a question on heirarchy

hello SQUIDers,

let this be the scenario:
  
A B C = server running SQUID.

                   +---+ +---+
                   | A |<-alias:proxy->| B |
                   +---+ +---+
                        \\ //
                         \\ //
                          \\ //
                           \\ +---+//
                              | C |
                              +---+
                                ||
                            +-------+
                            |browser|
                            +-------+

i have a server named "A" and another named "B". they have the same
alias/nickname which is "proxy". SQUID is executing in both server. i had
announced to my clients "C" to use "proxy" as their parent in their squid.conf:

     cache_host proxy parent 8080 3120
     acl browser src <IP_blah>/255.255.255.255
     http_access allow browser

AND this is what i have specified in A's squid.conf:

     cache_host B sibling 8080 3130
     acl C src <IP_blah>/255.255.255.255
     icp_access allow C

and in B's squid.conf i had this:

     cache_host A sibling 8080 3130
     acl C src <IP_blah>/255.255.255.255
     icp_access allow C

because i want A and B on the same level in the local heirarchy .

in that way, once a client has queried our proxy server named "proxy" in their
squid.conf, it would go to either A or B because they have the same
nickname/alias, is this right?

and when it is not found on one server (e.g. A) it would query it's sibling
which is B as specified by B squid.conf (or vice versa), is this right?

is this "flow" possible or would it create any problem/confusion?

is this advisable or should i rename the alias/nickname of the other server?
(e.g. A=proxy1; B=proxy2)

is the current configuration of A and B sufficient that they would be able to
communicate with each other (e.g. query)?

"browser" sent a query(TCP) to C but the object was not found(TCP_MISS), then it
queries(ICP_QUERY) "proxy" (either A or b) for the object. what if the queried
server was A and it was not found, so it queries B, and assuming the object was
found, would the object be cached in A and/or in C and delivered to "browser" or
other???

what if the object still not yet found(UDP_MISS) in both A and B, it would fetch
the object from the source. where would it cache the object?

Any *more* info?

i'd appreciate it. THANKS!

Leo David

---------------------------------------------------------
Get Your *Web-Based* Free Email at http://www.hotmail.com
---------------------------------------------------------
Received on Tue Apr 15 1997 - 10:28:24 MDT