Just put two addresses on every server on the inside. Nat the outside
address to the inside address for that particular connection. Use PBR to
make sure return traffic uses the correct interface. For redundancy purposes
put a DNS server behind each connection, and only advertise services
available via that connection out each. Use a low DNS TTL for services that
are typically persistent. Mail the MX records will handle this naturally,
DNS will handle this naturally, a web server will need the DNS cache on the
client side to time out. For egress traffic from your own clients, use GLBP
to load balance the two connections.
David
-- http://dcp.dcptech.com -----Original Message----- From: nobody_at_groupstudy.com [mailto:nobody_at_groupstudy.com] On Behalf Of Alexander Lim Sent: Friday, September 28, 2012 10:29 PM To: Umesh Narayanan Cc: ccielab_at_groupstudy.com Subject: Re: Redundancy with two ISPs by configuring BGP Use Link Load Balancer: F5 or Radware. Regards, Alexander Lim On 28 Sep, 2012, at 7:29 PM, Umesh Narayanan <umesh.narayanan_at_gmail.com> wrote: > Hi Guys, > > I have two public IPs (two different ISPs) and i am planning to > configure redundancy between these IPs for my network. Here i can configure > HSRP or any redundany protocol but the challenge is for the mail or > webserver where the incoming traffic towards them will face issues as the > public IP will change. > > BGP will be an idea solution, my idea is to configure BGP multi homing > with one set of public IPs from one ISP to be peered with the other ISP > with private AS at our premise. Do let me if this will be an idea solution > or pls suggest any other solutions. > > Regards, > Umesh > > > Blogs and organic groups at http://www.ccie.net > > _______________________________________________________________________ > Subscription information may be found at: > http://www.groupstudy.com/list/CCIELab.html Blogs and organic groups at http://www.ccie.netReceived on Fri Sep 28 2012 - 22:42:55 ART
This archive was generated by hypermail 2.2.0 : Mon Oct 01 2012 - 06:40:29 ART