5G is good, but perhaps not worth upgrading just yet

New research is suggesting London 5G speeds are getting the promised boost, though the overall experience might disappoint a few.

Global Wireless Solutions, a US network benchmarking, analysis and testing firm, released its examination of the London networks of EE, Vodafone and O2, and while there is success evident in the first months, there is still plenty of work to be done.

“The spikes in the test data reveal that promises of faster speeds can be delivered, but ultimately, it’s the consistency and reliability that is most important to consumers,” said Paul Carter, CEO of Global Wireless Solutions.

“Based on the limited number of sites with 5G antennas combined with the distance constraints of higher frequency 5G signals, it’s going to be a challenge to get 5G access in buildings.

“Given that the mobile network operators have a significant rollout ahead of them to fully realise the potential of 5G, we might also benefit from a review of restrictions governing signal mast height and placement to allow more antenna sites in more convenient locations, rather than just placing them on rooftops.”

According to the analysis, the MNOs are delivering the high-speed download experience which has been promised through 5G, though only if you are standing in the right place.

At St Pauls Cathedral, EE’s network delivered instantaneous peaks of over 470 Mbps, while 330 Mbps from O2 at Victoria Station and 320 Mbps from Vodafone in Belgrave Square also demonstrated the eye-watering speeds of these networks. These are cherry-picked examples from numerous tests throughout the city, though the trend was encouraging; 5G is delivering remarkable download speed upgrades.

What is worth noting, it this is not the gigabit download speeds promised, though you have to bear in mind these networks are operating in the world of non-standalone 5G. More will be delivered in the future as the technology progresses and matures.

This is of course encouraging, however there are two elements which dampen the parade. Firstly, the availability of these download speeds and secondly, latency.

On the latency side, Global Wireless Solutions has indicated there is no meaningful upgrade from 4G connectivity. This is not entirely surprising, as without a 5G core the full-suite of latency services will not be available, though one might have expected an incremental upgrade.

Secondly, the team has noted the drop-off rate is high. By making use of higher-frequency airwaves for 5G connectivity, coverage will be shorter. There is no way around this, the laws of physics dictate the state of play here. However, as 5G is currently being built on existing passive infrastructure, designed for 4G spectrum with larger coverage cones, the problem is unavoidable.

Over the next couple of months, governments and regulators will have to be engaged to ensure the 5G experience can be delivered. Rules on the deployment of active infrastructure will have to be massaged, as relying on rooftop infrastructure to deliver connectivity will not work everywhere. This is a bureaucratic challenge, and one which is being discussed behind closed doors.

All of this presents an interesting challenge for the telcos; how do you engage the consumer with an experience which is wholly inconsistent?

The telcos will have to be very careful. Arguably, it is more damaging to steal a customer and not deliver on the experience than not to have the customer at all. Burnt bridges are very difficult to repair after all, especially with the core mobile connectivity offering becoming increasingly commoditised.

Ultimately, 5G will be a necessity for the consumer. Data consumption habits are aggressively growing and 4G will not be able to meet the demands, both in terms of speed and network congestion. That said, the 5G proposition does look hard to justify for the moment. Compatible devices are incredibly expensive, and the network experience looks very limited. It does not appear to be worth the extra expenditure just yet.

  • Private Networks in a 5G World

  • 5G Networking Digital Symposium

  • LIVE: Getting the Best out of 5G

  • 5G Ecosystem Digital Symposium

  • 2020 Vision Executive Summit

  • TechXLR8

  • BIG 5G Event

  • 5G World

  • 5G Latin America


  1. Avatar Dillon Musvamiri 18/11/2019 @ 1:53 pm

    The conclusion is very important for many Telcos – the amount of upgrades required from the power plant to the antennae should be justified by the user experience. It will take a while to reach or come close the laboratory speeds promised.
    I applaud the networks that have pioneered this as we can all learn from their experiences.

  2. Avatar Sypaseuth 19/11/2019 @ 4:23 am

    Worth or worthless should be considered by more elements, not just the telecom network itself.

    5G or any later Generation, will not just to serve the Phone or people, but more about Machine or Artificial element. From this figure, we have to consider our system as three major parts: Brain, Nerve, Sensory where Telecom network taking role of Nerve, AI and BigData taking role of the brain and IoT will be IoT. So the ecosystem will be composed of AI+BigData, 3/4/5G, IoT.

    The detail of 5G perspectives, may be considered by 3 dimensions Throughput(10G/node), Latency(1ms), Connecting(10,000connection/sqkm). Some may refer to Enhance Mobile Broadband(eMBB), Ultra-Reliable Low Latency Communication(uRLLC) and Massive Machine Type Communication (mMTC).

    Why we need those perspectives???? User diven???? Market driven???? Vendor driven????

    For Speed, probably not user belong to User, because we may not enjoy to pay for download whole the movie into the device by just one click and watching just 5 seconds of Introduction part. New Radio (NR) shall be implemented with big change in not just Radio Access Network (RAN), but also Transport layer to convey 5 time bigger throughput from Access layer to Core layer. If we have 100Gbps to serve 4G metro network, we have to build 500Gbps to support 5G.

    Latency clearly Market driven, think about self driving car, Augmented Reality (AR). But how to have that latency while 1ms for front haul, 1ms to cross to back haul, and another 1ms to cross to transport, still another 1ms to core layer and blablabla with the best practice is 5ms. The ideal network should put everything into Cloud platform except the front haul or the CPRI or only Optical fibre from Cloud to Radio Unit. Oh!!! we need to build new metro network just to carry the Optical Lambda to Radio Unit. Big network revolution.

    Connection, probably possible for single Radio Unit to open 10,000 sensors which means Radio Unit have to maintain 10,000 Connection to Cloud. But very costly when multiply Radio Unit to scale of thousands. The cloud platform spend hugely in Memory to collect data from IoT and respond within 1ms by AI and Bigdata. We may realize this function by distributing the AI+BigData into the Machine where the frequently data will be handled by Machine itself, Sound what??!! Yes, Robot.

    Finally, the easiest way to do is robotizing, That means we will minimize the Transport stage by more and more let the Machine to handle it’s job.

Leave a comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.