The future of 5G: disruptions, opportunities and a look ahead
The communications industry is changing rapidly in all directions. Disruption in the 5G ecosystem, lower barriers to participation, and new technologies all require novel approaches. Several leaders from the 5G ecosystem recently gathered to explore the challenges, opportunities, and industry dynamics created by network disaggregation; answer some of the big questions around designing and building disruptive network architectures; and discuss business strategies for the future of 5G. Joe Bush reports.
Network disaggregation is reshaping and causing far-reaching disruption in the 5G ecosystem. The change is enormous and occurring in all directions. Simply put, network segregation gives customers the freedom to choose hardware and software options and software operating systems from different vendors - triggered by the move to 5G in particular, where a very different ecosystem is starting to emerge. Barriers to participation in that ecosystem are being lowered, and that in turn will require new approaches, supply chains, and technologies.
The future of O-RAN
Open radio access network (RAN) is a term for industry-wide standards for RAN interfaces that support interoperation between vendors' equipment. The main goal for using Open-RAN is to have an interoperability standard for RAN elements such as non-proprietary white box hardware and software from different vendors.
To this end the O-RAN ALLIANCE was formed in 2018 by AT&T, China Mobile, Deutsche Telekom, NTT DOCOMO and Orange, with the mission to re-shape the RAN industry towards more intelligent, open, virtualised and fully interoperable mobile networks. The new O-RAN standards will enable a more competitive RAN supplier ecosystem with faster innovation to improve user experience. O-RAN based mobile networks will at the same time improve the efficiency of RAN deployments as well as operations by the mobile operators.
“We’re actually very excited about network disaggregation and O-RAN - we've been participating in this wireless space for a long time and seen the transitions,” said Greg Henderson, Senior Vice President, Vertical Markets, Analog Devices. “First and foremost, O-RAN comes from an opportunity of this mega trend of virtualisation. With edge computing and standard hardware capability expanding, many of the 5G networks have the possibility to be run in a virtualised way.”
As Henderson explained, this means a lot of the network can be virtualised and run as system level software (instead of having purpose-built standard hardware), providing a much more open network architecture. So by leveraging virtualisation, you can have a multi-vendor environment, and by creating standard interfaces, you can have more flexible network features with increased capability, and potentially a lot more ecosystem partners.
For the carrier, the capability of O-RAN provides opportunities for new business models, features and more rapid evolution of capabilities on the network. “We’re very excited about that and we’re investing in solutions that will enable this open ecosystem to evolve,” Henderson added.
What are the biggest challenges for O-RAN?
Sometimes opportunities are challenges, and sometimes challenges are opportunities. And multi-vendor interoperability is fundamentally a huge challenge for O-RAN. In order to make this vision work, there will have to be multiple facets of the network that come from different suppliers, and that network will need to perform at the system level requirements of 5G.
Henderson added: “5G networks have a very high-performance standard, so all of that interoperability needs to work across use cases. We see that as a big challenge and there are three main threats to managing that challenge. The first is development standards - defining the interfaces carefully and clearly, and making those open standard interfaces that everybody can operate to.
“The second is about interoperability, which is incumbent on us as the vendor ecosystem to make sure that we can demonstrate multi-vendor interoperability. And the third is ecosystem partnerships to make sure that we’re all working together.”
Early indicators for successful adoption
Henderson continued that when it comes to what would be seen as successful adoption, there are two sides of the coin. The first really goes to the heart of the carrier space, being that there are currently O-RAN networks being announced, trialled and deployed in Asia, the Americas and Europe. So the carriers are already leaning heavily towards rolling-out deployments.
The other factor is the solution and vendor community, where custom built products are currently being developed - we will start to see targeted product solutions that are aimed at these O-RAN interfaces and, from a semiconductor company’s standpoint, such as ADI, there will begin to be products and chipsets that are targeted towards operating in a custom-built way to meet O-RAN requirements, but also have the high performance levels that you need to meet 5G.
What will virtualisation mean for the radio unit?
The key here is that much of the network can be virtualised, but not all. “If you think about the OSI protocol stack,” continued Henderson, “the bottom level is called the physical layer for a reason – it’s because it’s where this digital content connects to the physical world, and the regular unit is fundamentally in that physical layer. Therefore, there are elements of the radio unit that you can’t virtualise as that's where you’re connecting to the physical world of the radio waves.”
As such there are elements of the radio that must be hardware. However, it should also be said that there are architectures around the radio unit that can be virtualised. Having standard data models for how you interface to the radio unit, as well as the management plane, can be open and standard, and run on an open processing architecture. So, while a large chunk of the radio unit will be hardware, you can also have a virtualised interface built around the data models and management planes.
Opportunities of disaggregation
These, as Henderson explained, are heavily aligned with the opportunities for 5G. Many of the benefits of 5G are yet to be seen of course, and the current use cases that have been deployed for 5G are predominantly in the consumer space, and are about driving more bandwidth and higher data rates.
This is of course extremely important, however, 5G also has the capabilities of network slicing and low latency, which is going to open up all kinds of new applications in the industrial and industrial IoT (IIoT) space and within automated networks and automation.
“Those business models and use cases aren't really foreseeing their full potential just yet. From our perspective, we believe that this open disaggregated ecosystem will allow those use cases to evolve more quickly,” Henderson added. “Because the open ecosystem is by definition, ‘open’, there's a lot more flexibility, and new vendors can come in with new capabilities that they can plug into those networks.”
It’s ADI’s belief that while the first phase of 5G is interesting, it will be the second phase that will not only change the market for operators, service providers and their suppliers, but for the wider world, and it will start to empower enterprises towards mobile transformations.
Navigating the shift
The move towards network disaggregation will mean customers will have the freedom to choose different solutions from different vendors, depending on their particular requirements, and mix and match those together.
When we think about this in terms of the IT industry, it sounds relatively simple, but in communications it will represent a monumental shift which is going to impact the role of every player - from the biggest to the smallest in the 5G ecosystem - and introduce quite a few new ones.
Joe Barry, VP Wireless Communications, Analog Devices, commented: “The changes will be transformational. We’re seeing, and will continue to see, really exciting opportunities for the industry around this trend of disaggregation. And when you look at disaggregation in terms of the network software and hardware, it offers operators options on flexibility in terms of their supply chain. It brings forth the approach of best-in-breed products from established and disruptive suppliers. When you look at O-RAN networks, it lets more partners contribute, especially in terms of innovation and co-creation.”
Barry added that there are a rapidly expanding number of applications that are coming online for 5G, for example around IoT, private networks, consumer, and B2B. And accelerating those and creating more value for the network will be hugely beneficial. Open compute has impacted a huge amount of change, and so he stressed that we could potentially see similar, if not more, change and impact from an O-RAN network’s perspective.
“Huge opportunities lie in the ability of the ecosystem to focus on lowering the total cost of ownership for the networks through economies of scale, tackling software flexibility and cloud infrastructure leverage,” he added. “We at ADI are doing our part in terms of how we are fostering a vibrant and rapidly expanding ecosystem through our partnership and collaboration efforts. If the industry can execute these goals, I think we can all enjoy an unprecedented an accelerated innovation in 5G, and the networking and infrastructure services that will be available in the future.”
“At Facebook we have been a strong believer and supporter of disaggregated networks from the early days of O-RAN,” added Jaydeep Ranade, Director, Wireless Engineering, Facebook Connectivity. “And so as one of the founding members of the Telecom Infra Project (TIP), we have been working very closely with vendors to ensure that networks run reliably and profitably.”
He added that from Facebook’s perspective, so much business impact is tied to people using its services and apps over their mobile devices, and so the mission of Facebook Connectivity is to bring more people online, to a faster internet. As such network desegregation presents multiple opportunities to positively change the business of deploying and running cellular networks, all while enhancing their reach and reliability.
He further highlighting three key areas of change and opportunity: better network economics; unlocking new revenue streams; and unleashing technical innovation. Disaggregated networks encourage a diverse and competitive supply chain, reducing vendor lock in and risk.
Open and standardised interfaces between components increase continuous integration, improve reliable deployment and simplify messy operational automation. This reduces OPEX in the process, and at the same time, virtualised functions can be distributed across different parts of the network, improving compute efficiency and reducing CAPEX.
New revenue streams are of particular interest to service providers. Compute capabilities and a software-based architecture can give rise to new services by providing network operators with the possibility of running other workloads at the edge as opposed to just voice and data.
And finally, from the perspective of innovation, network operators can choose from a multitude of low cost, open source, lightweight core functions (depending on the deployment), standard interfaces with enhanced OSS BSS capabilities for network management, and application of machine learning.
Raj Singh, EVP Processors Business Group, Marvell, did add a word of caution by saying: “As you disaggregate functions it’s important to remember that disaggregation also introduces some level of latency, and 5G of course has its own challenges with that. So you can't just disaggregate things for sake of it. You need to do this judiciously, and in ways that give the best benefit.”
He did go on to highlight that there are use cases that we haven’t even imagined yet. If you look at when the iPhone was first introduced, if you told people then about the applications that are available now, nobody would have believed you. We could well be on the same cusp of a pivot with 5G in cellular networks.
Pitfalls and potential of O-RAN
Business and end consumers have gotten used to a certain level of quality and throughput. Therefore, going forward it will be important to maintain the same capacity, reliability and feature parity that exists in networks today.
Singh added: “There’s a pain point for massive MIMO in O-RAN and, at least until very recently, truly O-RAN compliant massive MIMO that were based on SOCs rather than FPGAs, were few and far between. We are very pleased that with our partnership with ADI we were recently able to announce a full-blown massive MIMO platform using our respective silicon, along with all the software to fill that gap.”
Security will obviously be an issue with O-RAN, more so, Singh explained, as we start looking at front hauls that are packet-based rather than IQ data, which are more vulnerable to attacks and invasions. There are currently a number of initiatives, both in the O-RAN Alliance working groups and elsewhere, where Marvell and others are looking at how to secure that front haul gateway.
He added that while there are a number of factors that will go towards creating the right ecosystem, the biggest issue will be interoperability between systems. We are moving from a world where everything came vertically from a single vendor, to one where we will be able to plug and play all these pieces together. As such, interoperability labs are appearing in a variety of locations around the world, which can only be a good thing for the industry.
Ranade added: “The biggest challenge we’re seeing from a business and technical perspective is what we are calling the unboxing experience for the service providers. The service providers today are used to dealing with a one-stop-shop as they procure and operate network components. So, this new segregated network is going to create a very different operational model.”
Therefore, unless the setup and operation of these networks is made easy, simple, and reliable, it will be very difficult to imagine a wide deployment. Therefore, interoperability, certification, and re-certification will be key.
What will also be crucial, he explained, is creating a healthy vendor base. “What currently exists is a captive ecosystem. There are a few incumbents who basically have created no incentive for the vendors to be able to do anything on their own so we want to break this cycle. But to do that, these vendors have to be able to invest money, they have to have economies of scale to be able to get the same cost price point targets. It’s a chicken and egg situation. So Facebook Connectivity is trying to break that cycle and we are hoping to put some of our resources behind this ecosystem in order to get it going.
“For this to work there has to be a clear standard open interfaces. Unless these components can be made interchangeable, and they can operate with each other in a very easy manner, it will be very difficult for people to be able to plug-and-play, and place different components.”