UMTS NodeB macro networks often consist of hundreds, or at most a few thousand, NodeB macrocell, microcellular, and picocell base stations covering a typical market. They are totally planned—that is, the location and coverage footprint of each NodeB has been carefully selected, analyzed, and optimized in the context of all the other NodeB units that surround it. The network design remains basically static. UMTS femtocells on the other hand, while they may be owned by an individual, operate in a licensed spectrum owned by a cellular/PCS operator often using the same frequency channel as the macro network, and are part of the operator’s overall network. There could be hundreds of thousands of femtocells randomly located throughout the overall macro network deployment and potentially on the same channel with the macrocell network. Each day, hundreds of femtocells may be added, removed, or moved to different locations. The addition of each HNB femtocell potentially changes the dynamics of the network design
Femtocells provide a great improvement in indoor coverage for voice calls in areas where macrocell coverage is weak or spotty. For high-speed mobile broadband data services, the advantages of deploying femtocells for indoor coverage are even more compelling than for pure voice services, both in terms of greatly improved coverage as well as the overall user experience. A typical user can expect to achieve broadband data rates, throughout their house, of approximately three to five times that of the existing wireless macro network. Femtocells deliver indoor coverage with high data rates and with extremely high reliability, a combination that is hard to achieve using macrocells alone. Measurements have demonstrated that femtocell data rates are higher and significantly more consistent than the macrocell rates, exceeding 3.5 to 5 Mbps in almost all locations of the home on a statistically reliable basis. The bottom line is that femtocells provide a dramatic improvement in indoor UMTS data performance when compared with using the existing macrocell network alone.
While universal mobile broadband is the goal and there are numerous advantages to deploying a network of HNB UMTS femtocells, a number of significant issues and challenges exist in deploying a femtocell network as an underlay to the existing macro/micro network architecture.
Because UMTS uses 5-MHz channel allocations, and most operators have two or at most three UMTS carriers, deploying femtocells on the same frequency with the macrocell network is a fact of life. Therefore, minimizing and managing interference between femtocells and the macrocell network is of prime concern to carriers. As the number of femtocells grows into the millions, managing interference between femtocells will become a primary challenge, which means centralized planning and control will evolve into distributed interference management architectures. This must all be done in the context of a dynamic network with femtocells constantly entering and leaving the network.
A common interference scenario is the interference caused to macrocell (or other femtocells) UEs by femtocells operating on the same or adjacent channel. This scenario is of greatest concern when the femtocell is located near the edge of coverage of the overlaid macrocell. An example of the interference impact to macrocell UEs due to the presence of femtocells is illustrated in the illustration below;
The figure shows a macrocell UE being interfered with by two femtocells f1 and f2 operating on the same channel, and one femtocell f3, operating on an adjacent channel. Note that f3’s coverage region is drawn smaller than that of f1 and f2 to indicate that its interference effect on the macrocell UE is discounted by 33 dB (the adjacent channel rejection). The combined effect of the three femtocells results in a fairly large zone in which macro UEs face considerable interference or may be totally excluded from service unless they move on their own, or are redirected to a different UMTS frequency or different radio access technology (e.g., GSM). The effect of interference from femtocells can be such that surrounding each femtocell is a “dead zone” or “black hole” in which a UE, operating on the same frequency, that is restricted from registering or using the femtocell will not be able to see the macro network due to forward link interference from the femtocell. If the macrocell UE is on an active call as it enters the femtocell zone, it must be redirected either to 2G GSM service or to another UMTS carrier, and perform a hard hand-over. Otherwise, the call will drop when the macrocell SINR drops below the minimum needed to sustain service. When an idle mobile approaches the femtocell zone, the situation is even more complicated and is dependent on the version of the UE device and on the access control mode of the femtocell. If the femtocell is operating using “open access,” in which any user may register onto the femtocell while idle, then the problem is relatively simple. Idle users perform same frequency cell-reselection, register, and are granted service from the femtocell. In closed access, in which only a limited list of users (family and friends) may use the femtocell, unauthorized users will be rejected, and depending on the UE behavior will need to do an inter-frequency search for another UMTS carrier or move to GSM.
The above scenario shows that from the perspective of femtocell UEs, signals emanating from nearby macrocell base stations transmitting on the same or an adjacent frequency constitute downlink interference.
Now suppose that a scenario when the femtocell is located close (in terms of propagation loss) to the macrocell and is also directly linked to the uplink interference scenario. Unlike interference from other femtocells, interference from macrocells is constant and is something the femtocell network has no control over. Thus, the only way femtocells can combat this interference is to increase their transmit power to satisfy the target coverage radius or to move to an unoccupied frequency, if available. This is one of the key functions of the automatic RF planning and provisioning function. As the density of femtocell deployments grow, there will be increasing requirements for more advanced and scalable algorithms for managing the interference between femtocells, and in the process a migration from centrally planned RF provisioning algorithms to more distributed algorithms will occur. To facilitate this, femtocells will need to be able to directly communicate with each other in a completely distributed manner. The presence of such autonomous communication opens up several possibilities for “co-operative” interference management.
