Add slides for the defense
They contain _a lot_ of commented lines… I don't care.master
parent
38b03f9260
commit
7941e4fd7b
Binary file not shown.
|
After Width: | Height: | Size: 13 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 211 KiB |
Binary file not shown.
|
After Width: | Height: | Size: 31 KiB |
@ -0,0 +1,289 @@
|
||||
\documentclass{beamer}
|
||||
|
||||
\usetheme{Madrid}
|
||||
% Ref: https://tex.stackexchange.com/a/692
|
||||
\setbeamertemplate{navigation symbols}{}
|
||||
|
||||
% We may as well create a PDF/A, why not. (I am too lazy to invent my
|
||||
% own header, instead I just pick line from the thesis' one :-))
|
||||
|
||||
\usepackage[a-2u]{pdfx}
|
||||
\usepackage[utf8]{inputenc}
|
||||
\usepackage{lmodern}
|
||||
|
||||
\usepackage{xcolor}
|
||||
\usepackage{listings}
|
||||
|
||||
\hypersetup{unicode}
|
||||
\hypersetup{breaklinks=true}
|
||||
|
||||
\def\uv#1{``#1''}
|
||||
|
||||
|
||||
\title{Visualising OSPF topology}
|
||||
\subtitle{Motivation and design of Birdvisu} %FIXME?
|
||||
\author{Pavel Turinský}
|
||||
\date{7 September 2023}
|
||||
%\date{07-09-2023} % Looks weird imho :-/
|
||||
|
||||
\begin{document}
|
||||
|
||||
\begin{frame}
|
||||
\titlepage
|
||||
\end{frame}
|
||||
|
||||
\iffalse % Maybe drop, in that case, comment the next line.
|
||||
%\else
|
||||
\begin{frame}
|
||||
\frametitle{Outline}
|
||||
\begin{itemize}
|
||||
\item Introduction
|
||||
\begin{itemize}
|
||||
\item What is Birdvisu
|
||||
\item Reasons for creating Birdvisu
|
||||
\item Goals
|
||||
\end{itemize}
|
||||
% intro:
|
||||
%% what
|
||||
%% why (comparison)
|
||||
%% goals
|
||||
\item Technologies
|
||||
\begin{itemize}
|
||||
\item OSPF overview
|
||||
\item Interaction with BIRD
|
||||
\end{itemize}
|
||||
% tech stack
|
||||
%% OSPF summary + history
|
||||
%% BIRD: why and how
|
||||
%%% the dump format
|
||||
%%% note gennet
|
||||
\item Design
|
||||
\begin{itemize}
|
||||
\item Retrieving topologies
|
||||
\item Annotation
|
||||
\item Drawing
|
||||
\end{itemize}
|
||||
% design
|
||||
%% the three phases
|
||||
\item Results
|
||||
\begin{itemize}
|
||||
\item Small networks
|
||||
\item Large: czela.net
|
||||
\end{itemize}
|
||||
% results
|
||||
%% czela,net
|
||||
%%% known deficiencies
|
||||
|
||||
\item Future
|
||||
% future?
|
||||
%% exports, integration
|
||||
%% saving layouts
|
||||
% gennet?
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
\fi
|
||||
|
||||
%TODO: slide about Birdvisu is, very highlevel
|
||||
|
||||
|
||||
\begin{frame}
|
||||
\frametitle{Project goals}
|
||||
The Birdvisu project aims to present the network topology, as known
|
||||
by OSPF, to the administrator and allow them to analyse it.
|
||||
|
||||
\bigskip
|
||||
|
||||
Goals:
|
||||
\begin{itemize}
|
||||
\item As independent on other services as possible
|
||||
\item Easy deployment
|
||||
\begin{itemize}
|
||||
\item Runnable on admin's laptop
|
||||
\end{itemize}
|
||||
\item No changes to other hosts required
|
||||
\item It should be easy to create new analysis tools
|
||||
\item Target: administrators of homelabs to middle-sized systems
|
||||
\end{itemize}
|
||||
|
||||
\medskip
|
||||
|
||||
To our knowledge, there is no such application yet.
|
||||
|
||||
\medskip
|
||||
|
||||
Birdvisu is implemented in Python 3 and Qt 6.
|
||||
\end{frame}
|
||||
% \begin{frame}
|
||||
% %\frametitle{Why visualise/analyse topologies}
|
||||
% \frametitle{Motivation and goals}
|
||||
% Existing analysis/monitoring tools:
|
||||
% \begin{itemize}
|
||||
% \item Need to collect data on individual hosts,
|
||||
% \item do may not have accurate data (network splits)
|
||||
% \item or need to be deployed on a server
|
||||
% \end{itemize}
|
||||
% \dots yet in systems routed by OSPF each router knows all of this.
|
||||
%
|
||||
% \vfill
|
||||
%
|
||||
% The aim of Birdvisu is to present this information to
|
||||
% administrators.
|
||||
% \begin{itemize}
|
||||
% \item single-host, easy-to-run, few dependencies on services
|
||||
% \item Target: administrators of homelabs to middle-sized systems
|
||||
% \end{itemize}
|
||||
% \end{frame}
|
||||
|
||||
% TODO: wanted features?
|
||||
|
||||
%%% design
|
||||
|
||||
\begin{frame}
|
||||
\frametitle{Birdvisu design}
|
||||
% uwu uwu FIXME TODO
|
||||
The program consists of three main steps:
|
||||
\begin{itemize}
|
||||
\item Obtaining the topologies
|
||||
% shenanigans?
|
||||
% non-matchability of uwu uwu.
|
||||
\begin{itemize}
|
||||
\item Reference (from file), current state
|
||||
\item Representation: weighted directed multigraph
|
||||
% \item Reference and current one merged into a single object
|
||||
% \item Extracted from BIRD or loaded from saved files
|
||||
% \item Networks can have various forms: orieiwnted weighted multigraph
|
||||
\end{itemize}
|
||||
\item Annotating
|
||||
\begin{itemize}
|
||||
\item Extending the topology with analysis results, other data, \dots
|
||||
% \item Many single-purpose annotators
|
||||
% \item e.g. compute topology differences, find shortest paths
|
||||
\end{itemize}
|
||||
\item Displaying the result
|
||||
% own try of layouting alg
|
||||
% each highlight has its StyleAnnotator
|
||||
\begin{itemize}
|
||||
\item Reduction of the topology to a simple graph
|
||||
\item Highlight relevant annotator results
|
||||
|
||||
\medskip
|
||||
|
||||
\item Alternatively export the annotated topology (in future)
|
||||
% \item Use Qt's features: dragging, interaction
|
||||
% \item Layout loaded from file or determined by ad-hoc
|
||||
% algorithm
|
||||
% \item Reuses Annotator design
|
||||
\end{itemize}
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
|
||||
\begin{frame}
|
||||
\frametitle{OSPF}
|
||||
\begin{itemize}
|
||||
\item Link-state IGP: routers share information about the current state of the system
|
||||
\item Long evolution (1989), multiple extensions, rather complex
|
||||
\item Different versions for IPv4 and IPv6 routing
|
||||
\begin{itemize}
|
||||
\item Same idea, quite different internals
|
||||
\end{itemize}
|
||||
\end{itemize}
|
||||
|
||||
\bigskip
|
||||
|
||||
Instead of implementing OSPF ourselves, we extract the state from existing routing daemon.
|
||||
\end{frame}
|
||||
|
||||
\begin{frame}
|
||||
\frametitle{BIRD Internet Routing Daemon}
|
||||
\begin{itemize}
|
||||
\item Originally SW project at this faculty, now maintained by CZ.NIC
|
||||
\item Experience from deployment at home, Dept{.} of Applied Mathematics
|
||||
\item One of two maintained OSPF daemons for general-purpose HW %FIXME: check BrE
|
||||
\begin{itemize}
|
||||
\item future extensions more likely to be implemented
|
||||
\end{itemize}
|
||||
\end{itemize}
|
||||
|
||||
\bigskip
|
||||
|
||||
\begin{itemize}
|
||||
\item Currently no machine-readable export
|
||||
\begin{itemize}
|
||||
\item The output for humans contains sufficient information
|
||||
\end{itemize}
|
||||
\item No dynamic output -- needs polling
|
||||
\end{itemize}
|
||||
|
||||
% logo? snippet of output?
|
||||
\end{frame}
|
||||
|
||||
\begin{frame}
|
||||
\frametitle{Annotators}
|
||||
\begin{itemize}
|
||||
\item Can add arbitrary$^*$ data to vertices and edges
|
||||
\begin{itemize}
|
||||
\item Exportable in future
|
||||
\end{itemize}
|
||||
\item Can depend on each other
|
||||
\item Idea: many simple and single-purpose annotators
|
||||
\item Uses: analysis, providing additional information, \dots
|
||||
\begin{itemize}
|
||||
\item e.g. comparing topologies, finding shortest path DAGs
|
||||
\end{itemize}
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
|
||||
\begin{frame}
|
||||
\frametitle{Visualisation}
|
||||
\begin{itemize}
|
||||
\item Reuse Annotators for styling vertices and edges
|
||||
\item Reduce topology to a simple graph by taking the most relevant edge
|
||||
\begin{itemize}
|
||||
\item lightest edge with positive cost
|
||||
\end{itemize}
|
||||
\item Layout can be loaded from a file
|
||||
\begin{itemize}
|
||||
\item New vertices placed to proximity of neighbours
|
||||
\end{itemize}
|
||||
\item Qt provides UI features: dragging, context menus
|
||||
\end{itemize}
|
||||
\end{frame}
|
||||
|
||||
% Results and examples
|
||||
|
||||
\begin{frame}
|
||||
\frametitle{Small network example: Gennet}
|
||||
\begin{center}
|
||||
\includegraphics[keepaspectratio,height=0.7\textheight]{gennet.png}
|
||||
\\
|
||||
Highlighting differences from expectation on a test network (Gennet)
|
||||
\end{center}
|
||||
\end{frame}
|
||||
|
||||
\begin{frame}
|
||||
\frametitle{Medium-sized network: Czela.net}
|
||||
Community network system in Čelákovice, 45 routers, 210 networks (incl. external), about 1600 people
|
||||
\begin{center}
|
||||
\includegraphics[keepaspectratio,height=0.6\textheight]{czmess.png}
|
||||
\\
|
||||
Our layouting approach is not great without any initial placements.
|
||||
\end{center}
|
||||
\end{frame}
|
||||
|
||||
\begin{frame}
|
||||
\frametitle{Fixing Czela.net}
|
||||
\begin{center}
|
||||
\includegraphics[keepaspectratio,height=0.7\textheight]{czbug.png}
|
||||
\\
|
||||
Misconfiguration, found after some dragging of vertices
|
||||
\end{center}
|
||||
\end{frame}
|
||||
|
||||
\begin{frame}
|
||||
\begin{center}
|
||||
\begin{Large}
|
||||
Thank you!
|
||||
\end{Large}
|
||||
\end{center}
|
||||
\end{frame}
|
||||
\end{document}
|
||||
Loading…
Reference in New Issue