Re: [quicwg/base-drafts] Stateless reset oracle defense (#1554)

[ianswett <[e-mail]>]

ianswett commented on this pull request.

Some editorial suggestions

> @@ -4730,6 +4727,32 @@ receiver, an off-path attacker will need to race the duplicate packet against
 the original to be successful in this attack.  Therefore, QUIC receivers ignore
 ECN codepoints set in duplicate packets (see {{using-ecn}}).
 
+## Stateless Reset Oracle {#reset-oracle}
+
+An attacker that can cause a server to emit a stateless reset token can force
+the closure of any connection that uses the same connection ID and method of
+producing the token.  This creates a possible denial of service attack on

I think of this as a reset injection attack.  Do you think denial of service is a better characterization?

> @@ -4730,6 +4727,32 @@ receiver, an off-path attacker will need to race the duplicate packet against
 the original to be successful in this attack.  Therefore, QUIC receivers ignore
 ECN codepoints set in duplicate packets (see {{using-ecn}}).
 
+## Stateless Reset Oracle {#reset-oracle}
+
+An attacker that can cause a server to emit a stateless reset token can force
+the closure of any connection that uses the same connection ID and method of
+producing the token.  This creates a possible denial of service attack on
+existing connections if the attacker is able to cause a stateless reset token to
+be generated for a connection with a selected connection ID.  That is, an attack
+is possible if an attacker can alter packet routing so that a packet is received
+by an instance that uses the same stateless reset key, but no connection state
+for the connection ID.
+
+To defend against this style of denial service, endpoints that operate in
+clusters and that share a static key for stateless reset (see {{reset-token}})

I think "that operate in clusters and" is unnecessary.  If they share a static key for stateless reset, it's an issue.

> @@ -4730,6 +4727,32 @@ receiver, an off-path attacker will need to race the duplicate packet against
 the original to be successful in this attack.  Therefore, QUIC receivers ignore
 ECN codepoints set in duplicate packets (see {{using-ecn}}).
 
+## Stateless Reset Oracle {#reset-oracle}
+
+An attacker that can cause a server to emit a stateless reset token can force
+the closure of any connection that uses the same connection ID and method of
+producing the token.  This creates a possible denial of service attack on

I find this paragraph difficult to read.  How about reordering it and rewording it slightly:

"Stateless resets create a possible denial of service(reset injection) attack on existing connections if an attacker is able to cause a stateless reset token to be generated for a connection with a selected connection ID.  An attacker that can cause this token to be generated and appear to be on-path can reset any active connection."

I'd move the concept from the last sentence to the end of the next paragraph.  "If a connection ID may be routed to different instances that share a static key, for example by changing the destination IP or port, then an attacker can cause the server to create a stateless reset token."

-- 
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/quicwg/base-drafts/pull/1554#pullrequestreview-141579246