OverviewOutbound on one Exinda is the inbound for the other Exinda. So, it is necessary to control "Outbound" traffic as the best practice.
It is necessary to control inbound traffic in the following scenarios
1> For a different set of policies for "Inbound and outbound" on an Exinda
2> When the outbound throughput of one Exinda is not the same as "Inbound" of the other Exinda
When Virtual Circuits are different sizes on the core Exinda and the remote Exinda, the following is recommended: for the side with the larger VC communicating to the smaller one, setup only Outbound. On the smaller VC, setup Inbound/Outbound, or Both. The likely cause of this scenario would be the occurrence of a single core VC connecting to multiple remote VCs/devices. Here is an example of why this is a good idea:
Core Exinda VC to Remote5 is 10Mbps
Remote5 Exinda VC is 2.5Mbps
Outbound Web Traffic Guaranteed Bandwidth on Core Exinda VC is 30% (that is 30% of 10Mbps)
In this situation, the core VC could potentially send up to 3Mbps of Web traffic at once to the remote VC. If there is no policy on the on the remote VC limiting Inbound traffic, it could become saturated with incoming.
For instances where the Virtual Circuit is the same size on the core Exinda as it is on the remote Exinda, use only Outbound. This will be fine for accelerating/non-accelerating. A benefit of using only Outbound VCs would be that it will remove any unnecessary memory/CPU usage from having extra VCs for Inbound traffic. That may not be a big concern for the remote devices, but it may help with the core device's performance depending on the number of VCs/policies.
To have less VCs/policies on the core Exinda, use only Outbound on it.
With a remote VC which is the same size as the core VC, use only Outbound on policies on both the Exindas.
With a remote VC which is a smaller size than the core VC, use Inbound/Outbound, or Both on the remote end. The core end needs outbound for sure.
Please note: If there is not a VC or policy configured for "Inbound" traffic, the traffic will fall in the "Auto Catch-all" policy of the Exinda. This will be seen in the Real-time monitor, control and virtual circuit graphs as "Auto Catch-all" Virtual Circuit and "Auto Catch-all" policy.
If monitoring of Inbound traffic based on the VC/policy is critical, then configure relative VC's and policies for it.
Presence or absence of a VC will not affect any user based or application based monitoring.