Skip to content

Commit

Permalink
Preliminary slides
Browse files Browse the repository at this point in the history
  • Loading branch information
Christian Amsüss committed Jul 24, 2024
1 parent ebad514 commit d1aeed5
Show file tree
Hide file tree
Showing 4 changed files with 132 additions and 0 deletions.
3 changes: 3 additions & 0 deletions .gitmodules
Original file line number Diff line number Diff line change
Expand Up @@ -4,3 +4,6 @@
[submodule "2024-03-20-ietf119/shared"]
path = presentations/2024-03-20-ietf119/shared
url = https://codeberg.org/chrysn/presentation-shared
[submodule "2024-07-24-ietf120/shared"]
path = presentations/2024-07-24-ietf120/shared
url = https://codeberg.org/chrysn/presentation-shared
1 change: 1 addition & 0 deletions presentations/2024-07-24-ietf120/Makefile
Original file line number Diff line number Diff line change
@@ -0,0 +1 @@
include shared/Makefile
1 change: 1 addition & 0 deletions presentations/2024-07-24-ietf120/shared
Submodule shared added at 2f063a
127 changes: 127 additions & 0 deletions presentations/2024-07-24-ietf120/slides.tex
Original file line number Diff line number Diff line change
@@ -0,0 +1,127 @@
\input{shared/ietf-slides.tex}

% about the presentation
\title[CoAP Transport Indication]{CoAP Transport Indication}
\subtitle{\ietfdraft{ietf-core-transport-indication-06}}
\author{\textit{Christian~Amsüss}, Martine~Lenders}
\date{2024-07-24 \\CoRE at IETF 120 in Vancouver}

\begin{document}

\frame{\titlepage}

\begin{frame}{Goals}\Large
\begin{enumerate}
\color{gray}
\item Enablement of transport discovery
\item No Aliasing
\item Optimization (no cost per request)
\item Proxy usability
\item Proxy announcement
\end{enumerate}

\bigskip

\begin{itemize}
\color{gray}
\item Give way forward after \texttt{coap://} and \texttt{coap+tcp} diverged
\end{itemize}
\end{frame}

\begin{frame}{Recent changes worth discussing}\Large
\begin{itemize}
\item Scope of has-proxy relation
\item A lot about SVCB
\end{itemize}
\end{frame}

\begin{frame}{Open question: Scope of \texttt{has-proxy}}\large
\color{gray}
``only through link relations''

\begin{itemize}
\color{gray}
\item URIs regarded as opaque
\item Relations are explicit
\item Allows to exclude indvidiual resources from transport indication
\item Uses RFC~6690 \texttt{rel=hosts} which is not very clear
\item Hard to keep track of what works where
\end{itemize}

\color{black}
\begin{center}
vs.
\end{center}

``Applies per Origin''

\begin{itemize}
\item HTTP's mechanism
\item Simple
\item Way shorter wording in terminology section
\end{itemize}
\end{frame}

\begin{frame}{Recent changes worth discussing}\Large
\begin{itemize}
\color{gray}
\item Scope of has-proxy relation
\color{black}
\item A lot about SVCB
\end{itemize}

\bigskip

\ldots without moving away from ``transports are proxies'' as a guiding principle
\end{frame}

\begin{frame}[fragile]{SVCB components}
% * DNR describes handling of SVCB data -- generalized
% * SVCB handling was mushy: mix of DNR actual-use-case, next-ip-based-should and what-if history
% * plot: flows
figer WIP / to be updated in iteration:

\begin{verbatim}
A / AAAA
_____ ___ _coap SVCB
/ \ \
v v
URI <---> full message <----> SVCB param set
1 (transport proto etc) (_coap or _dns,
_________________-> latter w/ path)
/ /
[ SVCB from DDR ] [ SVCBish from RA / DHCP ]
for future msgs can only go left with SVCB literals
\end{verbatim}

\end{frame}

\begin{frame}{SVCB records for name resolutions}\Large
\begin{itemize}
\item Not retroactively activated.
\item Applications can opt in.
\end{itemize}

\bigskip

Questions:

\begin{itemize}
\item Is the above too cautious?
\item \texttt{\_coap SVCB} or \texttt{COAP} RR?
% main reason for HTTPS are catch-all zones
\item Extra \texttt{\_coaps SVCB}?
% it's not like it has as precise semantics as HTTPS in practice
\end{itemize}
\end{frame}

\begin{frame}{Next steps}\Large
\begin{itemize}
\item Who would review this?
\item Follow SVCB-parameters literals\footnote{Necessary if we want the next IP based CoAP transport to work without a new scheme, or are we happy if that can't use literals}?
\item Who would implement enough of this to interop test?\footnote{Document is probably useful as theoretical background for new transports on its own, but I doubt that's all we want.}
\end{itemize}
\end{frame}

\end{document}

0 comments on commit d1aeed5

Please sign in to comment.