The looming spectrum shortage: worse before it gets better
Deloitte predicts that although additional spectrum will continue to be made available in 2013 in many global markets, spectrum exhaustion will continue to exacerbate in many countries, especially in dense urban areas. End users will continue to see performance impacts as a result, primarily in the form of lower speeds, but also through inability to access networks and dropped calls or sessions. The reason is simple demand for spectrum will exceed supply. Demand for wireless bandwidth continues to grow in leaps and bounds, but supply is relatively constrained. By 2014 the US alone may suffer a 275 MHz spectral “deficit”1 .
To be clear, a spectrum shortage is highly analogous to a crowded highway: it doesn’t just “stop working” like an electrical grid that goes down in a storm with no power to anyone across wide areas, sometimes lasting for days. Instead, the likely outcome of the predicted shortage will be most intense in cities, on certain networks (those with the most subscribers) and in peak wireless hours. Users can expect wireless ‘rush hours’ to be characterized by two to three times as many failed attempts to connect, three to four times as many dropped calls or frozen web browsing, and both 3G and 4G speeds 50-90 percent lower than expected. In the worst situations, download speeds may be under 1Mbit/s for lengthy periods of time, making video streaming impossible and even web browsing difficult.
The cellular device market sends and receives in the portion of the electromagnetic spectrum ranging from 600 MHz to 3600 MHz. These bands are strictly regulated by national governments and allocated for specific purposes. Spectrum is like land: no more can be made, it is difficult to share and not all spectrum bands are created equal. 900 MHz is the spectral equivalent of beachfront property: transmissions on that frequency band go further, penetrate buildings better and have good capacity. On the other hand, 3500 MHz is a bit like desert scrub land: radio waves in this band have shorter range, poor in building performance, and are even vulnerable to bad weather.
The demand for additional spectrum is tied directly to the seemingly insatiable consumption of wireless broadband communications. Wireless traffic has more than doubled each year since 2009 and the increasing penetration of smartphones and tablets only serves to exacerbate the problem2 . Today, the average smartphone drives 35 times more traffic than a typical cellphone. It is expected that by 2016 wireless traffic will have increased 50 fold from 2012.
So why can governments not just increase the “supply” of spectrum? The supply of radio spectrum can be improved by only two methods.
The first is accomplished through the allocation or reallocation of frequency bands to operators. As the distribution of spectrum can have major economic implications for competition and accessibility, many governments put significant effort into developing and governing the models used for assigning spectrum. For the last decade, the auction approach has become most prevalent, both as a source of government revenue and a relatively transparent method of allocating scarce and valuable resources. However, these auctions often incorporate additional rules to encourage new entrants, provide coverage of rural and lower income areas, provide support to aging public safety communication networks, and mandate that a certain percentage of the population be covered. While auctions allow additional capacity to be made available, such as the repurposing of analogue TV bands occurring in many countries, they do not occur rapidly, and are lagging behind surging demand.
The second method is to make more efficient use of limited spectrum. Fourth generation (4G) technologies such as LTE have substantially improved spectral efficiency. LTE is almost 16 times better than 3G at moving a bit of data over a Hertz of spectrum. However in the seven years it has taken to develop and widely deploy this new technology, wireless traffic increased 30-fold. Telecommunication equipment vendors simply can’t invent new technologies fast enough to meet growing demand. The emerging LTE-Advance standard expects to further double3 spectral efficiency over LTE. This is wonderful, but results in adding less than a year of additional capacity at current growth rates.
Emerging technologies, such as Heterogeneous Networks or HetNets, have the potential to address some of these concerns. HetNets consist of a series of wireless access layers, protocols, and equipment allowing mobile devices to seamlessly move between wireless networks of various types. Voice calls and data sessions can be maintained without interruption as devices move between macrocells (covering dozens of km), microcells (covering kms), picocells (100s of meters) and femtocells (tens of meters) and back. Emerging technologies and standards can extend HetNets across WiFi, Mesh and Ad-Hoc wireless networks as well.
As little as two years ago, this lack of portability was not a significant issue for most end users. Mobile devices connected to the cellular network were typically used in such a way that virtually all traffic was managed by a traditional macrocell, usually located on a tower. The problem is that moving between these networks is sometimes not transparent to either the end user or the network provider. Users may need to: manually identify and select a different network (either microcell or wifi); provide necessary credentials to authenticate onto the new network; and re-establish a session with the application.
HetNet is based around more intelligent devices and networks that can monitor the current wireless environment for available networks and single quality and, when appropriate, automatically select, authenticate and hand over current sessions without user intervention.
At this time, some of the technologies needed to deliver HetNet services have yet to be widely deployed. Further, HetNets require changes to the end user device, access points and the network core making adoption more complex and expensive.
Internet Protocol Version 6 (IPv6) is the successor to the widely used IPv4, which has “only” 4.2 billion unique addresses. Almost all of those are in use, requiring addresses to be re-used. While that helps mitigate the shortage of addresses under IPv4, it means that new addresses must often be dynamically assigned, making it difficult to maintain existing session information and to determine exactly where devices are located. IPv6 has 1038 addresses (or enough to give every star in the known universe a trillion IP addresses) and will be able to give each device a unique identifier, and simplify the handover processes.
VoLTE (Voice over LTE) is another set of technologies and standards that will enable HetNet by allowing voice traffic to be carried over 4G networks. Today, most 4G networks use LTE for data and fall back to 2G and 3G networks for voice. This increases the complexity of moving calls between networks as there may be multiple voice and data sessions than need to be managed using very different methods and technologies. VoLTE handles voice calls as another data session (containing audio information) allowing much easier movement between networks.
The standards behind HetNet have been under development for several years (the IEEE 802.21 working group was established in 2004). However, since HetNets span networks defined by multiple standards bodies (including IEEE, 3GPP, 3GPP2, ITU-T and IETF) a number divergent attempts at standardizing network interoperability have occurred delaying widespread adoption. Although current initiatives have begun to show progress (such as 802.11u) there is still much activity. The Wi-Fi Alliance’s Hotspot 2.0 program4 began administering the Passpoint™ certification process in June 2012, which covers mobile devices and hotspots that automatically select and authenticate access to Wi-Fi networks using a devices SIM card. At present, only a limited number of certified devices are available. In parallel, the Wireless Broadband Alliance, as part of its own Next Generation Hotspot initiative5 , is working closely with the WiFi Alliance to validate certified devices in real world conditions. Phase 2 trials with several global carriers began in Q4 of 20126 . It is expected that many carriers are waiting on the outcome of these trials before making significant investments in HetNet related infrastructure.
While strong progress is being made towards making HetNet services a reality, it may take most of 2013 to resolve these challenges. Foundational technologies will continue to be rolled out, standards compliant equipment will become widely available and business concerns will be ironed out. Some markets will see the introduction of limited HetNet capabilities and limited pilots. Additional acquisitions in the WiFi service provider and equipment market are likely as lagging carriers and manufacturers look to quickly build their footprint or expand their product offerings.
There is another approach, called cognitive radio, where the device detects all parts of the wireless spectrum and dynamically alters its transmission or reception parameters according to which bits are currently not being used. This allows much more data to be sent over a given spectrum band at a given point in time7 . Also known as dynamic spectrum management, it does work in labs today. But it is likely years to decades away from adoption.
While progress is being made towards making additional spectrum available, and considerable effort is being made to improve spectral efficiency, demand for wireless bandwidth will likely attempt to outstrip these improvements in supply for at least several years. Major metropolitan areas in some geographies should expect to see continued deterioration in end user experience. The other alternative is that carriers may want to increase what they charge for data and speeds: if spectrum truly is a scarce resource, then using price to signal its value is likely to reduce demand to the point where service standards do not fall.
Regulators may wish to accelerate and streamline their spectrum allocation process. Although auctions are an equitable and transparent process that also raises money for the treasury, the process around them can take years or even decades. Further, they can look at allocating larger spectrum blocks (the blocks have tended to be somewhat fragmented) and encouraging solutions that promote spectrum sharing, particularly at the higher frequency bands.
Carriers will likely want to do even more with wifi, as well as find picocell and femtocell business models that lead to more rapid adoption. One possible strategy is that instead of having consumers pay for the small cells (which they have been loath to do) carriers can respond to complaints of poor coverage by paying for the femtocells themselves, and thinking of it as a customer retention tool and associated cost. Any coverage of adjacent areas and cellular offload is just gravy.
Spectrum isn’t just needed for smartphones and tablets: as 4K TV rolls out, TV broadcasters may want to get back some of the spectrum that they gave away in the transition from analogue. Although compression is likely to improve the amount of bandwidth that 4K broadcast signals will require, it is unlikely to provide a true 4K signal in the current spectrum allocated for HD digital.
Ironically, in the short term, some customers may experience improved voice performance as delays in implementing VoLTE (Voice over LTE) will allow data traffic to migrate to 4G networks, while freeing up 3G networks to more effectively carry voice traffic.
There are some cities where macro-cell sizes are as small as they can usefully be: rooftop antennas and towers cannot be spaced any more closely. But that is not the case in all areas – sometimes the local resistance to new antennas is such that it can take years to erect a new tower8 . Streamlining the cell site approval process -- while continuing to allow for citizen input, of course – would help reduce some of the impact of spectrum scarcity.
Finally, along with cognitive radio, smart antenna technology with variable gain can correct for certain inefficiencies by directing signals toward devices generating or consuming traffic. In effect this shrinks the cell site by only occupying the spectrum in the direct line of sight between the tower and device. As other devices consume traffic, they can share that same spectrum by also taking advantage of the directionally focused antenna9 .
Source: Airwave overload? Addressing Spectrum Strategy Issues That Jeopardize U.S. Mobile Broadband Leadership, Deloitte Development LLC, September 2012. See: www.deloitte.com/us/spectrum
1 Source: Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2011–2016, Cisco, 14 February, 2012. See: http://www.cisco.com/en/US/solutions/collateral/ns341/ns525/ns537/ns705/ns827/white_paper_c11-520862.pdf
2 Source: 4G LTE-Advanced Technology Overview, Agilent Technologies. See: http://www.home.agilent.com/agilent/editorial.jspx?cc=IN&lc=eng&ckey=1905163&id=1905163
3 Source: Wi-Fi CERTIFIED Passpoint, Wi-Fi Alliance. See: http://www.wi-fi.org/discover-and-learn/wi-fi-certified-passpoint%E2%84%A2
4 Source: Next Generation Hotspot (NGH) Program, Wireless Broadband Alliance. See: http://www.wballiance.com/wba-initiatives/next-generation-hotspot/
5 Source: Major Telcos to Trial Next Generation Hotspots Using First Commercially Ready Equipment, 26 June, 2012. Wireless Broadband Alliance. See: http://www.wballiance.com/2012/06/26/major-telcos-to-trial-next-generation-hotspots-using-first-commercially-ready-equipment/
6 Source: Cognitive Radio, Wikipedia. See: http://en.wikipedia.org/wiki/Cognitive_radio
7 Source: Airwave overload? Addressing Spectrum Strategy Issues That Jeopardize U.S. Mobile Broadband Leadership, Page 32, Deloitte Development LLC, September 2012. See: http://www.deloitte.com/assets/Dcom-UnitedStates/Local%20Assets/Documents/TMT_us_tmt/us_tmt_Spectrum_Thought_Leadership_September_092512.PDF
8 Source: Airwave overload? Addressing Spectrum Strategy Issues That Jeopardize U.S. Mobile Broadband Leadership, Page 23, Deloitte Development LLC, September 2012. See: http://www.deloitte.com/assets/Dcom-UnitedStates/Local%20Assets/Documents/TMT_us_tmt/us_tmt_Spectrum_Thought_Leadership_September_092512.PDF