aboutsummaryrefslogtreecommitdiff
path: root/chapter/4
diff options
context:
space:
mode:
authorcnnrznn <cnnrznn@gmail.com>2016-12-06 14:51:02 -0500
committerGitHub <noreply@github.com>2016-12-06 14:51:02 -0500
commitf34fe47ae54d8790961d12d571df719bb1535b2d (patch)
treec9dd06550c34869518054f4802475d38923357dd /chapter/4
parent5adc3c9e9bf6998658240b6849ffac281e39199f (diff)
Update dist-langs.md
Diffstat (limited to 'chapter/4')
-rw-r--r--chapter/4/dist-langs.md16
1 files changed, 16 insertions, 0 deletions
diff --git a/chapter/4/dist-langs.md b/chapter/4/dist-langs.md
index 5bfea62..7b1412b 100644
--- a/chapter/4/dist-langs.md
+++ b/chapter/4/dist-langs.md
@@ -4,6 +4,13 @@ title: "Distributed Programming Languages"
by: "A Systems Person"
---
+### Problems of Distributed Programming
+
+* Partial failure
+* Consistency
+* Efficiency
+* Scallability
+
### Three major approaches to distributed languages:
#### Shared Memory
@@ -40,6 +47,15 @@ Some languages that use this model are:
* MapReduce (Spark, Hadoop, etc.)
* Orleans (Wait, what??)
+#### Which is best? Why?
+
+Why is MR all-the-rage?
+
+* Dataflow / MapReduce fundamentally changed the programming style for distributed systems
+* Other models (Actor, DSM) tried to mask distribution
+* By changing the style, programs need necessarily consider communication patterns (disk, network)
+* Although, system may still handle fault tolerance
+
### Why GPL's not DSL's?
* problem of domain-composition