Server Based Solutions for Self-Organizing Networks
Re-imagining the network
There is no unified view of how these network components are to be re-imagined. There are, however, a range of competing ideas, initiatives, and solutions:
• For example, the European Horizon 2020 COHERENT project, of which CommAgility is a participant, is looking to develop a unified programmable control framework for managing heterogeneous networks, (HetNets) which will be key to 5G success.
• Working on the signal chain from the air interface to the EPC, the first key element is the Remote Radio Head (RRH). Traditionally the RRH was a single sector radio mounted up the antenna mast supporting a point-to-point link back to the eNodeB located at the base of the mast. Common Public Radio Interface (CPRI) has become the protocol of choice for this link, connected over optical fiber.
The evolution of the SDN supports the reconfiguration of these RRHs using Software-defined Radio (SDR) techniques to support multiple bands at multiple bandwidths, thus allowing each RRH to play its part in the HetNet. Where a channel experiences interference, the RRH frequency may be moved. In areas of low usage, the bandwidth of the supported channel can be reduced and that bandwidth re-allocated to a busier cell without creating adjacent channel interference.
RRHs can be turned off and on depending on local demand, being used as sniffer channels to feedback details of the RF environment to improve overall network coverage.
Moving the eNodeB inside the data center allows the support of multiple RRH connections, not just the traditional three-sector model. This Distributed Antenna System (DAS) model as pictured in Figure 1 brings in multiple RRH connections to the data center, which can then be switched to a high-density eNodeB pool where the wireless data processing can be shared across a central resource.
On arrival at the data center, the point-to-point CPRI technology needs to be adapted to a switchable architecture in order to terminate the traffic at available resources. This is easier to envisage with Ethernet-based technologies, or technologies built on switched fabrics such as RapioIO, but point-to-point fabrics will need some level of custom switching based solution.
For example, FPGAs could be used to discretely switch point-to-point channels under a proprietary control protocol.
Once traffic arrives at the eNodeB pool, there is much greater flexibility than at the mast-located eNodeB. Protocol stacks with visibility of multiple RRHs can aggregate network statistics from Operators, Administration and Management (OAM) data which can be used to optimize the performance of the network as a whole. For example, gathering intelligence of CQI and Sounding Reference Signal (SRS) measurements, user bandwidth demand, and Quality of Service (QoS) parameters from the EPC.
Intelligent schedulers at the eNodeB support the use of a central resource with greater processing bandwidth to improve the overall network performance, for example focusing capacity on-demand hot spots. There is the ability to switch the operational protocols of the RRH operating in specific areas, for example increasing the RRH operational bandwidth or moving to narrow-band LTE protocols where there is a concentration of IoT devices.
Organizations such as the Small Cell Forum are working to standardize how SONs deal with HetNets work with intelligent schedulers. Issues such as handover and security gateway issues in a multi-vendor SON need to be addressed. Development of a common API will go a long way to supporting this.
A Distributed Antenna System (DAS) model brings in multiple RRH connections to the data center, which can then be switched to a high-density eNodeB pool where the wireless data processing can be shared across a central resource.
There is no unified view of how these network components are to be re-imagined. There are, however, a range of competing ideas, initiatives, and solutions:
• For example, the European Horizon 2020 COHERENT project, of which CommAgility is a participant, is looking to develop a unified programmable control framework for managing heterogeneous networks, (HetNets) which will be key to 5G success.
- A key initiative is to provide a simplified abstracted network view to support coordinated network resource allocation across all network types. This will lead to interface development to support programmable control and coordination to support new services.
• Working on the signal chain from the air interface to the EPC, the first key element is the Remote Radio Head (RRH). Traditionally the RRH was a single sector radio mounted up the antenna mast supporting a point-to-point link back to the eNodeB located at the base of the mast. Common Public Radio Interface (CPRI) has become the protocol of choice for this link, connected over optical fiber.
- Extensions to CPRI such as the Open Radio Interface (ORI) have been developed to support the concept of a distributed base station architecture. This technology and others are designed to replace the eNodeB at the mast with a front-haul technology from the data center to the RRH in a so-called Cloud RAN (C-RAN). ORI maintains the latency and delay timing requirements of the digital RF interface required by LTE, while striving to achieve data compression of at least 50% to facilitate more efficient use of the front haul.
- Newer contenders such as Radio over Ethernet (RoE), IEEE1904.3, are attempting to leverage the lower cost and ubiquitous nature of Ethernet Over Passive Networks (EOPN) to achieve the same result. In deployments where there is no existing dark fiber for use, the limitations of Ethernet may be an attractive compromise.
The evolution of the SDN supports the reconfiguration of these RRHs using Software-defined Radio (SDR) techniques to support multiple bands at multiple bandwidths, thus allowing each RRH to play its part in the HetNet. Where a channel experiences interference, the RRH frequency may be moved. In areas of low usage, the bandwidth of the supported channel can be reduced and that bandwidth re-allocated to a busier cell without creating adjacent channel interference.
RRHs can be turned off and on depending on local demand, being used as sniffer channels to feedback details of the RF environment to improve overall network coverage.
Moving the eNodeB inside the data center allows the support of multiple RRH connections, not just the traditional three-sector model. This Distributed Antenna System (DAS) model as pictured in Figure 1 brings in multiple RRH connections to the data center, which can then be switched to a high-density eNodeB pool where the wireless data processing can be shared across a central resource.
On arrival at the data center, the point-to-point CPRI technology needs to be adapted to a switchable architecture in order to terminate the traffic at available resources. This is easier to envisage with Ethernet-based technologies, or technologies built on switched fabrics such as RapioIO, but point-to-point fabrics will need some level of custom switching based solution.
For example, FPGAs could be used to discretely switch point-to-point channels under a proprietary control protocol.
Once traffic arrives at the eNodeB pool, there is much greater flexibility than at the mast-located eNodeB. Protocol stacks with visibility of multiple RRHs can aggregate network statistics from Operators, Administration and Management (OAM) data which can be used to optimize the performance of the network as a whole. For example, gathering intelligence of CQI and Sounding Reference Signal (SRS) measurements, user bandwidth demand, and Quality of Service (QoS) parameters from the EPC.
Intelligent schedulers at the eNodeB support the use of a central resource with greater processing bandwidth to improve the overall network performance, for example focusing capacity on-demand hot spots. There is the ability to switch the operational protocols of the RRH operating in specific areas, for example increasing the RRH operational bandwidth or moving to narrow-band LTE protocols where there is a concentration of IoT devices.
Organizations such as the Small Cell Forum are working to standardize how SONs deal with HetNets work with intelligent schedulers. Issues such as handover and security gateway issues in a multi-vendor SON need to be addressed. Development of a common API will go a long way to supporting this.
A Distributed Antenna System (DAS) model brings in multiple RRH connections to the data center, which can then be switched to a high-density eNodeB pool where the wireless data processing can be shared across a central resource.
Read full article
Hide full article
Discussion (0 comments)