It’s become accepted throughout the industry that while Enterprise Small Cells can cater for moderately sized buildings with moderate traffic levels, you really need DAS for the largest and busiest environments such as stadiums and airports. Does the transition to LTE and C-RAN present a disruptive challenge to this traditional thinking?
What are the practical limits of a Small Cell site deployment?
High performance wireless service is increasingly important in large public venues. Stadiums, malls and airports are examples where users expect reliable and consistent cellular service. It’s no longer enough just to deliver voice and text, users expect to make full use of a wide range of apps and related data services as part of their 4G plan.
The traditional approach for this type of large public venue has been to deploy a full DAS solution, with project budgets stretching into millions of dollars per site. Most large venues already have some 3G DAS in place, and will be considering what’s involved in upgrading to add 4G. Typically, this would be a major and expensive undertaking, requiring costly redesign and re-engineering of the 2G/3G system.
Until recently, small cell solutions have been discounted for these large venues. Concerns raised include the lack of co-ordination between cell boundaries, whether each cell can deliver enough capacity, and duplication of equipment to support multiple network operators.
Some small cell vendors argue that techniques such as centralised baseband control (C-RAN) allow them to meet the technical performance required without the high cost of a major DAS upgrade. This could fundamentally shift the industry perception of where performance boundary between the small cell and DAS lies.
What would it look like?
Airvana’s OneCell, an LTE small cell solution designed for larger enterprises and buildings, comprises a central Baseband Controller and multiple radio nodes called Radio Points which together form a single cell, not unlike a DAS with its central base station, distributed radio heads and multiple antennas. However unlike DAS, it has intelligence in the endpoints that enables, for example, uplink and downlink signal combining and higher-order MIMO for improved signal quality and device battery life.
Typical DAS designs split a large site such as a stadium into anything from 20 to 50 sectors, each operating as an independent physical cell and driven by a macrocell carrier module. Each sector can have up to 20 antennas, so that there are several hundred individual antennas installed. However, these sectors are statically defined and there is no coordination between them, so interference between them typically reduces performance to 0.5bps/Hz.
By contrast, a C-RAN system will actively control each radio points and can use the same physical cell identifiers (PCI codes) across them. Although it is a single PCI, virtual sectors are created each millisecond based on where active users are located to increase capacity. This dynamic coordination across the radio points results in lower interference and higher average spectral efficiency of around 6 bps/Hz on the downlink, and roughly 2 bps/Hz on the uplink..
The outcome is that where a typical DAS deployment might have 32 or more physical sectors, OneCell could achieve the same throughput with far fewer. Unlike DAS, radio coverage overlap between adjacent radio nodes can be a benefit, because the Radio Points can coordinate with one another, allowing a better user experience with faster data rates.
As an example, the 68,000 seat Seattle Seahawks stadium recently installed a very large DAS to accommodate usage in the 2014 National Football Conference championships. The table below compares the DAS as described in the Seattle Times to a OneCell C-RAN small cell system designed for equivalent capacity.
|DAS (2014)||C-RAN Small Cell (OneCell)|
|Number of Antennas||700||300|
|Number of Physical Sectors||47||8|
|Number of Virtual Sectors||Not applicable||32|
Sources: At CenturyLink Field, Seahawks fans want a win and a signal, Seattle Times, 18 January 2014, and Airvana estimates.
This example illustrates the potential the C-RAN small cells bring to simplify network designs for large installations.
Dealing with multi-operator
Stadiums and large venues are very much biased towards multi-operator deployments. DAS systems are inherently designed to cater for multiple network operators, whereas most Small Cells are locked to a single network. However, OneCell supports multi-operator deployments by providing operator- and frequency-independent Baseband Controllers and Radio Points that can be provisioned for particular operator or frequency band in software. In addition, multiple operators can share common IP/Ethernet cabling and fronthaul.
Choice of DAS replacement or Small Cell overlay
Many stadiums have installed DAS systems over the last 3-5 years, designed primarily for the peak demand of 3G traffic during that time. There’s now growing demand to cater for LTE. We’re seeing DAS designs with 40 or more physical sectors to handle a 100,000 seat stadium. Upgrading the entire DAS to handle LTE is a significant change to the 2G/3G footprint, with a full re-design. Many more radio nodes would be needed for LTE than 3G but all DAS nodes would carry both. This creates a lot of cell boundaries, requiring a lot of parameter tuning and technical design work to make it work well. It also needs a lot of macrocell equipment from each operator connected to it, taking up physical space and consuming significant amounts of power and air conditioning.
An alternative is to leave the DAS system in place for 2G/3G service and install a separate small cell overlay specifically for LTE. Design and deployment of a single LTE overlay would be less disruptive, lower cost and be quicker to go live. This approach also extends the lifetime of the existing 3G DAS system, improving return on investment. U.S. mobile network operator Nex-Tech Wireless recently used OneCell in this overlay strategy to bring 4G service to a coliseum in Kansas without disrupting the existing 3G DAS.
Why not just invest in Wi-Fi?
Wi-Fi is an important part of the solution but not a complete one. This UK example reported that 36% of spectators had Wi-Fi turned on, and of these only 35% connected to the Wi-Fi network meaning that fewer than 13% of all spectators were actually served by the Wi-Fi network.
Many smartphone apps are configured to initiate backups, synchronise photos, download updates etc. only when connected to Wi-Fi. This skews the traffic profile between cellular and Wi-Fi users.
The outcome is that a substantial proportion of the overall data traffic served at a stadium is carried over Wi-Fi [e.g. Superbowl 2015 reported 6.23TB Wi-Fi vs 6.56TB cellular], but much of it isn’t urgent and could be deferred. The top applications by bandwidth used in the 2015 Superbowl were Apple (mobile updates), Facebook, Dropbox and Snapchat. Arguably, the Apple and Dropbox traffic had little to do with users’ experience at the stadium. The cost of that service, including the broadband internet connection, is almost entirely paid for by the venue owner.
It’s not just stadiums
The case put forward for a separate LTE small cell overlay isn’t constrained to sport stadiums. It’s equally applicable to any large venue including airports, shopping malls, concert halls etc.
While I don’t expect DAS systems to be replaced overnight by small cells, there is now a clear choice for operators to make in the transition to LTE. The technical proficiency and capability of small cells with centralised baseband controllers provides a compelling alternative for delivering LTE capacity and performance.
Our thanks to Zach Lovell and Sohil Thakkar, Airvana Senior Product Managers, for sharing their insights during the preparation of this article.
Airvana are a sponsor of ThinkSmallCell.