Google Stadia: The implications for 5G and Edge Computing
On Tuesday March 19th, Google announced its cloud-based gaming platform, Stadia. The initial announcement and discussion was aimed at the gaming developer community whose help and buy-in Google will need to make Stadia possible. Few game titles means few users and fewer revenues. While I am not a big game player (except for a couple on my phone I play waiting for a plane), I do appreciate the revenue potential of games and the appeal. In fact, the ‘entertainment’ category in our 5G revenue forecast includes a sizable chunk from 5G-enabled gaming.
So I read the announcement of Google Stadia with some interest, especially the part about 4K video quality and needing a low latency broadband connection. Google says the service will work on a 25 Mbps broadband connection. iGR has just published its annual ‘Home Broadband and Wi-Fi’ forecast report and you may be surprised to know that not all homes in the U.S. can get 25 Mbps. In fact, the bottom quartile of users (ranked by speed) ends at just 12 Mbps; this means that the bottom 25 percent of homes in the U.S. have just 12 Mbps or less for their broadband connection.
This week, T-Mobile also announced a trial of 50,000 homes for its new LTE-based home Internet service offering about 50 Mbps. The full launch is now scheduled for the second half of the year. So the good news for a lot of homes is that they will be able to get a bump in their Internet speeds for a reasonable $50 a month.
OK, so what does this mean for Google Stadia? The good news is that T-Mobile’s new home Internet service should be able to support both a Stadia player (25 Mbps) and a Netflix stream, plus a little browsing, all at the same time. So the family should be happy. But for a conventional mobile LTE connection, speeds may be insufficient. I just tested my LTE connection from my office (using an iPhone XR) and got a download of just 9.10 Mbps. I have seen speeds higher than this, but I would say that 10 - 15 Mbps is typical. Far short of the 25 Mbps Google says Stadia needs. (Note that when I walked outside my office, I got 62 mbps and 36ms latency).
Of bigger concern for LTE is latency. The test in my office gave a latency of 45 ms and jitter of 410 ms. Compare this to my broadband connection on my office iMac that shows a ping of 21 ms (still not great).
The serious gamers out there will say “play Stadia at home on a broadband connection” and I would agree - serious gamers are going to do just that. But as more homes become ‘mobile Internet only’ (a surprising number of homes are dumping wired Internet and using mobile-only), move to a T-Mobile Internet service or adopt 5G for the home, latency is going to become a bigger issue for an increasing number of gamers. Bandwidth will be a lesser issue for some.
Latency will become a real issue if future versions of cloud gaming platforms add AR/VR support, as some console platforms are doing now. Now we need to get latency of 10 ms or lower to avoid motion sickness.
The answer is two-fold: 5G gives an increase in bandwidth we all know about; and edge computing architectures will reduce the latency. It was interesting that Google made no mention of 5G or edge compute in its announcement. iGR believes this support will come; it would be easy to see a partnership between Google and one of the major mobile operators to couple cloud gaming, 5G and edge compute into an integrated product. As we discussed in the last Opinion when we talked about 5G pricing, this could be bundled as one service. The same applies to Microsoft, who has the planned Project xCloud.
This seems like an ideal partnership: Google, Microsoft, Nvidia, Sony etc. partner with select mobile operators to enable cloud gaming. The mobile operators invest in edge computing and 5G and fully enable cloud gaming across mobile networks to a wide range of devices in a wide range of locations. This would generate revenue for all parties and provide a strong initial application for edge compute in the mobile network. To iGR, this seems like a natural evolution. The next big question is which cloud platforms will partner with which mobile operators? That remains to be seen but we have some ideas.