Uber cut its corporate network costs in half, HPE says. Video calls got faster.
Hewlett Packard Enterprise published the customer film on 1 October. In it, an Uber infrastructure engineer says the company runs more than 200 offices in five regions for over 30,000 employees.
He says office traffic used to travel over point-to-point links to a data centre first, which "was introducing a massive delay". Most of Uber's corporate servers have since moved to the cloud, he says, and Uber deployed HPE's Mist for its corporate network. He describes Mist as one platform for SD-WAN, Wi-Fi and wired networking that uses AI to correlate events.
HPE's description of the film puts the result at 50% lower network infrastructure costs and latency down from 100 ms to 8–10 ms. The engineer says users noticed it immediately, especially on video calls.
The film does not say what the 50% is measured against or over what period. It also does not separate the effect of dropping the data-centre detour from that of the AI tooling. The figures are the vendor's, and the ride scenes are its promotional footage.
Which office applications still route through your data centre, and why?
🎥 Video: HPE, official YouTube channel (customer film) · narration: AI voice
What the captions say
- Office traffic once went via data centres.
- 200 offices, five regions, 30,000 staff.
- Most servers moved to the cloud.
- Uber switched to HPE's Mist SD-WAN.
- HPE says costs fell 50%.
- Lag fell from 100 to 8–10 milliseconds.
- Staff noticed it most on video calls.
- Baseline for the 50%? The film doesn't say.
