Hi there, guest! If you have an account, then please sign in or register in order to see more.
Real DDoS Attacks We Absorbed: 40 Gbps and 21 Waves
5 min read
ESAGAMES Team
Most DDoS write-ups use round numbers and stock graphs. This one uses the figures our mitigation platform recorded for two attacks on ESAGAMES customers in mid-November 2026: a 26-hour campaign of 21 waves against a single VDS, and two back-to-back floods of up to 40.4 Gbps against a game server. In both cases the services stayed online.
How we measured
Every figure below is the peak our mitigation platform recorded for each attack wave, from sampled flow data (1:5,000 sampling), so treat them as close estimates rather than exact counts.
Gbps is volume and Mpps is millions of packets per second. Both matter: volume fills links, packets per second exhaust routers and servers.
Times are UTC. Customer names, IP addresses and services are not disclosed.
Case 1: one VDS, 21 waves in 26 hours
Recorded
Target
One VDS (Linux), one IPv4 address
Window
14 Nov 17:50 to 15 Nov 20:02 UTC
Waves
21, from 1 to 13 minutes each, about 115 minutes of attack traffic in total
Biggest by volume
17.7 Gbps / 13.0 Mpps (14 Nov 19:23, UDP flood)
Biggest by packets
30.2 Mpps at 13.7 Gbps (14 Nov 17:50, TCP SYN + UDP, about 57-byte packets)
Typical wave
7 to 15 Gbps of UDP, 160 to 180-byte packets
Result
The VDS stayed online
This was not one big hit but a campaign: a burst every 5 to 15 minutes on the first evening, a pause, a single wave after midnight, then a second evening of five waves. Three tiny UDP fragmentation probes (under 0.1 Gbps) between 19:28 and 19:34 on the first evening look like the attacker checking whether the protection was still active before trying again.
Campaigns like this are common against game communities: the goal is to wear the target down, not to win in one shot. That is why protection has to be always-on, with no “switch it on” delay for each new wave.
Case 2: a game server, two floods 7 minutes apart
Recorded
Target
One game server
Wave 1
15 Nov 12:49 UTC: 28.6 Gbps / 19.8 Mpps, UDP + UDP fragmentation, about 2 minutes
Wave 2
15 Nov 12:56 UTC: 40.4 Gbps / 27.9 Mpps, same vectors, about 2 minutes
Average packet
About 181 bytes
Result
The server stayed online
The second wave came seven minutes after the first and was 41% larger: a classic “it didn't work, try harder” escalation. 40 Gbps is far beyond what a single server's uplink can take on its own, which is why the filtering has to happen upstream, before the traffic reaches the node.
Why packets per second matter as much as Gbps
On 16 November at 12:32 UTC another game server received a UDP flood of only 15.8 Gbps but 47.3 million packets per second, the highest packet rate of the period, with an average packet of about 42 bytes. Small-packet floods like this barely look “big” in Gbps, but they are often harder on routers, firewalls and game server processes than a larger flood of big packets. When a host only quotes Gbps, ask about packets per second too.
What this means for youAttacks on gaming and hosting customers are frequent, repeated and escalating, so protection has to be always-on, and you should look at both numbers: volume (Gbps) and packet rate (Mpps). Our network filters at the edge of AS214918, across 17.8 Tbps+ of scrubbing capacity, with in-house XDP filtering. See how our Anti-DDoS works.
Running a community that keeps getting hit? Our DDoS-protected VDS and game server hosting in Frankfurt sit behind the same filtering described above.
Host behind a protected network
Game servers and VDS in Frankfurt, filtered at the edge of AS214918 by default.
The largest single wave reached 40.4 Gbps and 27.9 million packets per second against a game server. The highest packet rate was 47.3 million packets per second at 15.8 Gbps, and one VDS received 21 separate waves in 26 hours.
Did the attacked services go offline?
No. In both cases described here the services stayed online while the attacks were filtered.
Why look at packets per second and not only Gbps?
Because floods of small packets can overwhelm routers, firewalls and game server processes even at modest volume. The 47.3 Mpps flood in this article was only 15.8 Gbps.
BepInEx, QoL, building and content mods — the best Valheim mods to run on a dedicated server this year, with a starter pack and sync rules that stop join failures.
Super Structures, Spyglass, Cryopods and more — the best ARK mods to run on a server this year, with load order and the ASE-vs-ASA differences that trip people up.
SourceMod does not support CS2 — not in 2026, not through a bridge. Here is what actually replaced it.
Action required
Let's talk about privacy.
We use essential cookies to keep checkout, login sessions and security flows working. Optional cookies help us understand performance, remember interface choices and enable external support widgets only when you allow them.