when it comes to addressing SP move to CLOS more important than RIFT (which is more of a nice to have if you want to get your CLOS up easily and yes, RIFT does a much better flood reduction _and_ balancing than anything else on the table if you dig into it) is flood reflection which allows you to scale out your network when the IGP maxes out _without_ the problem of creating a single point of failure in the "abstracted CLOS" which will lead over time to all the "let's not use chassis" problems like NSR again. And flood reflection allows also TE within the abstracted CLOS without a massive refactoring of e.g. RSVP-TE. May not be the more popular thing these days to talk about RSVP-TE but the amount of deployment of that and business riding on it is massive.
Thanks Tony. As IP/MPLS is already the incumbent for many existing networks/domains, 5G is perhaps the most interesting opportunity for SR beyond the hyperscalers. Not just the 5G ecosystem has wrapped its arms around SR, but also, because, for reasons you have implied, it is easier to do something new, when you are doing something new. 5G/6G are likely opportunities to consider any number of new approaches, whether it be virtualized packet core or SR. We shall see...
when it comes to addressing SP move to CLOS more important than RIFT (which is more of a nice to have if you want to get your CLOS up easily and yes, RIFT does a much better flood reduction _and_ balancing than anything else on the table if you dig into it) is flood reflection which allows you to scale out your network when the IGP maxes out _without_ the problem of creating a single point of failure in the "abstracted CLOS" which will lead over time to all the "let's not use chassis" problems like NSR again. And flood reflection allows also TE within the abstracted CLOS without a massive refactoring of e.g. RSVP-TE. May not be the more popular thing these days to talk about RSVP-TE but the amount of deployment of that and business riding on it is massive.
Thanks Tony. As IP/MPLS is already the incumbent for many existing networks/domains, 5G is perhaps the most interesting opportunity for SR beyond the hyperscalers. Not just the 5G ecosystem has wrapped its arms around SR, but also, because, for reasons you have implied, it is easier to do something new, when you are doing something new. 5G/6G are likely opportunities to consider any number of new approaches, whether it be virtualized packet core or SR. We shall see...