Yusheng Wang <yshw...@vmware.com> wrote on 06/19/2016 11:12:49 PM:
> From: Yusheng Wang <yshw...@vmware.com> > To: Hui Kang/Watson/IBM@IBMUS > Cc: "dev@openvswitch.org" <dev@openvswitch.org> > Date: 06/19/2016 11:13 PM > Subject: Re: [ovs-dev] [PATCH] OVN: initial patch of datalog engine > > > > The paper explicitly says the nlog implements the incremental computation > > for any state change in the input table, "incremental computation allows us > > to recompute only the affected state and push the delta down to the network > > edge." Therefore, IMO your implementation is applicable to either OVN or > > the NVP [2]. Please correct me if I am wrong. > > You are correct. They are going to have similar behavior with > respect to computation > method. What I meant previously is that we could not compare the > performance just > based on the computation engine. Other factors, such as how database > interacts with > the engine and how the computation results are dispensed to hosts, > will also have impact. Thanks for the explanation and I agree with your above comments. In current OVN implementation, the ovn-controller takes the responsibility of generating local openflow rules in each hypervisor, translating from the logical flows in the southbound database. If all the work is performed at the nlog engine and pushed to other hypervisors, I suspect the nlog node could become the bottleneck. I am very interested in this work and please let me know if I could help in any case. Thanks. - Hui > > > Some error message is as follows: > > Applying: OVN: datalog man page > > /root/Downloads/ovs/.git/rebase-apply/patch:34: trailing whitespace. > > Thanks for pointing out this. I will remove those white spaces in later patch. > > -- Yusheng > _______________________________________________ dev mailing list dev@openvswitch.org http://openvswitch.org/mailman/listinfo/dev
