As noted in the previous article, there are considerable hurdles to jump for operators which want to deploy radical new architectures – including 5G NR Standalone or a cloud-native core and RAN – in order to exploit the full potential of 5G. But some operators are making bold moves towards the new architectures, either because they are relatively greenfield players like Rakuten, Dish and Reliance Jio, or they feel are operating in markets, like the USA, where significant competitive pressure is partly related to the quality and advanced state of the network. AT&T and Verizon are very advanced deployers of virtualized, software-defined networks and white box hardware, and now T-Mobile USA is hurtling towards 5G Standalone (SA).
TMO said it completed a significant trial – of an SA data session running on a multivendor network, with core elements from Nokia and Cisco, RAN from Ericsson, and devices based on chips from MediaTek. This may not be ushering in new vendors – an important goal for some MNOs in pursuing open interoperable networks – but it shows operators’ determination not to be locked into single-vendor, end-to-end systems as they often have been in 4G, thanks to semi-proprietary implementations of key interfaces such as CPRI and X2.
The data session took place at TMO’s Bellevue, Washington lab. CTO Neville Ray called it a “major 5G breakthrough” and added a push to the FCC to approve the proposed merger with Sprint, saying: “5G brings a new era in wireless, and if our merger with Sprint is approved, the New T-Mobile will bring together the resources and vision necessary to ensure America has a network that’s second to none.”
However forward-thinking they are, the big US operators still have to consider the challenges and costs of supporting and integrating legacy networks and undergoing the organizational change to enable a fully cloud-native approach. For new mobile players, the only challenge is to deploy an advanced new network, and that can support very different costs and timescales, making it easier to adopt modern techniques, from the webscale world, with less upheaval and risk.
For instance, Dish Network, which will receive network and spectrum assets from Sprint assuming its merger with TMO proceeds, will use an infrastructure MVNO agreement with New T-Mobile as a springboard to build its own national 5G network. It is already weighing up ways to reduce the cost of that deployment by making extensive use of virtualization and software-defined networking (SDN) to slash total cost of ownership (TCO).
Dish chairman Charlie Ergen has said he aims to learn lessons from Japan’s Rakuten Mobile, which is designing a cloud-native greenfield network in Japan using equipment, software and services from a long list of vendors. This makes sense for a new entrant with a web mentality, but is nonetheless a challenging task. Red Hat, HPE and Cisco, which are providing the NFV infrastructure, operating system and integration, have spoken of the challenges of implementing something as demanding and complex as a RAN on cloud hardware. Currently, this requires many cycles of tuning and optimization – some way from the fully automated, instantly reconfigurable dream of the fully cloud-native network of the future.
Dish, like Rakuten, could reduce costs and increase supplier competition by adding newer vendors alongside those with the proven track record. Rakuten is using small cells from Airspan and RAN virtual network functions from Altiostar alongside equipment from Nokia, NEC and other majors. Dish could also consider the effect of the rising trade wars between the USA and other countries, and the current government’s push for greater technological self-sufficiency, and select US-based companies like Mavenir, Parallel Wireless or JMA Wireless, all of which offer vRANs.
But of course, many of these companies have limited track record of engineering and supporting large-scale networks, rather than discrete small cell systems. They have a good shot at densification deals to deploy metro small cell networks, but it seems more likely that most MNOs will turn to the established vendors, with their huge engineering resource and experience, for the nationwide RANs. So the key to a disaggregated network with entirely new cost base will be to use collective MNO power to push the large OEMs to support fully open interfaces such as ORAN (see separate item), to enable them to mix and match large and small suppliers within open multivendor systems.
The new entrants, such as Dish, Jio and Rakuten, are taking some of the burden of driving this change among the vendor community, since they do have the luxury of starting from scratch with their architecture and supplier selection. But they will also have a higher level of risk because they lack experience of deploying large-scale mobile networks, which may make them conservative in some key areas of the platform (like Rakuten in the radio).
In India, Reliance Jio has been skilful at managing its suppliers and driving them to work closely with it, to co-develop very low cost network components such as small cells and in-building routers, as well as a homegrown self-optimizing network (SON) system which it may put into open source.
According to Jio’s new CTO, Shyam Mardikar, (who replaced Tareq Amin when he jumped ship to Rakuten), there are five pillars to Jio’s roadmap for 5G. He told the ET Telecom 5G conference in Delhi that these are “high speed connectivity, a virtualized and distributed core, closer cooperation with our customers, a robust and affordable device ecosystem, and a powerful radio offering with larger spectrum”.
But he said backhaul was a big potential challenge in India, saying: “Backhaul is potentially the hardest challenge. If we don’t sort the backhaul out properly, the 5G experience will be like driving a Ferrari on a road that is completely clogged up with roadworks and traffic. In that scenario, the Ferrari becomes completely useless.”
A choice of low cost smartphones will also be essential to Jio, which recently became India’s second largest MNO by subscriber numbers (see separate item) on the back of disruptive combinations of cheap services with bundled web services.
“Where are the 5G devices? How do we get that side of the equation up and running? If we don’t focus on the consumer with 5G it will be like putting the cart before the horse. It’s that fundamental,” Mardikar said.