Carnegie Mellon University

Alt Text for Image

Network Latency Segmentation Project

Contact Name

The code behind:

“Segmenting Latency in a Private 4G LTE Network”, Sophie Smith, Ishan Darwhekar, James Blakley, Thomas Eiszler, Jan Harkes, Carnegie Mellon University Technical Report CMU-CS-22-115, May 2022.


Many edge-native applications require low and predictable end-to-end network latency. In practice, many user-interactive edge applications must deliver less than 50ms round trip times (RTT) from the client device to an edge cloudlet and back to the client device to achieve acceptable user experience. More intensive interactive applications like virtual reality require less than 20ms RTTs. However, commercial 4G LTE networks fail to reliably meet these thresholds. This report summarizes the results of our measurement of the sources of network latency in an operational private outdoor CBRS 4G LTE network to provide a baseline for future network latency optimization.

NetworkLatencySegmentation on Github