On Wed, Mar 21, 2001 at 11:34:46PM +0800, Ian C. Sison wrote:
> But won't this increase overall latency - testing for congestion and
> making it the basis of a routing decision?
I haven't seen it yet, but still it is the best method I've seen for
inbound load balancing where BGP cannot be used... the only other way I've
thought of doing this is having two squids bound to different IP's,
and a third one which the users point to, which selects from one of the two.
> A lot of sites block ICMP
> nowadays, and it may actually fool such devices into thinking that the
> site is down... And doing a tcp port 80 probe ? Eh di kunin mo na lang
> yung buong page, nag connect ka na rin lang di ba?
A port 80 is just like a ping, open then close right away.
> Do you really think that the extra features of this appliance is worth
> the cost?
I don't know much it costs, so I don't know yet...
> So far, the only net appliance i think is worth the megabucks are the
> foundry L4 switches, and even that can be duplicated in functionality on a
> lesser scale by the linux virtual server.
L4 switches are more of switches+some intelligence; I would consider something
like the F5 "BigIP" as an appliance, in fact it is allegedly BSD based.
> Point being, unless your requirements are "wire-speed" networking
> functionality with special feature requirements, you don't need this, or
> a cisco, or a foundry, and definitely not one of those 'cache engines'..
The cache engines do have an advantage - not having the Unix file system.
squid is working on this aspect, though.
--
http://www.internet.org.ph The Philippine Internet Resource
Mobile Voice/Messaging: +63-917-810-9728
_
Philippine Linux Users Group. Web site and archives at http://plug.linux.org.ph
To leave: send "unsubscribe" in the body to [EMAIL PROTECTED]
To subscribe to the Linux Newbies' List: send "subscribe" in the body to
[EMAIL PROTECTED]
