VIDEO: Dan Pitt Exec Chair, Open Networking Foundation #Interop

i'm looking at uh all these nice boxes and switches and vendor signs and when you look at this dan what does this mean to you when you see all these names on all these boxes it says to me that these companies are actually investing in open flow and it's nearing more and more commercial reality great now when you onf started a year ago would we have seen a demo with this many different vendor names all over it probably not i would say so there was an open flow lab at interop a year ago sure i don't remember everything it was built in four years any um and there were a number of vendors they weren't really interoperating most of these are products product level now yeah and they're inter-operating so we have more more switch vendors we have more controllers and more applications yeah last year i remember a lot of switches and i remember the nec programmable flow controller on hardware and that was it if i'm looking at actual physical hardware switches the hardware controllers who are the hardware controllers now well is it still essentially just nec and then there's big switch yeah the first time i actually got it that's it uh i'm not quite sure how to classify liatus yet they're a new company from france so it still looks to me like the uh the openflow controller plane is an area where there may or may not be room for more vendor participation or is that an area where there may be risk because i i see some vendors whether it's hp or otherwise who are now baking that into their own proprietary solutions i think the controller space the software space in general above open flow is really right for innovation and for a lot of market entrance because we don't even know how to define the scope or limit the scope we have a controller it's controlled software yeah uh we talked about the so-called northbound api above the controller we don't even know what latitude that is we're talking about northbound we don't know what's contained in a controller what are apps or services on top of the controller for example look at the uh some of the traffic engineering you'll see here look at some of the load balancing look at the topology discovery is that part of the controller or is an app that runs on top of the controller look at floodlight around the corner yeah that's kind of a bare bones controller from big switch which makes a commercial controller so it's an it's a space that is still having a great deal of exciting experimentation sounds good when you look at and i know we talked about this not that long ago when you look at open flow you just talked about some of the potentials what do you see as the as the challenges or potential barriers to adoption of any right now well i think defenders are trying to figure out what to build and the enterprises are trying to figure out what to buy they know it's going to help them they haven't figured out exactly where yeah and so i think the biggest question is what comes out first so what does come out first everyone everyone is first saying okay our switches are going to be open flow compatible and then on top they're saying okay our controller will we'll do this or that and then where's the where's the middle ground so we're looking at seeing where the uh where the use cases are strongest so the large-scale data centers yeah um we're starting to see a lot of interest in carriers for enterprise services and enterprises themselves to really lower their operating expense yeah and so i think you're going to see a lot more automated configuration all the things that take too much human effort right now and cost money to cause errors sure and then the other thing that people are usually worried about whenever there's a new networking technology is uh rip and replace when you when you look at onf and what you have to do with software defined networking is there a mantra to try and prevent rip and replace or is rip and replace the only way forward for networks no this is actually one of the nicest radical changes in networking that it does not require rip and replace yeah if you look at the hp announcement from february they announced open flow support on 16 families of switches with a free software download stuff that's already in the field you don't even have to take that stuff out of the field you can deploy it one switch at a time you can have part of the network being open flow and part of it not and they're orthogonal and that allows you organic gradual growth of the capability with existing equipment or with a slow introduction of new equipment great now i know you recently had the open networking foundation or the conference the summit some things came out of that it was all kinds of interesting announcements when you look at the next set of conferences and summits that you have are there certain key milestones you're looking to achieve in terms of either consensus among vendors for a protocol or or some other sort of larger direction well i think we have consensus on the protocol and that's being standardized and we just issued a new version of that approved a new version of that i think what we'll see over time is more commercial announcements from both the vendors and the service providers yeah and we'll and we'll have more citations of installations and enterprises know that we can vote like nec has two here today for sure genesis hosting in chicago and midnight express in japan we'll have uh more enterprises that are willing to say hey i've done it here it is you can talk about me got it good uh and then on the uh risks of proprietary versus open i know open flow itself is is an open protocol there may or may not be some proprietary overlays that people try and do for control how do you prevent any part of the ons mission whether it's open flow or otherwise from becoming proprietary lock-in so you know we're all about open interfaces in an open marketplace and we're trying to make sure that the protocol itself has enough of the basic features that it satisfies a large part of the market and that the innovation takes place mostly in software and therefore is easily interchangeable we're also doing a lot to promote interoperability between the existing implementations so that gives customers confidence that they have some equipment they buy other equipment it'll work together and if they want to change one for another it'll still work with what they have so it's really about customer assurance the market will ultimately decide how much it wants to go proprietary and how much it wants to go standard it's mostly voted for standards

This transcript was generated automatically from the video's captions and may contain errors.