what will be accepted in volume.
Up to now, the consensus has broadly agreed on three main femtocell system interfaces:
- Compliance with the 3G radio interface to mobile phones. Femtocell vendors promote the advantage of working with existing 3G phones and are compatible with earlier “Release 99” specification models, through to the latest HSDPA smartphones capable of up to 7Mbit/s data transfer rates. This has pretty much always been the case and isn’t changing. We can expect future compatibility with later releases of the 3G specification which include higher speed uplink rates (although this might not be of much use with a broadband DSL line today), improved power consumption and lower latency. However, some optimisations have been agreed which will improve the femtocell experience when used with later “Release 8” handsets incorporating these changes.
- Compliance with TR-069 remote management systems. Pretty much all femtocell vendors have announced compliance with this standard, which originated from the DSL Forum. It provides software updates, configuration and diagnostic capability. Operators can therefore manage several different femtocell vendor’s products (and DSL modem types) from the same management systems.
- Compliance with the Iu interface: This connects mobile operators' Radio Access Networks (RAN) to their core switching systems (the voice switches and data routers). Whilst there is a small band of enthusiasts for a “next generation” core network interface directly using the SIP protocol, this would be compatible with neither existing mobile phones or existing core networks. Operators often buy their core networks and radio access networks from different vendors, hence are used to integrating at this interface point.
What has been introduced is a new interface point called the Iu-h, which is between the femtocell and the femtocell gateway. The purpose of this is to allow operators to purchase femtocells from different suppliers and connect them to the same femtocell gateway. This avoids lock-in to any one vendor and thus gives them more opportunity to drive prices down and maintain innovative new feature development. There is also a new protocol (HNBAP - Home Node B Application Protocol) which defines the IP signalling messages sent between these two nodes.
From analysis and interpretation of the documents from recent standards meetings, my understanding is that the scope of the proposal approved:
- will require some changes from all vendors
- is not directly based on UMA
- will require some subscriber provisioning in the femtocell gateway, breaking a long standing architectural principle that subscriber specific information was not held in the RAN (Radio Access Network).
- still needs a lot of technical detail to be specified and agreed, with a target completion date of December 2008.
The 3GPP standards body uses different names for femtocells. This might not be a bad thing - we’ve all seen those press articles introducing the topic, and usually starting with some joke about the name being confused with prison rooms for ladies. The new exciting names introduced into the formal specifications are Home Node B (HNB) and HNB Gateway (HNB-GW). There’s also a subtle naming difference with a Home eNode B, which seems to be reserved for 4G LTE based systems. Whilst some vendors have dreamt up some good brand names for their femtocells (I think Airvana’s HubBub wins my vote), standards organisations really aren’t the place for imaginative new product names, so lets hope that's where these ones remain.
[Ed Note: By the way, when 3G standards first developed they couldn’t agree on the names for the basestation controller and basestation, so they called them Node A and Node B pending agreement on better names. Node A became the RNC – Radio Network Controller – but a better name for Node B was never agreed, so it has stuck ever since].
Some enhancements to the handset have also been agreed in principle in subcommittee – these will need to be specified with all the technical detail and approved by the committees. A few of interesting ones are:
- Incorporating a “white list” of preferred femtocells into the handset. One problem has been telling the handset which other basestations to keep monitoring for use with handover. In a normal outdoor cellular system, each cellsite continuously broadcasts a list of suitable neighbours. In recent years, these lists have had to be expanded beyond the original limit of 32, but clearly can’t cope with the hundreds or even thousands of femtocells that could be co-located with an outdoor cellsite. Additionally, there wouldn’t be time or battery power to scan large numbers of potential handover candidates. Instead, the handset will be provisioned with a short list of potential femtocells which it can check and quickly handover to as you enter a femtocell zone.
- Allocating a set of CDMA scrambling codes for use by femtocells. This is similar to allocating a set of frequencies within the spectrum under the control of an operator. The codeset will be broadcast by the neighbouring outdoor cellsites.
- Each femtocell can broadcast a text string with its own “identity”, similar to that available from today’s WiFi access points.
There's still a lot of technical details to be worked through, but once the principles have been agreed this can often be done in months.