Sadly, our hosting situation currently has 7 points of failure (devices, not counting cables which are also single failure points). 3 of the single points of failure are before the cable reaches our rack. 2 of our failure points are switches, the other 2 are our firewall and load balancer. --Dan
On Fri, 11 Jul 2003 12:35:42 -0600, "Soren Harward" <[EMAIL PROTECTED]> said: > On Fri 11 Jul 2003 at 11:53:36, Travis Stevenson said: > > Out of curiosity how would do the tunnel machine. Would it be a proxy? How > > would you tell the difference in which host to go to? > > You can do it a number of different ways. You can use Apache's > mod_proxy, set up in reverse, to serve from redundant machines. You can > use a box with IP forwarding rules (via iptables on Linux or pf on > OpenBSD). You could do a reverse squid proxy. You can use a > specialized load-balancer like Foundry's ServerIron or Cisco's CSS. The > disadvantage of these are that they're still single points of failure; > if you lose the load balancer, then everything goes down. Round-robin > DNS (pointing one hostname to multiple IP's on multiple machines) is > much more robust, because if you lose one machine, the others will take > over. > > -- > Soren Harward > [EMAIL PROTECTED] > > ____________________ > BYU Unix Users Group > http://uug.byu.edu/ > ___________________________________________________________________ > List Info: http://uug.byu.edu/cgi-bin/mailman/listinfo/uug-list > ____________________ BYU Unix Users Group http://uug.byu.edu/ ___________________________________________________________________ List Info: http://uug.byu.edu/cgi-bin/mailman/listinfo/uug-list
