At MPLS World Congress 2013 and SDN Summit 2013 in Paris, the talk is less about MPLS and more about SDN. Behind the scenes, the talk is all about the world of open source controllers and the ongoing role of the ONF.
Since the publication of the consumer-driven Network Function Virtualization whitepaper in the Fall, there has been much discussion of the separation of planes (control, data, service and management) in currently integrated network platforms. Service plane separation will require the definition/derivation of active service-specific topologies which are very intertwined with and possibly advised by attempts to separate the control and data plane.
In spite of it's promises, the OpenFlow protocol hasn't brought about a universally adopted, standardized and open control/data plane separation. That may be due to the level of abstraction provided. Attempts currently percolating to in the IETF to provide network programmability will attempt a different level of abstraction (and potential separation of the control/data planes).
In the interim, NFV will probably rely on what seems to be working today ...on-ramps and overlays (which may not be pure control/data plane separation).
With a year or more under it's belt, the SDN wave has passed from Japan to the US and finally washed up in Europe. Two recent conferences in Darmstadt, Germany offer the opportunity to talk about standardization and SDN (against a backdrop of the ONF starting an Architecture Work Group).
It may not be heresy to participate in the standardization of OpenFlow when you work for Juniper Networks (after all, it is a network protocol and Juniper Networks is all about network protocols). Nor may it be heresy to relate just how popular this new protocol and the seemingly-related topic of SDN were at InterOp Tokyo (though it is worth noting). Perhaps the true heresy will come when we start to create hybrids of the OpenFlow and IETF protocol-based network solutions through cross-pollination.