274,153Events scored in 19 days
12%Of permitted traffic was malicious
8,934Critical & high events intercepted
3,016Unique critical & high sources
What the firewall saw
  • 200,434 events permitted, 74,659 blocked
  • Rules based on ports, protocols and policy
  • All 10 highest-scoring threat actors allowed through
What OneFirewall added
  • 24,132 malicious events found inside permitted traffic
  • 8,934 critical and high events intercepted
  • 23,654 critical events already blocked by the checkpoint-ip feed
The Challenge

A web application is never only a web application

Confidential Org's customers reach the service over HTTP and HTTPS. The same hosts also expose SSH, SFTP, FTP and RDP for administration and file exchange, protected by two Check Point gateways and two FortiGate firewalls. Three quarters of the traffic is inbound from the internet.

  • Firewall rules allow by design — a web service has to accept connections from anyone, so port-based policy can't tell a customer from a scanner.
  • Six exposed services, evenly attacked — HTTP, HTTPS, RDP, SSH, FTP and SFTP each carried around 46,000 events. Protecting port 443 alone would cover only a third of the attack surface.
  • Thousands of one-shot sources — 4,137 unique low-severity IPs produced 4,139 events: almost every source touched the perimeter once, then moved on. Reacting after the first hit is already too late.
  • No context on who is knocking — the firewall logs an allowed connection, not whether that IP was reported by other alliance members an hour earlier.
Traffic Mix

What crossed Confidential Org's perimeter

OneFirewall scored every event against alliance intelligence. 88% of permitted traffic was clean. The other 12% carried a crime score, and almost 9,000 of those events were critical or high.

200,434
Events permitted by the firewall
74,659
Events blocked by the firewall
23,654 by OneFirewall checkpoint-ip intelligence
176,302
Clean events
No threat match
24,132
Malicious events inside permitted traffic
Permitted vs blocked
Permitted · 200,434 (73%)Blocked · 74,659 (27%)
Clean vs malicious (of permitted)
Clean · 176,302 (88%)Malicious · 24,132 (12%)
Traffic by direction
Inbound · 206,636 (75%)Outbound · 68,457 (25%)
Severity

Malicious permitted traffic by crime score

Each event is scored 1–1000 by the OneFirewall Crime Score. Critical and high sources are the ones to block at the edge without review.

SeverityScore rangePermitted eventsShareUnique IPsAlready blocked
Critical175 - 10007,35730%2,34323,654
High120 - 1751,5777%6735,188
Medium60 - 12011,05946%2,62935,258
Low1 - 604,13917%4,1379,261

Already blocked: malicious events Confidential Org's firewalls had already stopped thanks to OneFirewall intelligence feeds.

Services, Geography & Devices

Where the pressure landed

Top services by events

Attackers spread their effort almost perfectly evenly. Web traffic (HTTP and HTTPS) was 91,925 events, but RDP, SSH, FTP and SFTP together drew twice that.

HTTP46,013
RDP45,990
HTTPS45,912
SSH45,848
FTP45,847
SFTP45,483

Top origin countries

The United States led origin countries by a wide margin, followed by Western European hosting hubs. Attack infrastructure is usually rented, not local.

United States865
The Netherlands277
United Kingdom181
Belgium175
China175
Germany155

Top devices by events

DeviceVendorEvents
Check Point gateway 1Check Point874
FortiGate firewall 1Fortinet844
Check Point gateway 2Check Point844
FortiGate firewall 2Fortinet387
Top 10 Threat Actors

Every one of them was allowed in

The ten highest-scoring sources in the period all scored 844 or more out of 1000. Confidential Org's firewall allowed every one of their connections. OneFirewall flagged every one as critical. Last octets are masked.

#Source IPCrime ScoreHitsFirewall actionServiceCountry
1195.178.110.x97121AllowHTTPThe Netherlands
2102.220.161.x9544AllowFTPNigeria
337.120.213.x9461AllowSSHSwitzerland
4119.28.84.x9357AllowSFTPHong Kong
5179.43.153.x8882AllowFTPSwitzerland
6156.227.234.x8887AllowHTTPJapan
7102.220.161.x8724AllowFTPNigeria
8195.178.110.x8461AllowHTTPSThe Netherlands
9107.150.123.x8451AllowSSHHong Kong
1092.118.39.x8441AllowRDPRomania
Two subnets, four of the top ten. 195.178.110.x (The Netherlands) and 102.220.161.x (Nigeria) each placed two different IPs in the top ten, hitting HTTP/HTTPS and FTP respectively. Attackers rotate addresses within the same rented range. Alliance intelligence tracks the range as well as the single IP, so the next address in the block is already known.
Findings

Five things Confidential Org's data tells us

Finding 1 — The firewall is not the filter

12% of allowed traffic was malicious

24,132 of 200,434 permitted events matched a threat. A firewall enforces policy. It doesn't know reputation. Without intelligence, those events reach the web application and the admin services behind it.

Finding 2 — The web is a third of the target

RDP and SSH were hit as hard as HTTP

The six exposed services each took between 45,483 and 46,013 events. Attackers sweep every open port with the same tooling. Web application protection must cover the whole host, not just the HTTP listener.

Finding 3 — The worst actors were quiet

Top-scoring IPs made as few as one connection

Five of the top ten threat actors made one or two connections in 19 days. Rate limits and anomaly detection never trigger on a single request. Only prior intelligence (who this IP is and what it did elsewhere) can stop the first hit.

Finding 4 — The long tail is huge

4,137 low-severity IPs, 4,139 events

Nearly one event per source. That's the signature of distributed scanning: a different address every time. Blocking it means using intelligence shared across 290+ alliance members, not waiting to see an IP twice.

Finding 5 — Intelligence already pays for itself

23,654 critical events stopped before they started

The checkpoint-ip feed had already blocked 23,654 critical events at Confidential Org's Check Point gateways, more than three times the critical events still arriving through permitted traffic.

The Results

What OneFirewall delivered for Confidential Org

8,934

Critical and high events intercepted that the firewall had permitted.

3,016

Distinct critical and high sources identified and ready to block at the edge.

23,654

Critical events already stopped automatically by the checkpoint-ip intelligence feed.

176,302

Clean events left untouched, so legitimate users of the web application kept full access.

6

Exposed services covered: HTTP, HTTPS, RDP, SSH, FTP and SFTP, all scored by the same intelligence.

2

Firewall vendors (Check Point and FortiGate) enforcing one shared blocklist, with no vendor lock-in.

“The firewall did exactly what it was configured to do: allow traffic to a public service. OneFirewall told us which of those allowed connections we should never have accepted.”

Confidential Org · Security operations

How much malicious traffic is your firewall allowing?

Run a free Proof of Value against your own traffic and get a report like Confidential Org's. No data leaves your perimeter.

Start a Proof of Value
Protecting a web application at the edge? See Web Application Protection, or read how a global SaaS platform stopped 35,000 daily attacks in 24 hours.