One of the great things about GSM and 3G was the common set of frequency bands adopted worldwide. While there were a few exceptions, you'd find 2G GSM at 900 and/or 1800MHz with 3G at 2100MHz in almost every country. The US/Canada operated at 1900MHz. This meant that the industry could build and ship the same models of phones that worked everywhere, bringing huge benefits to the entire supply chain resulting in reduced costs, complexity and time to market.
However, the multiple variants of the iPhone 5 indicate a disturbing trend ahead due to fragmentation in LTE implementations.
You might have thought that LTE would consolidate the industry further, having been adopted by both the GSM and CDMA camps – a single, global standard that the world can adopt.
The problem (which we've highlighted before) is one of spectrum choice. Unlike 3G, the industry (including regulators) has not been aligned in co-ordinating specific frequency bands. Responding to their customers' demand, the 3GPP standards body has defined more than 50 different frequency and mode options – far too many to consider incorporating into a single device.
This is most apparent when looking at the recent iPhone 5 device. This latest version of the iconic smartphone no longer comes in a single worldwide model, incorporating both 3G HSPA (used in the majority of the world) and CDMA (used by Verizon/Sprint/KDDI and a few others). No longer can you could switch between operators and/or technologies while keeping the same device.
The latest iPhone 5 comes with three physical variants for different operators:
- A1428 for AT&T which supports LTE in bands 4 and 17
- A1429 CDMA for Verizon which supports LTE in bands 1,3,5,13 and 25
- A1429 GSM for the rest of the world which supports LTE in bands 1,3 and 5
- EE in the UK, who's LTE network will be live before end 2012
- NTT Japan, who's LTE network unusually works at 2100MHz
Much of the reason being given for this approach relates to the level of maturity and integration of the components involved. With such diverse choice of frequencies and modes to choose from, the industry has not yet consolidated and optimised a common core set of choices. With other regions having their own variants, such as South America's choice for 450MHz, it seems unlikely there will be a single common iPhone model in the near future – if ever.
Do operators want it this way?
With high financial costs to operators – one analyst predicts some $10 billion of operator subsidies for the iPhone 5 by end 2012 in the US alone – it's said that some operators aren't too keen on their models being switched across to another network, even when outside the initial contract period. With Verizon and AT&T iPhone 5 models no longer working on each other's networks, that wish has been granted. I recall a similar intent being expressed in India by one Reliance Communications executive some time ago – the Indian network operator knew that even if their CDMA handsets were resold secondhand, they could only be used on their network and so any subsidies wouldn't benefit their competitors.
It seems strange for the industry to invest so many billions towards adopting the common worldwide standard of LTE and then waste the opportunity to drive convergence between operators and countries.
What are the implications for small cells?
Vendors of LTE capable small cells will need to have agile enough designs and manufacturing to be able to ship many different variants to each part of the world. While manufacturing techniques have evolved in recent years to cope with the logistics of this, I still believe that a common/single product is likely to be the cheapest and quickest to market.
The software stack and underlying broadband processing doesn't change much between these different frequency bands. Instead different passive components are required in the RF front end side, with some RF chipsets coping with a wider range of frequencies and bands.
This will inevitably mean LTE small cells being built to order, rather than shipped from stock. Once manufactured, it will be difficult and costly (probably uneconomic) to rework units for a different customer.
By contrast, I recall some years ago when a batch of GSM basestations was ready to ship from the factory while awaiting clearance from the finance department which didn't come through in time. The stock was quickly reallocated to another customer and shipped immediately elsewhere.
For residential femtocells, some may say this will make little difference. These units are often built to order with specific branding and packaging to suit the individual operator. In any case, the LTE small cell market is much more oriented on public access than residential today. For public access and enterprise units, where there may be no branding or differentiation, perhaps this will be more significant.
Locking in the customer, but locking out roaming
Small cells today are locked to one network, and aren't shared. They are used as a differentiator to establish better coverage and performance at each location. This also locks in the customer, giving the operator some financial reward for installing the small cells – whether at home, office or public space.
The trouble with this approach is that if in future, networks decide to share their capacity through national roaming, then this may not technically be supported. Examples are the recent merger in the UK of T-Mobile and Orange, now branded as EE. Where before these were competitors, they are now one business with shared spectrum, network and customers. They were able to turn on roaming between their networks at an early stage so that all customers from both networks could access all coverage and capacity. Telus/Bell Canada have a similar sharing arrangement for their 3G HSPA network. This won't work if devices aren't compatible with other networks due to specific spectrum choices.
It also reduces the chances of LTE being available when travelling abroad. Roaming revenues may be affected, although I suspect fallback to 3G will be adequate for most business users.
A common frequency?
An alternative is for the industry to agree on a common frequency for indoor/short range small cells. In the US, the FCC has given further indication of freeing up a common band at 3.5GHz exclusively for use by small cells.
This would be shared with the Federal government and be less of a "free for all" than Wi-Fi, with some control/co-ordination between cells.
Giving Wi-Fi a potential advantage
Every smartphone has Wi-Fi these days. It may not be the latest 802.11ac standard – the iPhone 5 boasts 802.11n – but these all share the same frequencies and are backward compatible. While I struggle with their utility for public access in locations where there are many competing Wi-Fi hotspots, they do work well in residential and enterprise situations.
Perhaps a more common scenario is when visiting a friend or another business with your iPhone 5, you may find that it won't work with the competing network. You may simply try to use Wi-Fi instead, which shares the same frequency and technology worldwide. With Next Generation Hotspot and related activities making it easier to access Wi-Fi, this may become an easier and/or more attractive option.
More investment in 3G small cells for now
This is why I think we'll see more operator investment in 3G small cells for the next few years. Initially, LTE networks will launch with macrocells to serve the smaller numbers of users attracted to the new service. Strong marketing will support the availability of high speed, high performance networks. Meanwhile, most of the installed base will continue to use (and overload) the existing 3G network. Wi-Fi also has its contribution to make, especially in locations without conflicting hotspots.
Operators will gain most by building out more 3G capacity, which is compatible with their customers' devices, than LTE at this stage because it will give the most visible performance increase to their existing users. In due course, LTE small cells (and multi-mode 3G/LTE/Wi-Fi) small cells will predominate.
So to cut a long story short, LTE risks fragmenting the market. Operators should continue to invest in 3G and support Wi-Fi in the medium term, using small cells in both of these technologies to deliver the substantial capacity and performance under load that their customers demand.