Summary Bullets:
- Mobile edge computing (MEC) is often cited as central to 5G core network evolutions, supporting the key 5G use cases: Critical Communications, Massive IoT, and Enhanced Mobile Broadband.
- 5G use cases may require MEC and edge computing, but MEC deployment does not need to wait until 5G arrives as signaled by operator trials.
While it’s not exactly clear what a 5G mobile core will look like, two topics almost always get raised when talking about it: MEC and network slicing. We’ll talk about network slicing in a future post. In the meantime, I wanted to dig into MEC.
After Nokia (well, Nokia Siemens Networks, back then) announced its Liquid Applications offer more than three years ago, at Mobile World Congress 2013, I recall asking them why they wouldn’t explicitly link it to NFV, since it was about hosting diverse applications on IT servers, just at the edge of the network – not the core. At that time, they explained that efforts at standardization might be expected going forward, but that their focus was on “applications” rather than “functions.” In other words, they didn’t see it aligned perfectly with NFV efforts. Nokia, of course, went on to play a major role in driving MEC, a technology supporting edge-based applications while still leveraging NFV specifications rather than “reinventing the wheel.” After MEC efforts got underway, then, I recall asking some smaller network infrastructure and software vendors how their RAN-linked gateway offers or portable software solutions were linked to MEC. The answer was that they weren’t linked. Today, they’re MEC members, speak at things like the MEC Congress and generally support the technology.
What’s the point of this introduction? It’s not to argue that we’re geniuses or fortune tellers. Instead it’s simply that, several years ago, there might have been confusion around the value of standardized edge computing, in general, and MEC specifically. Today, that confusion seems to be clearing as we get closer to 5G launches with use cases that require edge computing.
If you’ve made it this far, you probably know what MEC is. If not, here’s a quick rundown from ETSI, “MEC offers application developers and content providers cloud-computing capabilities and an IT service environment at the edge of the mobile network…characterized by ultra-low latency and high bandwidth as well as real-time access to radio network information that can be leveraged by applications.” In short, we’re talking about RAN-aware storage and compute at the base station or close to the network edge. The linkages to core 5G use cases, then, come in multiple forms.
- Critical Communications. Where 5G applications require low-latency, serving them up from the network edge – processing traffic and delivering applications closer to the user – will be a necessity. Consider autonomous driving, for example, where vehicles will need to share road hazard data in real-time, without the delay imposed by sharing this data further in the network core.
- Enhanced Mobile Broadband. Beyond critical communications, low latency will also be important for supporting a solid user experience around high-bandwidth applications like AR/VR. What’s more being RAN-aware, content served from the edge of the network can be delivered more effectively – conserving on backhaul bandwidth in the process (a potential issue given 5G RAN capacity).
- Massive IoT. IoT use cases depend on analytics capabilities, often time delivered in real-time for “data in motion.” Where MEC supports siting analytics at the network edge, IoT is supported.
Well before 5G arrives, however, we’re already seeing initial momentum in the form of trials and demos by carriers like Deutsche Telecom (assisted driving), China Mobile (edge video orchestration) and EE (video orchestration and crowd safety). Beyond the usual benefits, EE’s director of RAN – Mansoor Hanif – even pitched MEC as a solution for helping launch services more quickly at the RANWorld event earlier this year. In other words, 5G use cases may require MEC and edge computing, but you don’t need to wait for 5G to start deploying MEC. Or, perhaps more eloquently, as Caroline Chan (General Manager for Intel’s 5G Infrastructure Division) puts it, “MEC is justified with 4G and essential for 5G.”
Where the timing of 5G aligns with a move towards IT network principles and even open source technologies, no discussion of MEC would be complete without asking a very basic question: are MEC specifications completely necessary? Can’t service providers support edge computing use cases – edge clouds – without getting bogged down by ETSI standards? Edge computing might be key to delivering on 5G promises, but is MEC?
For its part, Ericsson has questioned MEC on this basis. And given the fact that a company like Intel will benefit from the proliferation of compute resources regardless of the specifications around them, you might expect them to be ambivalent on the topic. Nonetheless, when I had a chance to catch up with Intel’s Network Platform Solutions team at the recent SDN World Congress, they argued that MEC provides a platform for addressing three key service provider concerns: simplified service enablement, security, and network resource exposure. Can all three be enabled in the absence of MEC? Maybe. In a consistent way that will drive the ecosystem forward on the schedule at which 5G is maturing? Less likely.