From: Pranav Kantawala (p_kantawalaIZZATyahoo.com)
Date: Thu Jan 30 2003 - 14:09:03 EST
I'm trying to find out the recommended solution to
load balance Nodes on any subnet, without BIGIP
having an IP interface on that subnet. I'm using
'BIG-IP Kernel 4.2'.
For instance, I have the BIGIP on subnet S1, and Nodes
A and B on subnet S2 and S3 respectively.
Traffic from Nodes A and B to other nodes in the
network may bypass the BIGIP. I would like to
define a Virtual Server VS1 to load balance a service
on Nodes A and B.
I was able to achieve this using SNAT configuration as
1. I defined only one VLAN (external) on the BigIP.
2. I defined a default SNAT with a publicly accessible
IP address as the translation address.
3. I defined a pool, lb_pool with two nodes A and B.
4. I defined a Virtual Server VS1 that uses pool
The client is on a completely different subnet. When
client makes a connection to VS1, the BIGIP
translates the Source IP to SNAT address, and
Destination IP to Node Address (either A or B).
Thus, the Nodes always see the source as the SNAT
I had to define SNAT, because without it, the Virtual
Server definition instructs the BIGIP to
translate only the Destination IP address, and not the
Source IP address in packets. In doing so,
the return traffic from nodes back to the client could
bypass the BIGIP resulting in the client
dropping the connection.
Are there any limitations imposed by the Source IP
translation with SNAT? Is persistence
configuration affected by this (for example, Client IP
based persistence)? Are there any other
restrictions when using SNAT to load balance Nodes on
different subnets? If this is not the
recommended solution, could you please suggest an
Do you Yahoo!?
Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
The Load Balancing Mailing List
MRTG with SLB: http://vegan.net/MRTG
Hosted by: http://www.tokkisystems.com
This archive was generated by hypermail 2.1.4 : Thu Jan 30 2003 - 14:23:08 EST