As mobile usage in the developed countries moves rapidly from voice to applications, the level of call processing in the network must increase correspondingly. Applications such as video and navigation require hundreds of times more data than voice, and the network’s processing capabilities must be upgraded to handle it. Yet choices now being made about 4G infrastructures will significantly impact operators’ ability to scale processing power as needed.
In this article, we will briefly look at the need for more processing power in the network as 4G applications roll out, and compare architectural choices for providing scalable processing at the lowest cost.
Mobile data applications increase need for processing power
Mobile application use demands more data. Voice service requires only 13 kilobits or less of network bandwidth, but applications like video can demand hundreds of kilobits. A second issue is latency – voice service can accept a relatively high amount of latency in the network due to buffering systems in the network base station and the nature of speech, which typically includes lots of pauses. With data-intensive applications like video, however high latency in the network makes a big difference in the subscriber experience.
These variables – bandwidth and latency – directly impact the base station and are making it necessary to provide far more processing power in the network as providers deliver new data services. The question is, “What’s the best way to provide it?”
Distributed processing architecture is difficult to scale
The industry has traditionally relied on large mobile base stations and tower-mounted antennas to deliver service. Now, a new generation of femtocells, picocells, and microcells are emerging. Carriers need better coverage and higher capacity in the network. These small base stations are easy to deploy and can provide targeted coverage and capacity in specific locations. Some carriers envision deploying thousands of such devices in their 4G networks.
But it is extremely difficult to scale processing power in small cells that are broadly distributed throughout the network. There are two problems.
The first issue is the nature of small cell base station coverage. With so many cells in the network, a typical subscriber is being “heard” by more than one base station at a time. To process the call correctly, these base stations must exchange information about the call in order to consolidate the traffic and route it through one specific base station. This call overhead traffic can actually require more processing power than the call itself; in fact, call processing overhead can be 3-4 times as much data as data from the call itself. As the number of users being heard by multiple base stations rises, the call processing overhead multiplies exponentially.
The second issue is the obvious problem of having to upgrade processing power in hundreds or thousands of devices. The cost of travel time, labor, and components quickly becomes prohibitive, and the network scales inconsistently because base stations in one area may be upgraded at one time while base stations in another area must wait for installer or component availability. Centralized processing architecture provides simpler scalability and lower costs With the need to scale processing power in the network a given for 4G networks, there is a better way to design the network so it is easier to increase such processing power.
Rather than deploying hundreds or thousands of small cells, carriers can use distributed antenna systems (DAS) to distribute the signal from one cluster or “hotel” of base stations. Instead of small cells everywhere, the provider deploys remote antennas everywhere, all of them driven by a single, centralized group of base stations. This approach overcomes the disadvantages of small base station designs. All call processing is done at one site rather than many, so there is little or no call processing overhead on the backhaul network to be exchanged between remote base stations Secondly, the processing power in the network can be efficiently upgraded in one base station (or group of base stations) located at once place in the network, so the entire portion of the network covered by hundreds of remote antennas is upgraded all at once.
3G and 4G rollouts are driving the use of richer mobile applications, and we have seen only the beginning of higher data demands from subscribers on the network. To successfully provide and scale higher processing power in the network, centralized architectures are much more efficient and much more cost-effective.