The 5G Revolution Requires an Open Ecosystem

Thoughts about technology that is inclusive, trusted, and creates a more sustainable world

These posts represent my personal views on the future of the digital economy powered by the cloud and artificial intelligence. Unless otherwise indicated, they do not represent the official views of Microsoft.

Few technologies have had a more transformative impact on human society than mobile networks. They have enabled the world to go in a single generation from the revolutionary idea of “a computer on every desk” to the even more radical reality of “a computer with an Internet connection in the palm of every hand.” In barely more than a decade, modern ways of living and working have been revolutionized by the fact that nearly everyone is connected to a global data network at every minute of the day and has the compute power to exploit that connection for an infinity of purposes.

But technology never stops. Soon, 5G networks will put ever-faster Internet connections on every artifact of civilization, from the most sophisticated industrial machinery and critical infrastructure, to cars and household appliances, to the most insignificant of gadgets. I’ve written about 5G several times in recent months, focusing on the need for governments to make sure that deployment of 5G networks is not delayed by the pandemic. Today I want to look at how the technology underlying 5G networks is itself about to be transformed by the cloud revolution. This transformation has important implications for businesses and for government policy.

Source: SDxCentral

Mobile networks like 5G are really just collections of specialized computers that are connected to our phones and other user devices by large numbers of radio antennas scattered around the landscape. The basic principles that govern the evolution of mobile networks are the same as those that have always driven computer architectures. In both cases, the key trend is the shift from closed, proprietary systems where one company provides the entire system to an open ecosystem where many companies compete to provide components at different levels.

The idea of open ecosystems is familiar to us in the PC and server markets. Each of these vast markets is divided into a “stack” of  “layers,” with clearly defined public interfaces between the layers. In PCs and servers, there are public interfaces between the hardware, the operating system, and the applications. Different companies compete in each domain. Most end-user PCs run Windows, but you can easily run Linux on your PC if you choose, and many people—including many professional software developers—do just that. In the data center it’s the opposite: most servers these days run Linux, while many others run Windows Server. And on both PCs and servers, you can choose hardware that uses Intel, AMD, or ARM processors.

Now, this fundamental principle of dividing things into distinct layers connected by open interfaces is beginning to happen with the internal hardware and software that makes up 5G networks. The “stack” in a 5G network consists of four fundamental layers:

1. End-user devices, such as phones or Internet of Things (IoT) devices.
2. The Radio Access Network or RAN, which consists of cell sites and their subcomponents such as radios, antennas, computers, and software.
3. The core network, where server computers (traditionally located in an operator’s data center) perform connectivity, administrative, and value-added functions.
4. The countless applications provided by third parties that end-users access over the network. Here the “end-users” who use these applications include consumers, enterprises, and IoT-equipped artifacts such as cars, household appliances, factory machines, and critical infrastructure.

Layers 1 and 4 are the parts of a mobile network that users see. Layers 2 and 3 are the expensive hardware and software parts that network operators buy from specialized telecom equipment vendors such as Ericsson, Nokia, Samsung, and Huawei. Layers 1 and 4, the things we users really care about, would not be possible without layers 2 and 3, which are provided by the expertise and capital of the operators.

Traditionally the hardware and software in layers 2 and 3—respectively the network periphery and the network core—were sold to operators as integrated proprietary bundles from a single equipment vendor. But in the last few years, things have begun to change. Many operators now run their core network applications using virtualization software on industry-standard servers. Having taken this first step towards open architectures, some operators are now beginning to move the virtual machines running their core network applications out of their own data centers and into the cloud. This is exactly the path that large enterprises in other industries have followed over the past decade, which has seen the cloud rise up from almost nothing to a $266 billion global market in 2020. The migration of core network functions out of the operators’ own data centers is what the industry calls moving to cloud native.

On the Radio Access (RAN) side of the network a parallel movement is underway to define what the industry calls OpenRAN, that is, open interfaces between the radio components that connect with end devices and the compute components that do basic processing on the data signal at the edge of the network before passing it up to the core. The compute components of RAN networks are now being virtualized in the same way as the core network was, and that virtualization will make it possible for them to move to the cloud just as the core is doing.

Much of a mobile network consists of expensive capital assets: real estate, radios and their antennas, and spectrum. Cloud providers like Microsoft and Google do not propose to become mobile network operators or to run the operators’ businesses for them. Operators will continue to compete on the advanced services they provide to customers as well as owning the vital capital assets and of course branding and marketing. But cloud providers will unburden operators of the cost and complexity of owning and operating vast fleets of servers, both at the core and the periphery of their networks.

Until now we’ve focused on technology layers and stacks. But we should not lose sight of the higher level where things like government policies and national strategies come into play. The trend toward open 5G ecosystems will need support from governments in the form of regulatory mandates for open, standards-based technology and financial support for advanced R&D to develop this technology. As Diane Rinaldo, Executive Director of the recently formed Open RAN Policy Coalition puts it:

“As evidenced by the current global pandemic, vendor choice and flexibility in next-generation network deployments are necessary from a security and performance standpoint. By promoting policies that standardize and develop open interfaces, we can ensure interoperability and security across different players and potentially lower the barrier to entry for new innovators.”

5G networks are going to be a key enabler of the ongoing digital transformation of business and government over the next decade. It’s vitally important that industry and policymakers work together to ensure that these networks can develop efficiently in an open, standards-based ecosystem where diverse suppliers can compete and innovate.