Since the early ‘70s, the role of public mobile communication has changed from voice applications for privileged individuals in business and administrations to a ubiquitous service with significant relevance in everybody’s social and business life. Such a major transition was made possible by a sequence of innovations and the introduction of new technologies as specified by the 2G, 3G, LTE and LTE-Advanced releases.
However, those technology innovations come with a hefty price tag for the operators as they create a continuous need to upgrade networks, align architectures and introduce new technologies.
Let’s take a closer look at the mobile backhaul (MBH) network. An operator, who started installing mobile voice networks several decades ago, built a TDM-based backhaul network for the GSM infrastructure and invested in PDH multiplexers for grooming and aggregation of voice channels. GSM was applied almost globally and due to this became a huge success. However, rapid growth of Internet data traffic made it questionable whether GSM was prepared to support the expected increase of data volumes in mobile networks.
The industry concluded that GSM will not be able to satisfy future data hunger and started to develop a data-optimized mobile technology. 3G/WCDMA was born and it came with a different transport paradigm for the connectivity network. ATM became the converged transport technology for both voice and data, as it supported fixed bandwidth as well as best-effort connectivity services.
Not a big deal, ATM over E1s is a pragmatic and efficient solution. By adding ATM switches at the aggregation sites and introducing automated provisioning by means of PNNI protocols, the scene was set for replacing legacy TDM by ATM. Done – the MBH network was prepared for data growth.
Whoever believed that this was the end of the story was proven wrong. 3G was soon challenged by WiMAX and the mobile industry responded with an even better technology: LTE (Long Term Evolution). While this battle went down between the radio guys, the data community had some heated discussions on ATM versus IP/MPLS. ATM lost out. Due to this, LTE moved to IP, which became the connectivity technology of choice for LTE and “IP-fication” programs were initiated by many operators. The MBH network started its transformation into an IP-optimized transport infrastructure utilizing L2 and/or L3 technologies such as Carrier Ethernet and various flavors of MPLS.
With such a complex history, many are probably wondering whether we’ve finally reached the end of this network transformation journey.
Probably not. LTE(-A) not only introduces IP for inter-node connectivity but new features and services translate into more stringent synchronization, delay and jitter requirements. Growing inter-cell communication also puts a lot of load onto the network - especially with centralized architectures.
There is even more to consider as complexity with present network layers was inherited as mobile technologies evolved from 2G to 3G and LTE. Looking at protocol stacks such as the S1 interface at an eNodeB, you will be surprised to see several technologies being stacked - one over the other - providing similar functionality in parallel. IP user traffic is mapped into GTP-U which is transported over UDP mapped in IP. Such packets are frequently carried over MPLS using Ethernet interfaces featuring at least one level of VLAN identifiers. Is there room for simplification?
Wouldn’t we want to benefit from the virtually unlimited capacity of fiber and substitute technology layers by bandwidth? And in addition, solve some other issues with network synchronization, delay and jitter?
Sounds great, but how would that work?
Simply put, a radio base station consists of a Baseband Unit (BBU), which processes user and control data, and a Radio Frequency Unit (RU), which generates the radio signal. There are some nice advantages if those two units are locally split. The BBU could be pooled at central sites with other BBUs allowing almost unrestricted inter-cell communication, while the RU, now called Remote Radio Head (RRH), could be mounted at the antenna site requiring much less power compared to a full blown RBS. Such network architecture is especially favorable for LTE-Advanced as features require a tight coordination among neighboring cells which easily happens at central processing site where BBUs are pooled.
The good news: this architecture is supported by radio vendors and uses the CPRI interface, which was specified by industry cooperation between Ericsson AB, Huawei Technologies Co. Ltd, NEC Corporation, Alcatel Lucent and Nokia Siemens Networks GmbH & Co. KG.
This CPRI (Common Private Radio Interface) interface defines a simple signal structure. The digitized RF signals from various antennas can be aggregated, which allows the building of networks in chain, tree and ring topologies. The number of sectors per antenna site and the width of radio spectrum determine the required transmission line rate.
The digitization of baseband radio signals creates a lot of data resulting in line rates from 0.6 Gbit/s up to almost 10 Gbit/s, which demands fiber as the transmission medium. Different from the backhaul network, which is a Carrier Ethernet or MPLS network, this “fronthaul” network is a very simple optical transmission network based on transparent point-to-point connections. This is favorably implemented with DWDM/CWDM technology in order to maximize utilization of the fiber and to minimize installation cost.
As this link transports an analog signal in digital form, the transport system is now more or less agnostic to the radio technology and the same technology can be applied for GSM, 3G or 4G and even WiMAX. Hence, transparent optical transmission systems can easily be adapted to any mobile technology.
As changes in mobile technology do not impact transmission format and transmission technology, fronthaul networks are a sensible step towards a truly Self-Defined Radio Access Network. Such network construction practice forms a nice basis for spectral farming business models allowing an operator to seamlessly align mapping of mobile technologies to available spectrum without the need to care about the connectivity network.
To make a long story short: fronthaul networks will become the first mile access solution in mobile networks. Such design practice reduces complexity with technology layering and can be used for any mobile technology. Co-location of BBUs simplifies inter-cell communication which is a pre-requisite for efficient radio resource utilization. CWDM/DWDM transmission systems will efficiently utilize the fiber infrastructure providing essential activation and assurance functions. Fronthaul networks will complement backhaul networks and will become an essential step towards truly Self-Defined Radio Access Networks.