Hi All This mail is in context of discussion we had in controller hackers (weekly) meeting this week..
We have observed that there are two kind of datastore usage pattern in ODL 1. Low-rate events that result in lot of interactions among ODL modules. This could be for eg. a switch port-up event. 2. High-rate events with low interaction / almost pass-thru, that just installs state in switch. This could be for eg. installing subscriber filters for SFC classifier. We are working towards creating a reference app that allows support for the high-rate event pattern in an efficient manner. The plan is to use an existing opensource horizontally scalable datastore (such as Cassandra) and integrate it with ODL infrastructure. The integration goes beyond the software / libraries / jar integration. We intend to integrate the reference app in a manner so that restart scenarios such as DPN restart, ODL node restart and cluster restart, are properly handled. The restart scenario involves reconciliation of data/information between Controller and switch after the restart event. In this regards, we wanted to solicit feedback from community on this approach ( a lot of folks are not present in the hackers meeting :) ). In particular we wanted to know if such efforts been taken earlier.. and if so, is it possible to share the learnings.. With regards Ashutosh
_______________________________________________ controller-dev mailing list controller-dev@lists.opendaylight.org https://lists.opendaylight.org/mailman/listinfo/controller-dev
