IPFIX Sampling via Firewall Filter vs. IPFIX Sampling via Interface - Is there any difference?

Hello,

I have configured IPFIX on my MX960 via this guide:

https://www.juniper.net/techpubs/en_US/junos15.1/topics/task/configuration/services-ipfix-flow-template-flow-aggregation-configuring.html

And I have a question:

What is the difference between "applying" sampling to an interface and sampling via a firewall filter and then applying that filter to an interface?

Example:

set interfaces ae0 unit 0 family inet sampling input
set interfaces ae0 unit 0 family inet sampling output

vs.

set firewall family inet filter SAMPLE-ALL term 1 then sample
set firewall family inet filter SAMPLE-ALL term 1 then accept
set interfaces ae0 unit 0 family inet filter input SAMPLE-ALL
set interfaces ae0 unit 0 family inet filter output SAMPLE-ALL

Is there any difference between those two configurations?

The offical docs state the following:

https://www.juniper.net/techpubs/en_US/junos14.2/topics/usage-guidelines/services-configuring-traffic-sampling.html

• On the Routing Engine, using the sampled process. To select this method, use a filter (input or output) with a matching term that contains the then sample statement.
• On the Monitoring Services, Adaptive Services, or Multiservices PIC.
• On an inline data path without the need for a services Dense Port Concentrator (DPC). To do this inline active sampling, you define a sampling instance with specific properties. One Flexible PIC Concentrator (FPC) can support only one instance; for each instance, either services PIC-based sampling or inline sampling is supported per family. Inline sampling supports version 9 and IPFIX flow collection templates.

However, this explanation does not make any sense to me since in both cases ("firewall filter" and "family inet sampling") the sampled process on my routing engine seems to be very active.