Hi Jan, On Thu, May 22, 2014 at 12:19 PM, Jan Medved (jmedved) <[email protected]> wrote: > Hi Jamal, > > [..] > Jeff¹s email states: "Part of this consideration includes elements of > expediency such as existing open source tool chains and implementation > experience of I2RS-like mechanisms in other organizations and vendors.² >
> I don¹t think that Jeff is saying that there exists an open source I2RS > implementation - Agreed. My argument is that the only time you can make a claim about virtues of open source is if the implementation in fact exists. >he is saying that there is and *I2RS-like* mechanism in > open source (I think we all agree that NC/Y qualifies as *I2RS-like*). It does qualify like you say as *I2RS-like* (as does ForCES arch) - unfortunately since requirements are not clear, i am not sure how far we can stretch the *like* part. >The > point is that these open source NC/Y implementations (there are multiple) > will provide developers and protocols designers with a platform (and a > starting point) that will allow them to extend the existing functionality > towards I2RS, and to quickly prototype new protocol functionality that > will be defined in the I2RS WG. > I disagree. I dont think netconf/restconf/yang without acrobatics will resolve the requirements; not saying it couldnt be made to with a lot of effort and refactored. I dont consider the presence of open source to be a reasonable input unless the majority use the same code (reminds me of OF "interop" where 95% of the vendors have exactly the same implementation of OVS). cheers, jamal > > > Thanks, > Jan > >> > _______________________________________________ i2rs mailing list [email protected] https://www.ietf.org/mailman/listinfo/i2rs
