The Life of a Packet
When you’ve decided which services you need in your network, implementing them can seem like a daunting task. Which configuration do you apply first? Which addresses do you apply to a given service when applying a rule? How about if you have a class of service (CoS) applied? Where do packet filters fit into the packet flow?
Figure 8-13 and Figure 8-14 should help you sort through this mystery and decide where configuration processes occur. These diagrams are not meant to represent every possible scenario, but rather to give you a general feel for the complex processing that could be involved. The exact processing will depend on whether a next hop- or an interface-style service set is applied (for more on CoS concepts, refer to Chapter 9).
Figure 8-13. Incoming packet processing
The steps outlined in Figure 8-13 are as follows:
The packet enters the incoming interface.
The packet is classified by a behavior aggregate (BA) classifier.
The packet is processed by an input filter and policer (and may be reclassified).
The packet enters a service filter, and it is either serviced or skipped; if it is skipped, jump to step 7.
The input service is performed.
A post-service filter is applied.
The route lookup occurs.
If the result of the route lookup is a GRE packet, go to step 9; if not, send on for output processing.
De-encapsulate the GRE packet and go back to step 1.
After the ...
Get JUNOS Enterprise Routing now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
