Realizing Extreme Flexibility

Prof. Akihiro Nakao, Chairman, Network Architecture Committee

A basic concept of 5G mobile networks (here on, 5G) infrastructure is the idea of extreme flexibility. A key concept that describes how this will be realized is network softwarization. There are many ways to look at this issue, but one way to think about this technology is that inside each device there are many kinds of programs that can be called on to handle specific issues. In other words, the flexibility comes through the software. This form of quickly and flexibly evolve-able infrastructure is essential to 5G and therefore this is one path we are going forward on the way to realizing 5G itself.

Put in another way, extreme flexibility can be realized is through network slicing. Up until now, data has flown through pipes in a network regardless of what kind of data it is. With network slicing, data flows are diverted through different types of pipes depending on how the data is categorized. Which data flows through which pipe is not decided upon in advance. Instead these decisions are made dynamically and intelligently as the situation changes. If different types of data are to be sent and received together, it is difficult to manage the specific characteristics of individual flows of data. This is where slicing comes into to play. A slice is an independent and isolated virtual pipe implemented on top of a physical pipe. Network slicing is a technique to enable isolation and aggregation of various resources such as network resources (the ability to transmit data), computing resources (ability to process and analyze data), storage resources (the ability to store and retrieve data) into different slices to so as to allocate an isolated set of resources (and functions implemented atop) to each data flow for maintaining the specific characteristics of the data flow.

Up until now networks have been primarily concerned with data transfer. As 5G wireless technology is rapidly emerging, however, expectations are so high that it may also include edge computing, new technology to provide data processing near user equipment in order to utilize ultralow latency in wireless data transmission, e.g., to enable instantaneous feedback. For this to happen, we need to consider utilizing network slicing for computing resources and storage resources for a given application data flow as described above. Each type of industry, so called, vertical, may have common requirements for communication infrastructure, so it may make sense to allocate slices per vertical industry. Technically speaking, while content delivery application require high capacity and multicast, in the Internet of Things (IoT) scenario, sensors need to send out small amount of data spontaneously. Slice granularity depends on use cases, e.g., per vertical, per application, I am confident in saying we are leading the world in this way of thinking, network softwarization and network slicing.

The idea behind using slices in networks is thought to have begun with PlanetLab project at Princeton University in 2002, specifically, in the lab I belonged to. The research and development has progressed in the United States, Europe, as well as our own country of Japan. Recently, this concept has spread into the thinking about 5G mobile networking, especially as it must be extended to wireless networking. I had an intuition that the extension of slicing concept should inevitably come up when the performance of wireless network become comparable to that of fixed networks. This very concept should be realized in the infrastructure of 5G.

The details will be available on the 5GMF website soon, but for the network committee has identified four areas that we should focus our research on in order to quickly ascertain their suitability: First, network softwarization (with the idea of network slicing); second, network slicing in terms of mobile fronthaul/backhaul; third, a unified, intelligent management system; and fourth, mobile edge computing (data processing near user equipment). A major theme of this research, however, is the ability to have a high level of cooperation between wireless and fixed networks.

Now, the progress we have made with network slicing has reached a global level of discussion. However, we are currently at the stage where we have just turned on the ignition, so to speak. From the discussions in the research arena, where this used to seem like a faraway dream, we are now at the point where the 3GPP introduces this concept seriously to the world’s vendors and operators. We are in the middle of a great change in the meaning of this technology, and it is exciting to be a part of the 5GMF Network Architecture Committee at a time where our everyday work is a part of these great changes. The world is going to change and the level of excitement on this new technology will only increase once this concept enters the standards discussions.