By CalMax Systems Security Research, founder · Published April 17, 2026 · Last verified: October 28, 2026

A residential proxy VPN attack routes password guesses through rented home IP addresses, so each address sends only a few attempts and looks like an ordinary user. To detect it, alert on total failed logins across all source IPs. To limit it, block known attacker IPs at the firewall, require MFA, and require client certificates where possible.

ThreatListPro is one option for the blocking layer: it lists IPv4 sources seen attacking live production Palo Alto GlobalProtect VPN portals in the US. A residential address that has never attacked those portals can still reach yours before it is listed, so pair the list with login monitoring across all source IPs and with MFA.

VPN brute force defense has traditionally relied on a straightforward model: identify the IP addresses of known attackers, often hosted in datacenters and cloud providers, and block them at the firewall. That still works against attackers who send many attempts from a few servers.

Residential proxy networks weaken that model. The traffic looks legitimate because it is coming from consumer ISP address space. And because each IP sends only a few attempts before rotating, per-IP rate limits and lockout thresholds rarely trigger. This technique is also tracked in MITRE ATT&CK: residential endpoints used as relays appear under T1090.002 External Proxy, and the login attempts themselves under T1110.004 Credential Stuffing and T1110.003 Password Spraying.

What Are Residential Proxy Attacks?

A residential proxy network is a service that routes internet traffic through residential IP addresses: the addresses consumer ISPs assign to home connections and mobile devices. Commercial providers sell access to large pools of these addresses, marketed mainly for web scraping, ad verification, and market research. Attackers buy the same access, or use networks built from compromised devices.

The addresses come from several sources. Some come from apps that embed a proxy SDK, turning users’ devices into proxy exit nodes under terms few people read. Others come from browser extensions, free VPN apps, or compromised routers and IoT devices. The result is a large pool of IP addresses that belong to real ISP customers and often pass reputation checks designed to spot datacenter or hosting infrastructure.

For credential attacks, this changes the shape of the traffic. Instead of sending thousands of login attempts from one server that gets blocked quickly, the attacker spreads them across many residential IPs. Each IP sends a few attempts, below a typical lockout threshold, and the tool then rotates to a fresh address. The VPN portal sees what looks like many separate users mistyping their passwords, not one coordinated attack.

Why it is cheap enough to work: Residential proxy access is usually sold by traffic volume, and a login attempt uses very little traffic. Routing a credential campaign through residential addresses costs little compared with the value of one working VPN credential that grants internal network access.

Why Traditional Blocklists Miss Residential Proxies

Traditional IP blocklists are built on a specific assumption: malicious traffic comes from infrastructure that looks different from legitimate user traffic. Datacenter IP ranges are well-documented. Cloud provider address space is published in machine-readable formats by AWS, Azure, GCP, and others. Hosting providers register their IP blocks with regional internet registries. This makes datacenter-sourced attacks relatively easy to identify and block.

Residential proxy traffic weakens this model. Here’s why:

The result is attack traffic that looks like legitimate users when you examine one IP at a time. Detection has to look at the whole portal instead.

From our own list (November 9, 2026): We looked up the network behind every single attacker IP on the ThreatListPro list using Team Cymru’s IP-to-ASN lookup. About 4 in 10 are on networks run by ordinary home and mobile internet providers. Most of the rest are on hosting and data-center networks. A hosting network can often be blocked whole; a home ISP cannot, so the home and mobile addresses have to be blocked one at a time, after they have been seen attacking.

The “Low-and-Slow” Attack Pattern

Understanding the mechanics of a residential proxy brute force campaign shows why per-IP defenses miss it. Here is a step-by-step breakdown of a typical attack:

Step 1: Acquire Credentials and Proxy Access

The attacker obtains a credential list, typically username and password pairs leaked in earlier data breaches, or a list of usernames built from your email format for password spraying. They add residential proxy access, either bought from a provider or rented from a criminal proxy service.

Step 2: Distribute Across Residential IPs

The attacker configures a credential-testing tool (custom scripts, or tools such as Sentry MBA or OpenBullet) to send login attempts through a rotating pool of residential proxy IPs. The tool is set to send only a few attempts per IP before moving to a new address.

Step 3: Low-and-Slow Execution

The campaign runs over hours or days, with random delays between attempts to look like a person typing and retrying. In your logs, the result looks like this:

Step 4: Harvest Valid Credentials

When a credential pair succeeds, the attacker logs the working combination and the VPN portal it grants access to. They may not log in right away. Valid credentials can be stockpiled, sold, or used weeks later, when nobody connects the original brute force campaign to the eventual breach.

Detection Strategies

Detecting residential proxy attacks requires a fundamental shift in how you analyze VPN authentication logs. Per-IP analysis is insufficient. You need aggregate, behavioral, and fingerprint-based detection working together.

Aggregate Login Failure Analysis

The most useful starting point is shifting from per-IP failure counts to aggregate failure counts. Instead of only alerting when one IP fails many times, also alert when the portal as a whole sees far more failed logins, from far more unique IPs, than it normally does. Every organization has a baseline of failed logins from mistyped and expired passwords. A residential proxy campaign pushes the total well above it, even though no single IP stands out.

Configure your SIEM to learn your own normal failed-login rate and unique-source count per hour, and alert on sustained deviations. Pick the thresholds from your own data rather than from a generic number.

Geographic Anomaly Detection

When the attacker does not choose exit countries, residential proxy traffic arrives from far more countries than your users do. Legitimate remote workers are in predictable locations, so a sudden spread of failed logins across many countries is a useful signal. Because many proxy services let the buyer pick a country, treat a normal-looking geography as inconclusive, not as proof the traffic is clean.

Build a geo-baseline of your normal VPN authentication sources. Alert when the number of unique countries rises well above that baseline, or when traffic appears from countries where you have no employees or operations.

Timing Pattern Analysis

Automated tools, even with random delays, often produce timing patterns that differ from real users. Look for uniform intervals between login attempts from different source IPs—this suggests a single tool cycling through a proxy list. Real users do not arrive at your VPN portal in evenly-spaced waves from different countries.

Advanced detection correlates inter-arrival times of failed logins across all source IPs. Legitimate failed logins tend to cluster around your users’ working hours (for example, the Monday morning login rush). A campaign that runs steadily through nights and weekends stands out against that baseline.

TLS and User-Agent Fingerprinting

Most residential proxies tunnel the connection, so the TLS handshake and HTTP headers come from the attacker’s tool, not the proxy device. Many “different” residential IPs can therefore present the same JA4 or JA3 TLS fingerprint and the same User-Agent string.

There is a catch: your own users also share fingerprints if they all run the same managed VPN client. Record the fingerprints your real clients produce, then look for a fingerprint that is not one of them appearing from many unrelated IPs. VPN gateways rarely log TLS fingerprints themselves, so this usually needs a WAF, reverse proxy, or network sensor in front of the portal.

Username Pattern Analysis

Legitimate users attempt their own username. A brute force campaign tries multiple usernames from a single IP or, in the residential proxy case, sequences of different usernames from different IPs in the same credential list order. If your SIEM can correlate the order in which usernames appear across source IPs, you may detect the sequential pattern of a credential list being worked through.

Defense Stack for Residential Proxy Attacks

No single defense stops residential proxy attacks. You need layers that address different parts of the attack. Here is a five-layer stack, in a practical order of deployment:

Layer 1: VPN-Specific Blocklist

Start with a blocklist built from VPN brute force attacks, such as ThreatListPro. This layer refuses IPs that have already been seen attacking our own GlobalProtect VPN portals, whether they sit in a datacenter or a home ISP, before they reach your login page. It cannot block a fresh residential address that has not attacked anyone yet; the layers below handle those.

ThreatListPro is built from real attacks on live production Palo Alto GlobalProtect VPN portals in the US, covers 5.4M+ IPv4 addresses, and is checked for new attackers every 6 hours. It loads as an External Dynamic List (EDL) on your firewall in a few minutes. About 99% of its addresses were on none of 10 free lists (Nov 9, 2026).

Layer 2: Behavioral Detection

Deploy the aggregate analysis techniques described above: total failed login monitoring, geographic anomaly detection, timing pattern analysis, and TLS fingerprinting. This is your main detection layer for traffic that passes Layer 1, including residential proxy traffic that no static blocklist can catch in advance.

Feed these signals into your SIEM and create alerts. When a distributed campaign is detected, you can respond by requiring stronger authentication through your identity provider, temporarily restricting portal access to expected countries, or tightening access to known device populations during the attack.

Layer 3: MFA Everywhere

Multi-factor authentication means a guessed or stolen password is not enough on its own to connect. It is the most important long-term defense against credential attacks of every kind, not just residential proxy campaigns.

Deploy MFA on every VPN account with no exceptions. Authenticator app codes or push with number matching are a reasonable minimum; FIDO2/WebAuthn security keys resist phishing and live relay attacks best. MFA does not stop failed password attempts from counting toward Active Directory lockout, so see how to stop VPN account lockouts if campaigns are locking users out.

Layer 4: Client Certificate Authentication

Client certificate authentication takes password guessing alone off the table. Each authorized device receives a client certificate issued by your internal PKI, and the portal is configured to require it in addition to the user’s credentials, not as an alternative. A connection without a valid certificate fails before a password is checked.

This shifts the first check from “something you know” (a password that can be guessed) to “something you have” (a certificate on a managed device). A residential proxy campaign can still reach the portal, but credential guessing alone no longer gets in. For GlobalProtect settings, see the GlobalProtect hardening guide.

Layer 5: Residential Proxy Reputation Services

If you need more coverage against fresh residential exits, add a commercial IP reputation service that flags addresses currently acting as residential proxy exit nodes. This complements a VPN attack list, which only contains addresses that have already attacked.

The tradeoff is collateral damage. An address flagged as a proxy exit node today may be an ordinary user’s home connection tomorrow, after the proxy app is removed or the ISP reassigns the address. Use these signals to raise the authentication challenge (for example, an extra MFA step) rather than to hard-block traffic.

The layered approach: Layer 1 refuses addresses already seen attacking the VPN portals the list is built from. Layer 2 detects distributed campaigns that get through. Layer 3 makes a guessed password insufficient on its own. Layer 4 requires a device certificate before the password matters. Layer 5 adds residential proxy signals for addresses with no attack history yet.

How ThreatListPro Helps

ThreatListPro lists IP addresses observed attacking live production Palo Alto GlobalProtect VPN portals in the US. When a campaign uses datacenter or hosting IPs, or keeps reusing the same residential exits while it runs, those addresses are blocked at your firewall once they are on the list.

Residential proxy addresses are reassigned often, so residential entries matter most while a campaign is active. The list is checked for new attackers every 6 hours, which means a brand-new residential exit can reach your portal before it is listed.

That is why the list belongs in a stack rather than standing alone: ThreatListPro for addresses with observed VPN attack history, aggregate login monitoring for fresh addresses, MFA and client certificates so a guessed password is not enough, and optionally a residential proxy reputation service for extra challenges.

Deployment time: ThreatListPro loads as an External Dynamic List (EDL) on your firewall in a few minutes, with no software to install and no changes to your VPN configuration. Your firewall re-fetches the list on the schedule you set; the list republishes every 6 hours.

Frequently Asked Questions

What is a residential proxy attack on VPN?

A residential proxy attack on a VPN is a password-guessing campaign routed through home IP addresses rented from residential proxy networks. Each address sends only a few login attempts before the attacker rotates to another, so per-IP rate limits and lockouts rarely trigger, and the traffic comes from ordinary ISP address space instead of datacenter ranges.

Why can’t traditional IP blocklists stop residential proxy attacks?

Most blocklists are built from addresses with a history of abuse, and much of that history comes from datacenter and hosting ranges. Residential proxy addresses belong to consumer ISPs, so blocking whole ISP ranges would cut off real users, and a rotating address may have no abuse history when it first reaches your portal. Per-IP rate limits also miss them because each address sends only a few attempts.

How do I detect low-and-slow brute force attacks on my VPN?

Count failures across all source IPs, not per IP. Alert when total failed VPN logins rise well above your own baseline even though no single IP crosses your lockout threshold. Then look for many usernames tried from many new IPs, attempts from countries where you have no users, and many IPs presenting the same unfamiliar TLS fingerprint (JA4 or JA3) where a WAF or network sensor records it.

Does ThreatListPro block residential proxy IPs?

Partly. ThreatListPro lists IP addresses seen in real attacks on live production Palo Alto GlobalProtect VPN portals in the US, whatever network they come from, and it is checked for new attackers every 6 hours. A residential address that has not attacked those portals yet will not be on the list, so pair the list with aggregate login monitoring and MFA.

What is the best defense stack against residential proxy VPN attacks?

Layer the controls: (1) block known VPN attacker IPs at the firewall with a list such as ThreatListPro, (2) alert on total failed logins across all source IPs, (3) require MFA on every VPN account, (4) require a client certificate before the password is checked where your VPN supports it, and (5) optionally use a residential proxy reputation service to add an extra challenge rather than a hard block.