RIPE NCC observed DNS steering directing selected PlayStation 5 and Xbox GTA VI downloads across multiple content delivery networks (CDNs) during the November 2026 pre-load period. Its measurements describe those routes; researchers lacked visibility into network traffic loads and could not identify where or why service issues might have occurred. RIPE NCC’s study followed paths to selected download endpoints, not total internet traffic.

A route map shows where download requests were directed; understanding network disruptions also requires visibility into the traffic those networks carried.

How GTA 6 downloads moved across networks

DNS, the system that translates a website or service name into an internet address, helped direct console download requests to delivery endpoints. Location-aware DNS, sometimes called GeoDNS, can steer a request toward an endpoint based on where it originates. CDNs distribute content across networks and endpoints, so downloads need not all travel from one central server.

RIPE NCC observed Akamai and Fastly on both PS5 and Xbox paths. Its probes also encountered Google and Amazon on PS5 paths, and G-Core on Xbox paths. These were observations from RIPE Atlas probes, not market-share estimates for either platform.

What RIPE NCC measured

Starting November 12, 2026, RIPE NCC ran traceroute measurements every four hours to selected PlayStation and Xbox download-steering hostnames. Traceroute records the network path traffic takes toward a destination. The study tracked resolved IP addresses, paths, and latency over both IPv4 and IPv6, using RIPE Atlas, a network of more than 15,000 probes.

The probe hosts were concentrated most heavily in Europe and tended to be technically inclined users. That distribution matters when interpreting which delivery networks appeared in the sample: the observed mix describes those probes and routes, not all console downloads worldwide.

How often measured routes crossed an internet exchange

An internet exchange point (IXP) is a place where networks connect and exchange traffic. RIPE NCC recorded an IXP on different shares of measured paths depending on the console and IP version:

ConsoleIP versionMeasured paths crossing an IXP
PS5IPv411.2%
PS5IPv621.4%
XboxIPv419.9%
XboxIPv629%

These percentages count measured paths, not the share of download traffic carried by IXPs. They describe how the routes in the probe sample connected, rather than how much data moved over each route.

What the latency figures say

RIPE NCC measured a median latency of 8 milliseconds on paths that crossed an IXP and 10 milliseconds on other paths. Median latency is the middle value in a set of measurements, rather than an average.

The overall difference largely disappeared when RIPE NCC compared paths within the same network. The two-millisecond gap is therefore an observed association in the broader set of routes, not evidence that crossing an IXP caused a download to be faster.

Can these measurements tell us whether outages occurred?

RIPE NCC’s measurements characterized selected download paths, while the researchers lacked visibility into traffic loads moving into and out of networks, including transit links and private interconnects. They could not identify where or why network problems might have occurred, so the study does not provide a comprehensive account of service availability during the pre-load period.