diff options
Diffstat (limited to 'doc/talks/2024-02-03-fosdem/talk.tex')
| -rw-r--r-- | doc/talks/2024-02-03-fosdem/talk.tex | 94 |
1 files changed, 73 insertions, 21 deletions
diff --git a/doc/talks/2024-02-03-fosdem/talk.tex b/doc/talks/2024-02-03-fosdem/talk.tex index 6c3e47f0..cebdb6b6 100644 --- a/doc/talks/2024-02-03-fosdem/talk.tex +++ b/doc/talks/2024-02-03-fosdem/talk.tex @@ -1,5 +1,5 @@ \nonstopmode -\documentclass[aspectratio=169]{beamer} +\documentclass[aspectratio=169,xcolor={svgnames}]{beamer} \usepackage[utf8]{inputenc} % \usepackage[frenchb]{babel} \usepackage{amsmath} @@ -176,7 +176,12 @@ \begin{frame} \frametitle{CRDTs / weak consistency instead of consensus} - Consensus can be implemented reasonably well in practice, so why avoid it? + + \underline{Internally, Garage uses only CRDTs} (conflict-free replicated data types) + + \vspace{2em} + No Raft, ... Why? + \vspace{1em} \begin{itemize} \item<2-> \textbf{Software complexity} @@ -191,8 +196,6 @@ \item<6-> \textbf{Takes time to reconverge} when disrupted (e.g. node going down) \end{itemize} \end{itemize} - \vspace{2em} - \visible<7->{\underline{Internally, Garage uses only CRDTs} (conflict-free replicated data types)} \end{frame} \begin{frame} @@ -263,11 +266,9 @@ \vspace{1em} \item Replication modes with 1 or 2 copies / weaker consistency \vspace{1em} - \item Kubernetes integration + \item Kubernetes integration for node discovery \vspace{1em} \item Admin API (v0.7.2) - \vspace{1em} - \item Experimental K2V API (v0.7.2) \end{itemize} \end{frame} @@ -323,7 +324,8 @@ \end{itemize} \vspace{2em} - \textbf{LMDB:} very stable, good performance, reasonably small files on disk + \textbf{LMDB:} very stable, good performance, file size is reasonable\\ + \textbf{Sqlite} also available as a second choice \vspace{1em} Sled will be removed in Garage v1.0 @@ -417,15 +419,15 @@ \textbf{Partition} & \textbf{Node 1} & \textbf{Node 2} & \textbf{Node 3} \\ \hline \hline - Partition 0 & Io (jupiter) & Drosera (atuin) & Courgette (neptune) \\ + Partition 0 & df-ymk (bespin) & Abricot (scorpio) & Courgette (neptune) \\ \hline - Partition 1 & Datura (atuin) & Courgette (neptune) & Io (jupiter) \\ + Partition 1 & Ananas (scorpio) & Courgette (neptune) & df-ykl (bespin) \\ \hline - Partition 2 & Io(jupiter) & Celeri (neptune) & Drosera (atuin) \\ + Partition 2 & df-ymf (bespin) & Celeri (neptune) & Abricot (scorpio) \\ \hline \hspace{1em}$\vdots$ & \hspace{1em}$\vdots$ & \hspace{1em}$\vdots$ & \hspace{1em}$\vdots$ \\ \hline - Partition 255 & Concombre (neptune) & Io (jupiter) & Drosera (atuin) \\ + Partition 255 & Concombre (neptune) & df-ykl (bespin) & Abricot (scorpio) \\ \hline \end{tabular} \end{center} @@ -484,9 +486,9 @@ \vspace{1em} {\small - \textbf{Property:} If node $A$ did an operation $write(x)$ and received an OK response,\\ - \hspace{2cm} and node $B$ starts an operation $read()$ after $A$ received OK,\\ - \hspace{2cm} then $B$ will read a value $x' \sqsupseteq x$. + \textbf{Property:} If client 1 did an operation $write(x)$ and received an OK response,\\ + \hspace{2cm} and client 2 starts an operation $read()$ after client 1 received OK,\\ + \hspace{2cm} then client 2 will read a value $x' \sqsupseteq x$. } \vspace{1.5em} @@ -539,10 +541,53 @@ \item We rely on quorums $k > n/2$ within each partition:\\ $$n=3,~~~~~~~k\ge 2$$ \item<2-> When rebalancing, the set of nodes responsible for a partition can change:\\ - $$\{A, B, C\} \to \{A, D, E\}$$ - \vspace{.01em} - \item<3-> During the rebalancing, $D$ and $E$ don't yet have the data,\\ - ~~~~~~~~~~~~~~~~~~~and $B$ and $C$ want to get rid of the data to free up space\\ + + \vspace{1em} + \begin{minipage}{.04\linewidth}~ + \end{minipage} + \begin{minipage}{.40\linewidth} + {\tiny + \begin{tabular}{|l|l|l|l|} + \hline + \textbf{Partition} & \textbf{Node 1} & \textbf{Node 2} & \textbf{Node 3} \\ + \hline + \hline + Partition 0 & \textcolor{Crimson}{df-ymk} & Abricot & \textcolor{Crimson}{Courgette} \\ + \hline + Partition 1 & Ananas & \textcolor{Crimson}{Courgette} & \textcolor{Crimson}{df-ykl} \\ + \hline + Partition 2 & \textcolor{Crimson}{df-ymf} & \textcolor{Crimson}{Celeri} & Abricot \\ + \hline + \hspace{1em}$\dots$ & \hspace{1em}$\dots$ & \hspace{1em}$\dots$ & \hspace{1em}$\dots$ \\ + \hline + \end{tabular} + } + \end{minipage} + \begin{minipage}{.04\linewidth} + $\to$ + \end{minipage} + \begin{minipage}{.40\linewidth} + {\tiny + \begin{tabular}{|l|l|l|l|} + \hline + \textbf{Partition} & \textbf{Node 1} & \textbf{Node 2} & \textbf{Node 3} \\ + \hline + \hline + Partition 0 & \textcolor{ForestGreen}{Dahlia} & Abricot & \textcolor{ForestGreen}{Eucalyptus} \\ + \hline + Partition 1 & Ananas & \textcolor{ForestGreen}{Euphorbe} & \textcolor{ForestGreen}{Doradille} \\ + \hline + Partition 2 & \textcolor{ForestGreen}{Dahlia} & \textcolor{ForestGreen}{Echinops} & Abricot \\ + \hline + \hspace{1em}$\dots$ & \hspace{1em}$\dots$ & \hspace{1em}$\dots$ & \hspace{1em}$\dots$ \\ + \hline + \end{tabular} + } + \end{minipage} + + \vspace{2em} + \item<3-> During the rebalancing, new nodes don't yet have the data,\\ + ~~~~~~~~~~~~~~~~~~~and old nodes want to get rid of the data to free up space\\ \vspace{1.2em} $\to$ risk of inconsistency, \textbf{how to coordinate?} \end{itemize} @@ -589,7 +634,7 @@ \end{frame} \begin{frame} - \frametitle{Towards v1.0} + \frametitle{Towards v1.0...} Focus on \underline{security \& stability} \vspace{2em} \begin{itemize} @@ -603,6 +648,13 @@ \end{itemize} \end{frame} +\begin{frame} + \frametitle{...and beyond!} + \begin{center} + \includegraphics[width=.6\linewidth]{../assets/survey_requested_features.png} + \end{center} +\end{frame} + % ======================================== OPERATING % ======================================== OPERATING % ======================================== OPERATING @@ -684,7 +736,7 @@ \end{itemize} \vspace{.5em} \end{itemize} - Current deployments: $< 10$ TB, we don't have much experience with more + Our deployments: $< 10$ TB. Some people have done more! \end{frame} |