Using programmable logic devices is the most effective way to address the fluid and continuously improving algorithmic and connectivity function requirements in Cloud-RAN.
Initially, system vendors used proprietary protocols to distribute timing and synchronization information to remote radio heads. The Open Base Station Architecture Initiative (OBSAI) and Common Public Radio Interface (CPRI) standards were introduced to standardize connectivity and synchronization between the base station chassis and remote radio heads. These protocols enable dissemination of synchronization information along with the time division multiplexed-based data plane transport.
The stringent requirements of maintaining round-trip deterministic latency of less than 16 ns, along with the timing alignment error for transmit diversity chains to within 65 ns, has resulted in the use of dedicated fiber links between the base station chassis and remote radio heads. Point-to-point microwave links are also being used in some cases where fiber connectivity is not economical.
Dedicated fiber for connectivity between the base station chassis and the radio heads is extremely limiting and expensive. To optimize fiber connectivity, the remote radio heads are connected to the base station chassis using chain, tree, or star topologies. The CPRI and OBSAI standards support fibers up to 40 km long, but the majority of remote radio installations are constrained to distances of hundreds of meters from the base station chassis. A widespread rollout of Cloud-RAN-based distributed base stations would need a fiber reach of up to 40 km and -- more importantly -- over a shared network. Transition to a common Ethernet-based data transport protocol within the base station chassis and remote radio heads would be a vital step toward the use of shared networks. Deployment of fine-grain traffic engineering functions within the shared network would be another key requirement to forward data to remote radio heads in a prioritized way with careful management of intermediate node buffering to achieve desired deterministic latency resolution.
Connecting any baseband channel card to any remote radio heads in a Cloud-RAN topology would require a new set of hierarchical switch functions. Today, these switches reside in the base station chassis and allow data per each antenna carrier to be switched from any of 3-6 base band cards to any of 12 or so remote radio heads. Base band cards and remote radio heads multiply in a Cloud-RAN topology. Bigger and hierarchical switch functions will be required to achieve the desired connectivity.
Conceptual Cloud-RAN sharing network architecture using QoS/traffic engineering. (Click here for a larger image.)
Using programmable logic devices is the most effective way to address the fluid and continuously improving Cloud-RAN algorithmic and connectivity function requirements. Programmable logic devices are widely used in channel cards, radio heads, network nodes, and backhauling equipment.
The presence of this programmable capability in every node within the network can be leveraged to keep underlying algorithms and connectivity functions in lockstep via field upgrades. The Xilinx 28nm All Programmable SoC family, for example, integrates FPGA, CPU, DSP, and analog mixed-signal functions along with an optimal number of high-speed transceivers and IO interconnect in a single device. This processor-centric platform offers software, hardware, and IO programmability to build smarter switching and algorithmic functions to lay a strong foundation for truly self-healing, self-learning, and self-optimizing wireless network nodes.
Coupled with the Vivado design environment and tool suite, Xilinx solutions enables designers to deliver an unmatched time to integration, productivity, and quality of results. Xilinx tools and silicon technology are enriched by a robust ecosystem that allows for rapid innovation while providing better off-the-shelf solutions for old and new problems.
Productivity, performance, and time to market are of the essence. Xilinx 20nm UltraScale All Programmable devices deliver ASIC-class system-level performance for building high throughput and low latency networking and signal processing functions. This family is co-optimized with the Vivado design suite and UltraFAST design methodology to accelerate time to market. Xilinx is committed to innovating tools, silicon, and solutions and expanding a healthy ecosystem to support every major technology rollout in the wireless network.
Xilinx is well positioned to help steer solutions that enable widespread adoption of Cloud-RAN networks and move Cloud-RAN into a much-needed network platform for abstracting underlying heterogeneity for effective network monetization, ease of network deployment, and maintenance.