Could this be accomplished with 7.6 Event Policy
configuration?
http
://www.juniper.net/techpubs/software/junos/junos76/swconfig7
6-automation/html/event-policy-overview.html
-----Original Message-----
From: juniper-nsp-bounces puck.nether.net
[mailto:juniper-nsp-bouncespuck.nether.net] On Behalf
Of Ebben Aries
Sent: Wednesday, June 14, 2006 1:19 PM
To: juniper-nsppuck.nether.net
Subject: Re: [j-nsp] Best way to use transparent caching
with M/E-series
RPM is great for monitoring performance and pulling via SNMP
or NETCONF but I don't believe you can have the rpm trigger
an action based on a threshold which is what you are
attempting to accomplish.
--
ebb
-----Original Message-----
From: juniper-nsp-bouncespuck.nether.net
[mailto:juniper-nsp-bouncespuck.nether.net] On Behalf
Of Alex
Sent: Wednesday, June 14, 2006 10:14 AM
To: Guy Davies; Dean Albano
Cc: juniper-nsppuck.nether.net
Subject: Re: [j-nsp] Best way to use transparent caching
with M/E-series
RPM (Real Time Performance Monitoring) is able to send HTTP
probes (I think
these are GETs).
HTH
Cheers
Alex
----- Original Message -----
From: "Guy Davies" <aguydaviesgmail.com>
To: "Dean Albano" <dean.albano121media.com>
Cc: <juniper-nsppuck.nether.net>
Sent: Wednesday, June 14, 2006 2:26 PM
Subject: Re: [j-nsp] Best way to use transparent caching
with M/E-series
Hi Dean,
There's no "http checker" or the like. I
suppose if the cache device
is *directly connected* to the router and the interface goes
down,
then you could have a second, less preferred default route
in the
cache routing-instance that points to the same place as the
default
route in the inet.0 routing table. But that would rely on
the
interface on the router recognising that it was down so the
router no
longer had a path to 192.168.2.2 (in my example). That
would make
that static route unusable and the less preferred static
would be
used.
I haven't tested that so treat with due scepticism 
Rgds,
Guy
On 14/06/06, Dean Albano <dean.albano121media.com> wrote:
> Thanks Guy. We will give this a try. One question
though; what
> happens if the cache host goes down? Is there any
mechanism with
> Filter Based Forwarding to identify that the cache host
is down and
> route http traffic around it?
>
> Regards,
>
> Dean J. Albano
> Network Integration Consultant
>
>
> On Jun 14, 2006, at 8:37 AM, Guy Davies wrote:
>
> Hi Dean,
>
> You can do something similar on the M/T series by doing
filter based
> forwarding.
>
> You need a minimum of three interfaces.
>
> 1. The ingress interface from which http/https requests
will be
> received.
> 2. The egress interface to the Internet
> 3. The egress interface to the Caching device
>
> You identify ingress traffic that is http or https and
then apply a
> non-standard routing-instance (with a default pointing
to the Caching
> device) to that traffic. The traffic is then forwarded
to the caching
> device which is able to either respond direct or has to
make a query
> to the Internet. It is important to apply the filter
to the input of
> the ingress interface rather than the output of the
egress to the
> Internet or you'll catch the traffic from the caching
device too and
> send it back to the caching device.
>
> interfaces {
> ge-0/0/0 {
> description "Ingress";
> family inet {
> address 192.168.0.1/24;
> filter {
> input transparent-cache;
> }
> }
> }
> ge-0/1/0 {
> description "Egress";
> family inet {
> address 192.168.1.1/24;
> }
> }
> ge-0/2/0 {
> description "Cache";
> family inet {
> address 192.168.2.1/24;
> }
> }
> }
> firewall {
> family inet {
> filter transparent-cache {
> term http-https {
> from {
> protocol tcp;
> destination-port [ 80 443 ];
> }
> then {
> routing-instance cache;
> }
> }
> term default {
> then accept;
> }
> }
> }
> }
> routing-instances {
> cache {
> instance-type forwarding;
> routing-options {
> static {
> route 0.0.0.0/0 next-hop 192.168.2.2;
> }
> }
> }
> }
> routing-options {
> interface-routes {
> rib-group inet direct-if;
> }
> rib-groups {
> direct-if {
> import-rib [ inet.0 cache.inet.0 ];
> }
> }
> }
>
> Rgds,
>
> Guy
>
> On 14/06/06, Bjørn Mork <bjornmork.no> wrote:
> > Dean Albano <dean.albano121media.com> writes:
> >
> > > What would be the best way to do transparent
caching through either
> > > an M-series or E-series router?
> >
> > Put a L4-switch at either side of it?
> >
> >
> > Bjørn
> >
> > _______________________________________________
> > juniper-nsp mailing list juniper-nsppuck.nether.net
> > h
ttp://puck.nether.net/mailman/listinfo/juniper-nsp
> >
>
>
_______________________________________________
juniper-nsp mailing list juniper-nsppuck.nether.net
h
ttp://puck.nether.net/mailman/listinfo/juniper-nsp
_______________________________________________
juniper-nsp mailing list juniper-nsppuck.nether.net
h
ttp://puck.nether.net/mailman/listinfo/juniper-nsp
_______________________________________________
juniper-nsp mailing list juniper-nsppuck.nether.net
h
ttp://puck.nether.net/mailman/listinfo/juniper-nsp
_______________________________________________
juniper-nsp mailing list juniper-nsppuck.nether.net
h
ttp://puck.nether.net/mailman/listinfo/juniper-nsp
|