Overview

Usually, we use the Exinda's Edge Cache feature to peer with a third party upstream proxy to check the cache first before processing on the internet.

But what if we want that upstream proxy to be the Edge Cache (EC) of a Central Exinda? This article guides you on that procedure.

Process

NOTE: As of v7.4.4, there were no CLI commands to perform this in the Acceleration Edge-Cache CLI options.

On the Upstream Cache Exinda

1. Log in to the appliance and drop to the shell: _shell
2. Change the filesystem permissions to read-write: [admin@ex-240 ~]# remountrw
3. Open the squid.conf file: [admin@ex-240 ~]# vi /opt/tms/lib/md/templates/squid.conf
4. Change the http_port line as shown below:

Before:

# Bound ports
http_port 127.0.0.1:3128

After:

# Bound ports
http_port 3128

5. Save the file and restart EC.

On the Downstream Cache Exinda

1. Add the HO Exinda as a peer with HTTP port as 3128 and ICP port as 0.
2. Restart Edge Cache Service.
   
   

Confirmation 

1. Open a YouTube video.
2. Check the /var/log/squid/access.log file.
   It shows the log as querying the parent first as shown below:
   

   *1351115446.305 47 192.168.50.17 TCP_MISS/200 278401 
   GET http://s.ytimg.com/yts/swfbin/watch_as3-vflSguSsi.swf - FIRST_UP_PARENT/172.16.1.240 application/x-shockwave-flash*

NOTE:

The Exinda is now monitoring one HTTP conversation, as it was two different conversations:

• The one from the client IP to the Server IP.
• And the one to/from Exindas (Proxy'd traffic).

This generates incorrect monitoring due to duplicated HTTP flows.

When creating network objects for both Exindas:

• Declare them as internal; making sure internal-to-internal traffic is ignored solves the issue.