Enterprise Femtocells: Too Much of a Good Thing?
Everybody is talking about femtocells as if they’re some sort of savior for in-building cellular communications. Femtocells were developed to provide spot coverage and capacity in homes and small offices, and many enterprise users and pundits talk as if we can simply deploy a bunch of the devices to cover a larger office. It’s not that simple.
Fundamentally, a femtocell was designed to provide radio capacity and coverage over a specific area, like a Wi-Fi access point (AP). To cover a larger area, it would be necessary to deploy more femtocells, just as wireless LAN architects use multiple APs to cover large areas. But the mobile network is not like a private, unlicensed wireless LAN, and using multiple femtocells to provide coverage within an enterprise can cause a lot of problems. Here’s a list of the potential pitfalls:
Interference – Like wireless LAN APs, femtocells must be tuned and their deployments engineered to minimize interference with other femtocells. This means minimizing coverage “overlaps.” Where interference between femtocells occurs, performance can be comprised. But femtocell users also must worry about interference with the outside macro network. For example, femtocells must be placed close to the outer walls to ensure strong coverage along the inside building perimeter. However, due to “bleed out” to the outside, these femtocells will likely interfere with the external macro network – an event that is is rarely tolerated by wireless carriers – and it would be the enterprise’s responsibility to correct the problem.
Signal Dominance – One way to minimize hunting among signal sources is to establish a dominant signal so that the cellular device only hunts at the edges of a cell. But since femtocells have very low output power (milliwatts compared with 20 watts for a macro cell), users near the external walls may find their devices locking onto the macro cellular signal outside the building rather than the nearest femtocell. This defeats the purpose of having femtocells in the building to begin with.
Handoffs – If multiple femtocells are used to provide service for adjacent interior areas, users’ handsets must hand off the connection from one cell to the next as they move through the building. Handoffs reduce handset battery life (and thus the usefulness of the device), and they will occur much more frequently than in the macro network because of the femtocell’s small cell size.
Over-provisioning – Each femtocell provides capacity for just a handful of users within its coverage area, so conference rooms, cafeterias and other areas with high user density will need to be over-provisioned to provide enough capacity for peak usage times. These investments in extra femtocells will be stranded during times of low usage.
Management – There’s no doubt about who owns Wi-Fi access points, but the model isn’t so clear with femtocells. Consumer and SoHo femtocells are sold through retail channels, and it’s up to the consumer to manage the device and to provide backhaul for it via DSL. But enterprise IT managers may be far less enthusiastic about taking responsibility for quality of service and management of a femtocell network inside their buildings, and it is not at all clear whether mobile operators will want to do it, either.
Rather than using multiple femtocells, a better way to approach the problem is to use the signal from one femtocell (or “super femto”) and distribute it throughout the office with a distributed antenna system (DAS). Here’s why:
Interference – Since all RF channels are distributed to each antenna in the DAS (creating one large cell or coverage area), there will be no interference among coverage areas. In addition, DAS solutions can be designed so the signal is only directed inward toward the building core, eliminating any leakage into the macro network.
Signal dominance – DAS solutions can easily provide uniformly strong signals throughout the interior of a building so user devices don’t hunt between interior and macro network signal sources.
Handoffs – Since the area being covered by a DAS is essentially one large cell and all antennas broadcast the same RF channels, there are no handoff s and thus no reduced handset battery life.
Over-provisioning – All antennas in the DAS have access to all of the RF channels provided by the centralized femtocell. There is no need to be concerned with traffic engineering to handle areas with high user density because all of the system’s capacity is available throughout the coverage area (to every antenna). In fact, a combined femtocell-DAS solution can easily scale to add more capacity with the addition of one or more femtocells in a central location, if needed. Larger enterprises can use picocells or a micro cell base station instead of multiple femtocells.
Management – Centralized femtocells located in a data center are far easier to manage, and the DAS itself requires virtually no management other than to monitor it for antenna failures.
While I like the idea of small mobile base stations that reduce the cost of providing capacity within buildings, mobile network architects should understand that by themselves, multi-femtocell deployments raise as many issues as they solve unless paired with a DAS that delivers seamless coverage and eliminates key performance and management issues.
Spindler is vice president of product management for ADC’s wireless business.