Dear Antonio and Marco, > > In the attached .pdf, page 6, there are remarks about the "size of DNS > > replies". > > It points out that the authority and additional records mean higher > > overload. > > In that text you add a footnote (3) about "avoiding the inclusion of > > these sections". > > The footnote states, I quote : > > "It can be removed with dig using the options +noauthority +noadditional" > > > > I hope you're not too serious about this ! > > From dig help : > > +[no]answer (Control display of answer) > > +[no]authority (Control display of authority) > > +[no]additional (Control display of additional) > > → these options control the *display* of the sections, > > not if a name server should include them in the reply ! > > > > Yes, you are true, it is only for display issues. We were testing at the beginning > only as a proof of concept the capabilities of reducing the DNS overload reducing > the number of entries. > > Right now, we are in progress of our own implementation on Contiki in order to > select by ourself the replies to be included. As soon as we complete the Contiki > version and we evaluate it on the nodes, we will evaluate it over directly. > > Anyway, the dig options are interesting, since although it is true that it is only > removing in the display, the analysis of the overload presented is regarding to the > displayed authorities and answers. > Therefore, it continues being useful :-). > > Otherwise said : the (DNS) client has *no control* about what the name > > server will include in its answers. > > That control has to be exercised on the name server itself, if at all > > possible. > > (for Bind, look at the option : minimal-responses) > > > > Yes, our idea is similar to offer an implementation of minimal-responses for smart > things, and right now the analysis is focused on what is a minimal response for a > constrained node, i.e. > only the PTR record, only the TXT of the interested attribute...
Usually we deploy this option on recursive server in the internet, because the information is useless for end user. So it can reduce payload to client. I don’t know whether it could be supported on the client. Usually not do it on client in the internet, I think. As I know a lot of recursive servers use this option, such as Google's. One possible option is that we just use the recursive server with dig :-). But it can't solve the problem at all, research on minimal response is still need. Best regards, Yan _______________________________________________ Lwip mailing list [email protected] https://www.ietf.org/mailman/listinfo/lwip
