docs(journal): add more slides for journal presentation

doc/demo-journal/slides-journal.md

@@ -1,7 +1,7 @@
 ---
 title: Towards conveniently debuggable distributed systems
 author: Stevan Andjelkovic \and Daniel Gustafsson
-date: 31th Jan, 2022
+date: 24th Mar, 2022
 header-includes:
   - \definecolor{links}{HTML}{2A1B81}
   - \hypersetup{colorlinks,linkcolor=,urlcolor=links}
@@ -108,16 +108,70 @@ nocite: |
 * Three (virtual) files (clean, active, dirty)
 * Circular buffer implemented on top of `mmap`ed byte array
 * `recv` zero-copied straight to byte array (and persisted)
+* Lock- and wait-free concurrency
 
-# Built-in profiler/metrics
+# Design of the journal
+
+![Journal](./images/journal_0.png){ height=90% }
+
+# Design of the journal
+
+![Journal](./images/journal_1.png){ height=90% }
+
+# Design of the journal
+
+![Journal](./images/journal_2.png){ height=90% }
+
+# Design of the journal
+
+![Journal](./images/journal_3.png){ height=90% }
+
+# Design of the journal
+
+![Journal](./images/journal_4.png){ height=90% }
+
+# Design of the journal
+
+![Journal](./images/journal_5.png){ height=90% }
+
+# Design of the journal
+
+![Journal](./images/journal_6.png){ height=90% }
+
+# Design of the journal
 
-* Idea due to Tyler "sled" Neely
-* Counters
-* Histograms
+![Journal](./images/journal_7.png){ height=90% }
+
+# Design of the journal
+
+![Journal](./images/journal_8.png){ height=90% }
+
+# Design of the journal
+
+![Journal](./images/journal_9.png){ height=90% }
+
+# Design of the journal
+
+![Journal](./images/journal_10.png){ height=90% }
+
+# Design of the journal
+
+![Journal](./images/journal_11.png){ height=90% }
+
+# Design of the journal
+
+![Journal](./images/journal_12.png){ height=90% }
 
 # Design of the event loop
 
-* TODO
+1. Fork a webserver where the request handlers have concurrent write access to a
+   shared journal;
+
+2. Request handlers merely write the entire request into the journal (this gives
+   us a linearised sequence of requests);
+
+3. A separate "worker" thread reads the journal entires and updates the state of
+   the database.
 
 # Demo
 
@@ -143,6 +197,25 @@ nocite: |
   and journal using deterministic replay to show how the state machines change
   over time (whether the server is running or not).
 
+# Amdahl's law vs the Universal scalability law
+
+* *C(N) = N / (1 + a(N - 1) + ((b \* N) \* (N - 1)))*
+  + *C* = capacity or throughput
+  + *N* = number of processors
+  + *a* = contention penality (time under some kind of lock)
+  + *b* = coherence penality (time to agree, e.g. load the cache line with the shared reference)
+
+![USL](./images/amdahl_vs_usl.jpg){ width=60% }
+
+# Built-in profiler/metrics
+
+* Ideas due to Tyler "sled" Neely and Thompson et al
+* (Atomic) counters (an 64-bit signed integer)
+* Histograms (2 + 2^16 counters) uses `log` and `exp` for
+  compression/decompression
+* `mmap`ed bytearray (8 bytes per counter) allows for atomic update (`lock xadd`
+  in x86 asm) and access for different processes
+
 # Summary
 
 * We have shown how to use the journal to: