5G means 5th Generation wireless broadband technology (for cellular networks). It is a standard that is still under construction, but some wireless infrastructure providers like Ericsson are already running live tests with chip partners such as Qualcomm. There is no estimated time for the completion of the 5G standard, but this will probably emerge as we get to commercial rollouts in ~2020. The 3GPP is the main entity overseeing wireless standards like 5G and 4G LTE.
It should NOT be confused with 5G WiFi, which is a wireless local networking technology.
Because there is no final 5G standard yet, you may hear different things from different 5G actors, especially when it comes to specific details about the network infrastructure or performance benchmarks. However, it’s possible to see some common goals among most people involved in 5G.
5G at a glance
These are some high-level goals that are shared by most 5G actors and are heavily communicated. In essence, this is what makes 5G different from 4G LTE
- ~1ms network latency
- ~1Gbps per user
- 100X efficient as 4G
- 1000X more capacity as 4G
The goals of 5G
Of course, any new technology should be significantly faster. The theoretical speed for 5G is set by many to be 10 Gbps. That’s a 10X increase from 4G. Obviously, those numbers are theoretical, and each generation of network never reach those lofty goals in real-world from a user perspective (they sometimes do in the lab). The relative increase in speed does provide an interesting taste of what to expect.
During my visit to the Huawei HQ in China, someone mentioned that downloading 8GB in about ~5 seconds would be possible. If true, that would be very impressive. Let’s compare how long it would take to download 8GB using different speeds:
- 4G (80 Mbps): 14 mn
- 3G (7.2 Mbps): 2 hrs 29mn
- Modem 56K: 340 hrs 52mn
5G should be fast, even under overwhelming demand
“Peak speed” of a single 5G connection is less important as performance under heavy load from many users. We all know that sports events and other high-peak use cases can easily overwhelm existing networks. Places like Stadiums often need special 4G LTE accommodations and tuning. 5G is designed to behave much better “out of the box” and support speeds of tens of Mbps, even with tens of thousands of users. Since 4G LTE can be tuned to reach 1 Gbps of peak speed, the ability to provide high-speed to a much larger number of users is a 5G characteristic.
5G Latency is defined as the time it takes for data to go back and forth between the 5G device and the access point. The current latency for 4G is around 80ms-100ms, which means that only ten requests would already induce a wait time of 1 second. A web page can contain dozens, if not hundreds of requests.
In theory, and take this with a rather good pinch of salt, latency could be as low as 1ms. This would make mobile browsing as fast as desktop browsing. In reality, you never see 1ms of latency over regular web traffic (over the public Internet), even on desktop networks and machines.
Some 5G actors are aiming for a 5ms latency with as much as 100 devices per square meter. This is the kind of device density that we’re getting in the Ubergizmo office today (/joke).
As a comparison, private wired Ethernet networks can easily go “sub-milliseconds”, but wireless has unique challenges and advantages that wired networks could never offer.
Explosion in the number of connected devices
With the rapid rise of the Internet of Things in which many more objects can be connected to the Internet (sensors, appliances, cars, etc.), the overhead caused by each of them needs to be managed. A priority system also needs to be embedded from the ground up because many devices will be part of emergency services, and possibly new life-supporting, safeguarding and medical devices.
No-one can predict how many devices will come online with 5G, but the possibilities are endless. With a network that is pervasive enough, cheap enough and fast enough, it’s not even possible to imagine all the potential applications today, and it will take another decade to realize the full potential of 5G.
The low-latency would enable more “real-time” applications where “lag” would be bad. Any remote control (drones, machinery, medical, telepresence…) application could benefit from it.
Some companies also pushed the idea that autonomous vehicles could transmit “vision data” to be processed into the cloud, but more likely all vision processing will (and should!) happen onboard. Wireless broadband is simply not pervasive enough for that specific use case.
As data demand continues to explode, most 5G actors believe that the network has to become intelligent to meet the demand for capacity. While it’s easy to communicate and focus on peak speed, speed only aggravates the capacity problem if not managed properly.
Infrastructure providers often have some extra capacity margins, but this is very much an expensive, brute-force approach to the problem of managing peak capacity. The telecom company will either have too much margin and waste money, or not enough margin which leads to poor user experience.
Since the budget allocated to building infrastructure isn’t rising as fast as demand for capacity, higher efficiency is required to bridge the gap and to ensure an orderly evolution of the market.
With 5G, some actors want to tighten the work between fixed/wired networks and wireless networks. Most actors agree on the need of local, short-distance access points, but some want to see a wired fiber-optics backbone with mainly short wireless endpoints.
Some also want to see QoS (quality-of-service) which is a form of data-packets prioritization. This could lead to more net-neutrality debates, but it raises a good point: real-time systems could use a higher priority than buffered video streams.
The 5G network philosophy is formulated as Demand Attentive Networks (DAN).
5G Release Date / Deployments
Although wireless companies and chipmakers are very excited about the prospects of 5G, don’t expect to benefit directly from it right now.
Most 5G actors expect mass-deployments to happen around 2020 (Qualcomm executives confirmed it was the company’s expectation for now), and it will take a long time for the world to be fully equipped.
Right now, many places have yet to get 4G LTE deployed. However, leading markets have been enjoying its benefits for years now. The same thing will happen with 5G.
5G Data Plans
From a consumer standpoint, the data plans remain one of the biggest unknowns. With dizzying speeds, your 1GB-10GB plan could last mere minutes if we were to believe these peak numbers. Data plans had gone from “unlimited” when we had horribly slow 2G/3G to capped 1GB-10GB when we could blow past them relatively quickly.
What will happen with 5G? We can only assume that the data caps will be expanded somehow. By how much isn’t clear.
The fifth generation of wireless networks is designed to address some use cases that 4G LTE was not built for. The exponential rise in capacity demand, the explosion of the number of connected devices (therefore of simultaneous “users”), and the need for higher reliability and lower latency.
At the same time, 5G has to be much more efficient because economic realities dictate that it’s impossible to grow the capacity within the current framework and technologies. The network needs to become much faster.
Finally, 5G will use multiple bands and different types of network technologies (WiFi cooperation, unlicensed LTE, 6GHz+ bands, conventional bands) to optimize the network and radio congestion for each use case.