This is part of our ‘5G Anyhaul’ blogs series providing expert insights and best practices to help you prepare your transport network for 5G and support innovative services that create new revenue opportunities.
All good stories start “once a PON a time.” Every few years, the hero of broadband – PON – faces a new challenge. And our hero always comes through. From BPON, EPON, GPON to XG-PON and TWDM-PON, our hero always finds a new way to cope with growing demand for bandwidth and performance.
Next up is the challenge of 5G anyhaul.
In 5G, mobile network operators (MNOs) need to deploy and maintain a far greater number of radio units (RUs) than ever before – without breaking the bank.
Densification is forcing two significant changes. First is a move away from over-provisioned point-to-point fiber, which has been successfully used for 4G macrocells, and towards statistical multiplexing. Second is a new distributed access architecture where network functions previously found in RUs are virtualized and moved into central (CU) or distributed (DU) units. Without all the intelligence, large numbers of RUs are cheaper to buy, deploy and maintain. But this creates new types of mobile transport depending at which point in the 5G network the transport will occur: backhaul between the network core and CU; midhaul between the CU and DU; and fronthaul between the DU and RU.
These transport types, collectively known as mobile anyhaul, have markedly different performance requirements. While backhaul and midhaul have similar relaxed bandwidth and latency requirements, fronthaul requires much lower latency and higher bandwidth. PON networks must be able to support all these requirements. 25G PON, as the next step in PON evolution, hits the sweet spot for cost-effectively fulfilling these needs.
Interestingly, where previous generations of PON have been derived from long-haul optical networks, 25G PON comes by way of the data center. The eco-system around 25G is very mature so leveraging the high volumes of data center optical components will drive the cost-efficiency of 25G PON.
While 10G PON technologies are adequate for 5G backhaul and midhaul, 25G PON would enable more RUs to be connected on the same fiber, increasing cost efficiency. But also, as some operators move to more antenna layers and wider RF bandwidth, it is expected that transport capacity requirements will quickly surpass 10 Gb/s.
However, dynamic bandwidth allocation – a property that makes PON so effective for residential and business broadband – have a latency overhead of around 2 ms that poses a challenge for fronthaul. Fronthaul latency really needs to be sub 160 μs each way.
Nokia Bell Labs has had one breakthrough in this area with cooperative DBA (CO-DBA), which demonstrated the ability of a commercial PON network to support eCPRI fronthaul. CO-DBA puts the PON into proactive mode: 5G DUs have a traffic scheduler that, in effect, prewarns optical line terminals (OLT) when RUs will need resources, resulting in a significantly lower latency overhead. Efforts to standardize the CO-DBA protocol have begun with the ITU.
While 25G is a few years from commercial deployment, all the signs are that PON can yet again up its game and meet our insatiable demand for connectivity, this time by providing the cost-effective mobile anyhaul that is essential for the success of 5G.
Another PON story with a happy ending.
Visit our Broadband Anyhaul page to learn how we can help you prepare your transport network for 5G and support innovative services that create new revenue opportunities.
Share your thoughts on this topic by joining the Twitter discussion with @nokianetworks or @nokia using #5G, #Anyhaul, #fronthaul