The long-term evolution of LTE

LTE networks can provide connectivity to a variety of fixed and IoT devices. They are scalable, future-proof wireless networks, and make the perfect choice for a WISP, farm, or private LTE deployment. So why are LTE networks less commonplace than WiFi has been in the past decade?

To understand the reasons for that we need to look at the long-term evolution of open source implementations and how open systems dominate over time through iterative improvements that outscale proprietary ones. As open systems are able to innovate easier, faster, and more affordably, industries gravitate towards building on them. They build strong communities that contribute back as they innovate. Traditional proprietary LTE has not yet reached this point, and as a result remains largely inaccessible.

Why-Fi?

WiFi is an example of one technology that has succeeded as a direct result of having been implemented in an open and accessible way. It evolved to its current level of ubiquity due to its easy accessibility to developers. It’s easy to build on and implementations are iteratively developed, making improvements available to everyone.

Unencumbered by proprietary hurdles, developers have been free to iteratively improve upon open source software integrations of WiFi. The confidence inspiring accessibility provided by non-proprietary WiFi has had a lasting effect. It essentially led to WiFi becoming what is now an integral part of most handheld devices, computers, and embedded systems the world over. So why can’t LTE be that too?

Well, this is a possibility for LTE as well. The LTE standard itself is not proprietary after all. But implementations of LTE typically do rely on proprietary hardware, software, and firmware. These proprietary implementations have inherent drawbacks that negatively affect the ability of the industry itself to grow.

A part of this problem is when different companies try to compete with their own proprietary implementations. This can make it more difficult for devices and networks to interoperate and ultimately stifles development around it. The resulting negative effect on the economy ensures broken proprietary systems remain unfixed.

The ability of the industry to innovate and adopt new technologies in a stunted ecosystem is severely hindered as companies focus on protecting their proprietary technologies instead of collaborating with others to drive innovation.

This is especially problematic for smaller competitors. Smaller entities simply lack the same proprietary resources as larger, better funded ones. The costs of proprietary solutions are passed on to end users as well. They suffer the costs of those technologies, with insufficient open alternatives to adopt.

Essentially, the proprietary nature of these solutions makes them nearly impossible to innovate and compete with in any beneficial way. Growth doesn’t happen under those conditions. It happens through collaborative processes and wide scale adoption. It happens when developments are being made iteratively and shared back into the ecosystem they support. No proprietary built system or software can do that as effectively. This is why even today every major proprietary brand in the world depends on open source software for their operations. Open source is eating the world.

Open Source is Eating the World

Linux is a prime example of how open source is eating the world, and Linux has a huge appetite. Originally written as a hobby project by then 21 year old Linus Torvalds, it was up against a well funded industry standard: Unix. It not only triumphed, but continues to be an integral part of every system we use today. You can find Linux in Windows, Mac OS, Android, and pretty much everything else in one form or another. It simply makes too much sense not to base your work off of what works best. And what works best gets to be the best by being iteratively and collaboratively improved on.

One would have assumed a better backed and more prevalent Unix would have won out in the early days. But it was Linux’s values of being free to build on and contribute to by the whole community that cemented it as the victor.

If you map that history to the LTE industry, open source software implementations of LTE networks have the potential to bring about exponential growth through increased competition, innovation, and increased adoption, especially in developing countries. LTE can be as commonly integrated as WiFi is today in both consumer and business applications.

A competitive marketplace would form in private and public networks that could drive down costs for end-users as well. This would give providers flexibility in deployment and remove the long term commitment and vendor lock-in associated with proprietary implementations.

Hobbyist and business minded developers would also enjoy a lower barrier of entry. New use cases would open extra markets. This is something we saw with the rise of the Raspberry Pi. It is an open source single board computer kit that has become synonymous with DIY robotics and embedded systems. Suddenly, complex electronics products and projects could be engineered by anyone regardless of funding. Anyone could become an electronics engineer in their own home. This led to many novel innovations in handheld and embedded technologies and prototyping electronic products became even cheaper. These kinds of major evolutions are made possible only when the source is open and improvements are contributed back upstream.

Tele-qualms

Raspberry Pi’s disrupted the consumer electronics industry. Similarly, open source LTE implementations will disrupt the telecoms industry. The line would begin to blur between what and who can become telco. LTE would become democratized, allowing everyone to build their own networks. Imagine not needing to pay a telecom company a monthly service charge to use LTE on your own land.

The process for onboarding of new LTE networks would also naturally improve. It would no longer suffer bottlenecking by proprietary and monopolistic providers. Easier, self-directed deployments would be empowered by a supportive world-wide open source community. A global community of collaborating developers means faster iterative improvements and security fixes. Even the best current proprietary solutions would not be able to compete with that.

The dream of open systems and networks is a beautiful one, but proprietary providers are not going to volunteer to hand over monopoly willingly. As it always has been, the proprietary comes first, but it’s up to the community driven open implementations to follow. They will rise up and eventually become the status quo the proprietary systems depend on. It was this way with Linux and Android as well. It takes community development of open source implementations to make that future a reality.

We are now in the very early stages of that future, but open source software is already bringing the packet core to the edge in consumer devices. All that heavy lifting that used to be done through proprietary implementations is now being handled through open source software. The core network logic is being brought to the consumer and self-directed network operator directly.

Open Source Initiatives

Several open mobile core network initiatives exist today, each hoping to take us one step closer to that future.

1. OpenLTE: An open-source implementation of the LTE protocol stack, providing support for LTE data plane and control plane functions.
2. Osmo-LTE: A component-based implementation of LTE that is part of the Osmocom open-source cellular infrastructure project.
3. srsLTE / srs5G: A fully open-source LTE software suite that provides support for the LTE air interface, including both the user plane and control plane functions.
4. Magma: An open-source high-performance enhanced packet core developed to bring advanced packet core technology at a lower total cost of ownership to MNOs, ISPs, MSOs, and enterprises that need to launch mobile broadband services.
5. Open5GS: An open-source implementation of the 5G Core (5GC) network, which provides the control plane functions for 5G wireless communication systems. srs5G: An open-source 5G software suite that provides support for the 5G air interface, including both the user plane and control plane functions.
6. KeyLTE: Built upon and contributing back to Open5GS, it is designed for deployment of public and private LTE networks, agile and lightweight enough to embed the EPC (Evolved Packet Core) into a smart router.

While each has their own merits, it’s important to understand the metrics that matter the most for real world implementation. Platform scalability, security, and ease of deployment are important, but most important to becoming a standard is whether whether it is designed to receive iterative upstream contributions and grow through a strong developer community. This is in many ways influenced by its license.

To illustrate this, we’ll compare two well known but differently structured implementations and see how license choice and structure can affect adoption.

Magma

Magma Core is an open-source LTE network platform designed by Facebook with the hopes of democratizing internet access in the developing world. It was designed as a big company core solution with a lot of components and a big development team. If nothing else it is further evidence by major development companies that open source telecom is a necessity to grow and develop unencumbered. One issue with it though is due to how the design is structured to mimic the monolithic company that created it, it’s more difficult to implement meaning a higher barrier to make improvements.

The license also presents concerns for healthy adoption. Magma’s software license — the BSD-3 license — is considered to be permissive, which means that it allows derivative works to be released under a different license. This might be fine for some non-essential open source projects, but for critical network infrastructure software everyone relies on, it’s not ideal.

Aside from creating potential compatibility issues for projects that require all components to be released under the same license, the biggest issue is the lack of requirement or incentive to contribute back to it. This means that improvements made, if any, will end up benefiting some proprietary version in the wild instead of upstreaming back to make Magma better for everyone.

Simply put, it creates the perverse incentive for companies to take code, close it themselves, and not contribute back. While it’s well within their legal rights to do so, it has a parasitic effect on the health of the codebase and community using it and can ultimately lead to stagnated, broken, or increasingly inefficient and difficult to implement code — which is exactly where some see Magma today.

In the case of open LTE in an industry notorious for closing its code to maintain continued monopoly, openness would need to be baked in at the license level. Else it simply will not scale to the level of iterative development necessary.

Open5GS

On the opposite side of the spectrum in terms of side and complexity is Open5GS. Open5GS is an open-source implementation of the 5G Core (5GC) network. It was designed primarily as a research core that researchers would use for their experimentation — highly modular, compact, easier to use. This makes it a great choice for real world development as well, as the developmental focus is on outcomes rather than supporting and justifying pre-defined monolithic structures.

In terms of incentives for upstream contribution, Open5GS uses the Affero General Public License (AGPL). The AGPL license is a “copyleft” license, which requires derivative works to be released under the same license. This ensures that changes to the software are contributed back to the community, promoting collaboration and innovation.

One open LTE toolkit called KeyLTE chose Open5GS core to adapt for LTE networks. Instead of expensive, unwieldy legacy core topologies, networks could then be built easily and affordably in a variety of environments. Open5GS also benefits from KeyLTE contributing back to make the original code better with each iteration. This symbiotic relationship strengthens both technologies and the whole industry with continued permanent improvements. It’s these types of relationships that make a continually growing open ecosystem possible.

The True Long-Term Evolution

More open implementations of LTE will likely take this approach, making their decisions based on favorable licenses, community commitments, and highly compartmentalized, highly iterative codebases. The underlying benefits will be felt across the industry as it continues to open, and more people and businesses start seeing the benefits of switching to an open LTE for their network needs.

The ability to easily build upon solid open foundations is the real long-term evolution of LTE, and it is powered by open source implementations. While still relatively early in the evolutionary phase, open source has already shown an insatiable appetite and it’s definitely coming for LTE.