Explanation of Transparent Mode Deployment (Bridge or Proxy)


This article applies to:

  • WebDefend (All Versions)

Question:

  • What is the difference between bridge and proxy mode?

Information:

The Getting Started guide outlines setting a bridge IP address for the appliance, but it does not explain why you need one.

The WebDefend appliance is "transparent" from the protected web host's perspective (as it spoofs the client source IP). However, it is not totally transparent from a networking perspective (unlike some stealthy network IDS or other transparent devices).

Here is a little look "under the hood."

The appliance is a normal transparent bridge for all traffic initially. The default IPTables rules on the appliance bridge all traffic when you first boot up and have the external/internal Ethernet cables plugged in. Of course this setup does not perform any protective functions.

In order to set the appliance to act as a gateway for HTTP/HTTPS traffic, you use the Command Line Interface (CLI) to define the IP addresses of the back-end web application hosts that you want the appliance to act as a WAF for. When you do that, the appliance creates custom IPTables interception/redirection rules so that this traffic will not be bridged as other traffic but instead will be redirected to a bridged LAN network range that is internal to the appliance. This internal network is where Apache/ModSecurity is running to inspect and proxy the traffic. If the inbound/outbound data is allowed to pass, the appliance will spoof the SRC IP address and send the traffic to the other side of the bridge.

In short, for intercepted traffic the appliance functions as a "Transparent Proxy" instead of a "Transparent Bridge."

Another reason why WebDefend is not totally transparent is that the appliance has a "Bridge IP" address that is set appropriately for the network range that it is currently intercepting traffic for. The bridge IP on the appliance is used to initiate the connections to the back-end systems. This is why the Bridge IP address/netmask must be set appropriately so that it has access to all IP ranges. The Bridge IP address is also used for tasks such as DNS resolution, NTP syncs and sending email. Although it might be possible to use the Management Network for these tasks, experience shows that many Management networks do not have the access required. It is more practical to use the same network as the protected web apps, since they most likely already need to do NTP, send email and do DNS resolution.

Another point is that the default gateway specified in the CLI should always be specified for the bridge IP network and not for the management network. If you find that you have routing issues for the Management IP, then you can use the CLI to set a static route.


Last Modified 10/16/2013.
https://support.levelblue.com/kb/KnowledgebaseArticle14777.aspx