Author: adrian.chadd
Date: Sun Jul 12 20:18:28 2009
New Revision: 14185

Added:
    wiki/FeatureTproxySupport.wiki

Log:
Created wiki page through web user interface.

Added: wiki/FeatureTproxySupport.wiki
==============================================================================
--- (empty file)
+++ wiki/FeatureTproxySupport.wiki      Sun Jul 12 20:18:28 2009
@@ -0,0 +1,44 @@
+#summary TPROXY - fully transparent client and server proxying support
+
+= Introduction =
+
+TPROXY allows the proxy to spoof both the client and server side IPs,  
allowing for (mostly) fully transparent operation.
+
+The TPROXY support in Lusca is inherited partly from Squid-2 and partly  
from Squid-3.
+
+NOTE: this documentation is still under development.
+
+= Details =
+
+= Platform specifics =
+
+== Linux TPROXY2 ==
+
+== Linux TPROXY4 ==
+
+The TPROXY4 support is in the latest Linux-2.6 kernel (from version?)
+
+== FreeBSD ==
+
+The basic support is in FreeBSD-current. There's FreeBSD-7 patches  
available at http://tproxy.no-ip.org/.
+
+= Differences between Squid-2, Squid-3 and Lusca =
+
+Squid-2 only supports TPROXY2. There are patches available from Visolve  
which implement TPROXY4 support in Squid-2.7.
+
+Squid-3 supports TPROXY4. I'm not certain whether they support TPROXY2 at  
all anymore.
+
+Lusca supports all of the above.
+
+In Squid-3, the "interception/NAT" code which figures out the original  
destination IP address actually tries all of the possible compiled in  
options before failing. Lusca only tries the compiled in option. This means  
that Squid-3 can be compiled with --enable-linux-netfilter -and- TPROXY4  
with no ill effect.
+
+Lusca can only be compiled with one - it will get very confused on TPROXY4  
connections as it will try the netfilter-specific code on a TPROXY4  
connection, get an error, and not correctly complete the request.
+
+
+
+
+= Known issues =
+
+There may be an issue with at least FreeBSD to do with overlapping client  
IP selected ports and the spoofed client IP/port space on the proxy. I need  
to investigate that further and ensure things "work right". (No issue.)
+
+There's a known security issue with transparent interception to do with  
differing DNS and original destination IP. Fixing this via the 'simple'  
method (ie, including the original destination host IP as part of the cache  
key) will both negatively impact on cache hit rates and include large  
amounts of duplicate content into the cache.
\ No newline at end of file

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"lusca-commit" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/lusca-commit?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to