Short answer: stack two to five feeds that each cover a different threat. For a VPN portal, start with a VPN attacker list such as ThreatListPro plus Spamhaus DROP for hijacked networks. Add abuse.ch botnet C2 lists and a scored reputation feed only if you expose other services. A few feeds with little overlap do more than one giant list.
Why Stack Instead of Picking One?
No single IP blocklist covers every threat category. A VPN-specific feed like ThreatListPro targets the sources behind password spraying and credential stuffing against VPN logins. A reputation list like Spamhaus DROP blocks whole netblocks run by criminals. An aggregator like IPsum collects IPs that show up on several public lists at once. Each covers a different part of the problem.
Overlap between these sources can be small. About 99% of the 5.4M+ addresses on our list were on none of 10 free lists we checked (Nov 9, 2026). That is why a VPN-specific feed adds coverage that general lists do not.
Two common mistakes are using one giant aggregated list (more false positives and more entries for the firewall to hold) and stacking many feeds that pull from the same public sources (redundant entries for little new coverage). A short stack of complementary feeds avoids both.
Feed Comparison Table
Details for third-party feeds come from each provider's own page, linked in the first column (checked October 23, 2026). Sizes change daily, so check the provider before relying on a number.
| Feed | Best For | Update Freq | Size | Cost | Provider Notes |
|---|---|---|---|---|---|
| ThreatListPro | VPN brute force, password spraying, credential stuffing | Checked for new attackers every 6 hours | 5.4M+ IPv4 addresses in a few thousand entries | $19.99/mo (14 days free, no card) | Low false-positive risk; IPv4 only |
| Spamhaus DROP | Hijacked and criminal netblocks | Re-evaluated daily; fetch at most once per hour | Small set of netblocks (no fixed count published) | Free | EDROP was merged into DROP on April 10, 2024; use DROP alone |
| abuse.ch Feodo Tracker / SSLBL | Botnet command-and-control servers | Generated every 5 minutes | Varies; can be very small or empty | Free | Use the recommended list; abuse.ch says the aggressive version will cause false positives |
| IPsum Level 3 | General reputation (IPs on 3+ of 30+ public lists) | Daily (every 24 hours) | Varies daily | Free | Higher levels mean fewer, higher-confidence IPs |
| CrowdSec | Local detection plus community blocklists | Blocklists updated several times a day | Varies by blocklist | Free tier (up to 3 blocklists) + paid plans | Firewalls can pull lists over HTTPS via Blocklist Integrations |
| Emerging Threats compromised IPs | Known compromised hosts | Not stated on the download page | Varies | Free | Plain-text file compromised-ips.txt |
| AbuseIPDB | Community-reported abuse with confidence scores | Not stated; each download carries a generated-at timestamp | Free: up to 10,000 IPs; paid tiers up to 500,000 | Free + paid API | Free downloads are fixed at 100% confidence; setting a lower minimum is a subscriber feature |
| FireHOL Level 1 | Conservative general blocking | Rebuilt as its sources update | About 3,900 subnets (list header, Oct 23, 2026) | Free | Built from dshield, feodo, fullbogons and spamhaus_drop; the bogons include private ranges such as 10.0.0.0/8 and 192.168.0.0/16 |
| FireHOL Level 3 | Broader blocking of recent attackers | Rebuilt as its sources update | About 10,900 entries covering about 33,600 IPs (list header, Oct 23, 2026) | Free | IPs reported or detected in the last 30 days; broader than Level 1, which FireHOL describes as the minimum-false-positive option |
| Palo Alto Networks built-in EDLs | Bulletproof hosting, high-risk, known malicious and Tor exit IPs | Delivered through content updates | Set by Palo Alto Networks | Needs a Threat Prevention license | PAN-OS only; contents cannot be edited, but you can exclude entries |
Recommended Stack Tiers
Tier 1: Minimal (start here)
The two-feed stack for a VPN portal: little firewall overhead and low false-positive risk. It suits most environments, from home labs to enterprise networks.
- ThreatListPro — Blocks 5.4M+ IPv4 addresses (attacker IPs and ranges) built from real attacks on live production Palo Alto GlobalProtect VPN portals in the US. It is the feed in this stack aimed specifically at VPN credential stuffing and password spraying.
- Spamhaus DROP — Blocks netblocks that Spamhaus describes as "the worst of the worst IP traffic". These are whole networks, not single IPs, so the coverage is wider than the entry count suggests. The former EDROP list is now part of DROP.
Size: a few thousand ThreatListPro entries plus the DROP netblocks, small enough for any firewall that supports external lists. False positives: Low. Cost: $19.99/month (Spamhaus DROP is free).
Tier 2: Balanced (Production-Grade)
Adds botnet infrastructure blocking and broad reputation scoring. Appropriate for production networks running exposed services beyond VPN (web servers, mail, SSH, RDP).
- Everything in Tier 1, plus:
- abuse.ch Feodo Tracker + SSLBL — Botnet command-and-control server IPs, generated every 5 minutes. Feodo Tracker adds an IP only when it returns a valid botnet C2 response. These lists matter most for outbound traffic: block your own hosts from reaching them. Use the recommended lists, not the aggressive versions.
- IPsum Level 3 — IPs that appear on 3 or more of the 30+ public lists IPsum collects. Requiring several sources removes much of the noise that single lists contain. Rebuilt daily in the stamparm/ipsum GitHub project.
- CrowdSec (optional) — Its Security Engine watches your own logs, blocks attackers as they show up, and shares signals with the CrowdSec community. CrowdSec blocklists are also available as a hosted HTTPS list your firewall can pull. The free tier is limited to 3 blocklists.
Size: depends on the feeds you pick; check each list's current size before adding it. Cost: $19.99/month (the others listed here have free tiers). Operational overhead: Low (CrowdSec's Security Engine adds an agent to manage).
Tier 3: Enterprise (High-Security)
Full-stack threat intelligence with vendor-native feeds, geographic restrictions, and automated response. Requires a SOC, SIEM, or security automation platform.
- Everything in Tier 2, plus:
- Vendor-native intelligence — On PAN-OS, the built-in EDLs (bulletproof hosting, high-risk, known malicious and Tor exit IPs) come with a Threat Prevention license. Other vendors offer their own equivalents. See also our GlobalProtect hardening guide.
- GeoIP blocking — Restrict VPN access to the countries where your organization operates. Most firewalls support country-based address objects. Attackers can route through proxies inside your own country, so treat this as one layer, not a fix.
- SIEM-driven auto-tagging — Use Splunk, Microsoft Sentinel, or Palo Alto Cortex XSIAM to correlate login failures across users and IPs. Auto-tag offending IPs into Dynamic Address Groups (DAGs) for escalating blocks (for example 1h, 24h, 7d, 30d).
What NOT to Stack
Certain feed combinations cause more problems than they solve:
- Several aggregated lists at once (FireHOL Level 2 + Level 3 + IPsum) — These draw on many of the same public sources, so you add entries and false-positive risk for little new coverage.
- FireHOL Level 1 on internal interfaces — Its fullbogons source includes private ranges such as 10.0.0.0/8 and 192.168.0.0/16. Apply it only to traffic arriving from the internet, or exclude your own networks first.
- Ten free feeds "because they are free" — More feeds do not mean more security. A short stack with little overlap is easier to troubleshoot and keeps false positives down.
How to Deploy This Stack on Your Firewall
The specific steps depend on your firewall platform, but the pattern is the same everywhere:
- Create one external list object per feed (EDL, threat feed or URL table), each with its own URL and refresh interval. Respect each provider's limits: Spamhaus asks for no more than one fetch per hour, and SSLBL no more than one every 5 minutes.
- Reference the inbound lists (ThreatListPro, DROP, IPsum) as the source in a deny rule. On PAN-OS you add the EDL objects directly to the rule.
- Reference the botnet C2 lists (Feodo Tracker, SSLBL) as the destination in an outbound deny rule.
- Place the deny rules early in your security policy, before any allow rules for the affected zone.
- Commit and verify that entries are loading. To see whether a given address is already on our list, use the IP check tool.
For platform-specific setup instructions, see our firewall guides:
- Palo Alto GlobalProtect EDL Setup
- FortiGate External Threat Feed
- pfSense pfBlockerNG Configuration
- OPNsense URL Table Alias
- Cisco AnyConnect Security Intelligence
- SonicWall Dynamic Block List
Frequently Asked Questions
How many IP blocklist feeds should I use on my firewall?
Most environments do well with 2–5 feeds. A VPN-specific feed (ThreatListPro, 5.4M+ IPv4 addresses) plus Spamhaus DROP for hijacked networks covers a VPN portal. Add 1–3 more feeds only if you expose services beyond VPN, and only if each one covers a threat the others miss.
Should I use ThreatListPro alongside CrowdSec?
Yes, they work well together. ThreatListPro is a hosted list your firewall pulls, built from real attacks on live production Palo Alto GlobalProtect VPN portals in the US. CrowdSec's Security Engine detects attacks in your own logs as they happen and shares signals with its community. One blocks IPs already seen attacking VPN portals; the other reacts to what hits you.
Is IPsum a good free alternative to a paid blocklist?
IPsum is a free aggregator built from 30+ public lists and refreshed daily. Level 3 keeps IPs that appear on 3 or more of those lists, which cuts noise. It is not VPN-specific, so pair it with a VPN-focused feed such as ThreatListPro if VPN logins are your main concern.
What is the difference between ThreatListPro and Spamhaus DROP?
Spamhaus DROP lists netblocks that Spamhaus considers hijacked or run by criminals, so it blocks whole networks. ThreatListPro lists attacker IPs and ranges built from real attacks on live production Palo Alto GlobalProtect VPN portals in the US. They target different threats, which is why stacking them makes sense.
Will stacking multiple blocklists slow down my firewall?
A short stack stays small: ThreatListPro covers 5.4M+ addresses in a few thousand entries, and Spamhaus DROP is a small set of netblocks. The practical limit is how many external-list entries your firewall model supports, so check your vendor's documentation and avoid aggregated lists that add hundreds of thousands of entries.