4 ms·

Wow, this article is extremely disingenuous with the notion that P4 is the first vendor-agnostic way to control forwarding behavior in switches. OpenFlow has b

Wow, this article is extremely disingenuous with the notion that P4 is the first vendor-agnostic way to control forwarding behavior in switches. OpenFlow has been around for years now and is a vendor agnostic way to control forwarding behavior and it allows packet injection as well. I get that a startup has to pretend this is a huge foundational advancement, but P4 is really an incremental improvement to the state of SDN that openflow brought. Disappointing that they don't even acknowledge it once.

insaneirish 9y ago
OpenFlow and P4 address different levels of abstraction in switching. OpenFlow (in its original form, for simplicity's sake) says "Hey, we have a match/action table with these 12 fields. Here's a way to put stuff in those fields and the switch will match and act accordingly." P4 says "Hey, we have this piece of programmable silicon. Feel free to define the sorts of tables you want, the protocols you want to switch on, and we'll then give you tables and an API to program." You can write an OpenFlow pipeline in P4. Also, maybe take a look at the original OpenFlow paper [1] and the original P4 paper [2]. Nick McKeown, one of Barefoot's cofounders, is an author on both. [1]: http://ccr.sigcomm.org/online/files/p69-v38n2n-mckeown.pdf http://ccr.sigcomm.org/online/files/p69-v38n2n-mckeown.pdf [2]: http://yuba.stanford.edu/~nickm/papers/v44n3k2-bosshartA.pdf http://yuba.stanford.edu/~nickm/papers/v44n3k2-bosshartA.pdf
- johnvega 9y ago
  Would you consider this as "huge foundational advancement"?
- hueving 9y ago
  Yes, different levels of abstraction but P4 is not a foundational beginning of SDN like this post makes it out to be. Packet injection, etc were already possible with OpenFlow and this article pretends P4 is the first thing to enable that. P4 can certainly allow much more than openflow, but some of the use cases this article pretends like P4 is the first to solve were already solved by openflow, etc.

insaneirish 9y ago

OpenFlow and P4 address different levels of abstraction in switching. OpenFlow (in its original form, for simplicity's sake) says "Hey, we have a match/action table with these 12 fields. Here's a way to put stuff in those fields and the switch will match and act accordingly." P4 says "Hey, we have this piece of programmable silicon. Feel free to define the sorts of tables you want, the protocols you want to switch on, and we'll then give you tables and an API to program." You can write an OpenFlow pipeline in P4. Also, maybe take a look at the original OpenFlow paper [1] and the original P4 paper [2]. Nick McKeown, one of Barefoot's cofounders, is an author on both. [1]: http://ccr.sigcomm.org/online/files/p69-v38n2n-mckeown.pdf http://ccr.sigcomm.org/online/files/p69-v38n2n-mckeown.pdf [2]: http://yuba.stanford.edu/~nickm/papers/v44n3k2-bosshartA.pdf http://yuba.stanford.edu/~nickm/papers/v44n3k2-bosshartA.pdf

johnvega 9y ago
Would you consider this as "huge foundational advancement"?
hueving 9y ago
Yes, different levels of abstraction but P4 is not a foundational beginning of SDN like this post makes it out to be. Packet injection, etc were already possible with OpenFlow and this article pretends P4 is the first thing to enable that. P4 can certainly allow much more than openflow, but some of the use cases this article pretends like P4 is the first to solve were already solved by openflow, etc.