auto

notes/atlas.md

@@ -14,13 +14,13 @@ EPaxos but......faster.
 
 
 ### Slow Path
-size is only *f+1*!
+size is only *f+1*!      (flexible paxos)
 
 ## Recovery
 
-Background:
+Background, fast path iff:
 
-   *U<sub>Q</sub>dep = U(f)<sub>Q</sub>dep*        (line 15 in protocol)
+&nbsp;&nbsp;&nbsp;   *U<sub>Q</sub>dep = U(f)<sub>Q</sub>dep*        (line 15 in protocol)
 
 ![protocol](atlasFig2.png)
 
@@ -30,11 +30,13 @@ after coordinator and one other die, but how to tell whether the fast
 path was to be taken?
 
 How to distinguish 2a from this version:
-- "2" didn't report "c", so we should not take fast path
+- "2" didn't report "c", so fast path should not be taken
 - assume (1) and (4) die
+
 ![noC](atlasNoC.png)
 
-Come back to this.
+Slow path quorum is *f+1* = 3, *but all f+1 need to reply*, meaning
+2,3,5 have to respond, and get a *D* that does not include *c*.
 
 
 
@@ -60,8 +62,15 @@ Recovery states:
 
 ![recovery](atlasRecovery.png?1)
 
-### When do they die?
-- if coordinator dies after sending a commit that is received elsewhere
+## Performance
+
+![recovery](atlasPerf1.png?1)
+Note that Epaxos w/ *f=1* also always takes the fast path, though they
+don't mention this.
+
+![recovery](atlasPerf2.png?1)
+Apples to oranges here a bit. Epaxos fault tolerance increasing w/
+more sites, but Atlas not.
 
 
 ## Issues

notes/titan.md~

+# State-Machine Replication for Planet-Scale Systems (ATLAS)
+
+EPaxos but......faster.
+
+- fastpath quorum: $$\lfloor n/2 \rfloor + f$$
\ No newline at end of file