Looking back on the 2022 P4 Workshop I had two big takeaways: the first narrowly focuses on the proposed roadmap for P4, while the second looks at the broader context in which P4 is evolving. Both are connected to many of the themes we write about in these posts. The narrow takeaway is that P4 is now securely established as the language of choice for programming packet forwarding pipelines.
Larry - P4 has been around so long and I have seen (and asked for) the need to support stateful processing. why has that taken so long? at the same time, supporting L2/L3 switching along the lines of table-based OpenFlow is not very interesting IMO
Raj -- There are proposals to support flow state, and I don't know of any reason why a thoughtful reduction of those ideas wouldn't work, but it takes energy to make it a reality. My sense is that P4 is just now achieving the critical mass needed to make the next jump. But up to this point, the todo list has outpaced the available community resources.
Larry - P4 has been around so long and I have seen (and asked for) the need to support stateful processing. why has that taken so long? at the same time, supporting L2/L3 switching along the lines of table-based OpenFlow is not very interesting IMO
Raj -- There are proposals to support flow state, and I don't know of any reason why a thoughtful reduction of those ideas wouldn't work, but it takes energy to make it a reality. My sense is that P4 is just now achieving the critical mass needed to make the next jump. But up to this point, the todo list has outpaced the available community resources.
I'd rather like some form of FQ and AQM to make it cleanly into p4. https://github.com/ralfkundel/p4-codel/issues/2