Amit Jain, VP Product Management and Marketing for Spidercloud, shared his current views on Enterprise Small Cells in a recent webinar. Here, we summarise the key architecture choices they’ve taken and explain the reasoning behind them when addressing this rapidly evolving marketplace. This includes where an Enterprise Controller is needed, why 3G/LTE remains important and what’s required to support MulteFire and LAA.
Current State of Play in the Enterprise Small Cell Market
It seems we often hear that “this year” will be the year of the Small Cell, only to be disappointed and told the same story the following year. But the growing volume of Enterprise small cells is masked behind quite large numbers of residential units measured in the millions. There is increasing evidence of an uptake in market demand and deployment.
Part of the reason for the delay has been the initial glut of new spectrum and technology with LTE. Lower frequency bands have allowed existing outdoor cellsites to penetrate in-building and improve coverage. Wi-Fi offload has also helped, mostly for private data use rather than public ad-hoc. Nonetheless, performance inside many buildings remains less than ideal.
So we continue to see strong pent-up demand from the Enterprise sector, with property rental values strongly affected by the quality of cellular service available. Coverage alone isn’t enough – seamless, high speed data access has become the baseline requirement.
3G or LTE or both?
LTE only networks aren’t yet viable in most countries. VoLTE is not yet widely deployed and it’s common to fall back to 3G for voice calls. A substantial proportion of today’s smartphones/handsets are 3G only or support limited set of frequency bands.
While networks consider 3G to be a limited term investment, it’s still an important component of reliable service. The ideal solution today would support both 3G and LTE, with a simple transition to LTE-only in the medium term.
Our audience poll indicated that 75% of viewers thought multi-mode 3G/LTE small cells would be required for at least the next two years, and 43% thought up to 5 years.
The mix of 2G/3G/LTE varies considerably from region to region as exemplified in the chart below from Ericsson’s 2015 mobility report.
It’s also important to understand the implications and timescale of forthcoming developments, such as LAA, MulteFire and shared spectrum at 3.5GHz. The latter two are very suitable for multi-operator/neutral host and all are best delivered using small cells.
One size doesn’t fit all scenarios
Amit distinguishes between three Enterprise market segments based on floorspace.
- Small Office/Home Office. Just over half (53%) of all buildings fit into this category, which is best served by standalone consumer Femtocells and/or Wi-Fi.
- Small and Medium Enterprises (5,000 to 50,000 sq ft). 42% of buildings, best served by Enterprise small cells or indoor picocells.
- Medium to Large Enterprises/Venues (over 50,000 sq ft). Numbering only 5% of buildings but 50% of all floorspace. These justify a standalone network of small cells with an Enterprise Controller or multiple radio heads connected to a dedicated baseband unit.
When selecting the right architecture for each category, the factors to consider include:
- The total cost of ownership per subscriber, involving not just the product cost of the small cells but also installation, any software licences, site lease and backhaul.
- Best use of spectrum. Small cells reuse spectrum efficiently to achieve very high data density, hence the need to consider the number of small cells deployed and average throughput per cell.
- User experience. End users aren’t really concerned with how technology works, they just want a seamless, fast, reliable experience. Data speeds should be consistent; access and hand-in/out should be efficient; value added services easy to use.
Three alternative Enterprise Small Cell architectures
Amit describes three common in-building architectures and explained the differences between them. Broadly speaking, these are
- Standalone small cells, independently connected directly into the central network
- Small Cells with an on-site central controller, which consolidates and optimises the signalling traffic. The controller can also host applications for MEC (Mobile Edge Computing)
- Distributed Radio Heads with a centralised baseband controller. These require higher bandwidth connections to each radio head, shifting more of the low level processing to a central point.
For smaller sites with very few small cells, standalone deployment is the most cost-effective and straightforward. With larger sites, say 50,000 sq feet or more, a dedicated controller is justified for efficient operation.
Shared spectrum requires low level intelligence at the edge
When we look ahead at MulteFire and LAA, which use a modified form of LTE in the unlicensed 5GHz bands used today by Wi-Fi, we need to consider the technical limitations. Each radio head must implement “Listen Before Talk”, where a decision to transmit must be taken in a few microseconds based on whether other transmissions are heard at that time. This can’t be achieved from a central controller where the latency is measured in milliseconds, thus LBT is best implemented within the radio head itself. Amit believes that this requirement strongly favours dedicated small cell architectures rather than remote radio heads.
Our audience thought that LAA would have a significant part to play in the years ahead, with 40% voting for LTE/LAA taking the largest share in 5 years, but a significant number (24%) felt there were too many unknowns to take a firm view.
The webinar itself dives into much greater detail and explains these points in more depth, and includes a Q&A session addressing issues.
Watch the webinar recording (no registration required)