aboutsummaryrefslogtreecommitdiff
path: root/doc
diff options
context:
space:
mode:
authorAlex Auvolat <alex@adnab.me>2021-02-25 11:37:42 +0100
committerAlex Auvolat <alex@adnab.me>2021-02-25 11:37:42 +0100
commitd7e148d3027b7092d7de7e561665d2667199e9bc (patch)
tree73b5beae5bcf2dec3a41c8aec43da975f34e1d9d /doc
parent0522983aec2d7daccb47160bfcdf6c09c048a35e (diff)
downloadgarage-d7e148d3027b7092d7de7e561665d2667199e9bc.tar.gz
garage-d7e148d3027b7092d7de7e561665d2667199e9bc.zip
Description of MultiDC MagLev
Diffstat (limited to 'doc')
-rw-r--r--doc/Load_Balancing.md23
1 files changed, 17 insertions, 6 deletions
diff --git a/doc/Load_Balancing.md b/doc/Load_Balancing.md
index 1d508fa0..a348ebc4 100644
--- a/doc/Load_Balancing.md
+++ b/doc/Load_Balancing.md
@@ -42,7 +42,7 @@ The ring construction that selects `n_token` random positions for each nodes giv
is not well-balanced: the space between the tokens varies a lot, and some partitions are thus bigger than others.
This problem was demonstrated in the original Dynamo DB paper.
-To solve this, we want to apply a second method for partitionning our dataset:
+To solve this, we want to apply a better second method for partitionning our dataset:
1. fix an initially large number of partitions (say 1024) with evenly-spaced delimiters,
@@ -50,7 +50,9 @@ To solve this, we want to apply a second method for partitionning our dataset:
proportionnal to its capacity (which `n_tokens` represented in the first
method)
-I have studied two ways to do the attribution, in a way that is deterministic:
+For now we continue using the multi-DC ring walking described above.
+
+I have studied two ways to do the attribution of partitions to nodes, in a way that is deterministic:
- Min-hash: for each partition, select node that minimizes `hash(node, partition_number)`
- MagLev: see [here](https://blog.acolyer.org/2016/03/21/maglev-a-fast-and-reliable-software-network-load-balancer/)
@@ -67,7 +69,7 @@ for large values), however in both cases:
- the disruption in case of adding/removing a node is not as low as it can be,
as we show with the following method.
-A quick description of MagLev:
+A quick description of MagLev (backend = node, lookup table = ring):
> The basic idea of Maglev hashing is to assign a preference list of all the
> lookup table positions to each backend. Then all the backends take turns
@@ -143,12 +145,21 @@ removing grog moxi : 74.22% 20.61% 4.98% 0.20%
removing grog modi : 75.98% 18.36% 5.27% 0.39%
removing grisou geant : 46.97% 36.62% 15.04% 1.37%
removing grisou gipsie : 49.22% 36.52% 12.79% 1.46%
-on average: 62.94% 27.89% 8.61% 0.57% <-- Worse than custom method
+on average: 62.94% 27.89% 8.61% 0.57% <-- WORSE THAN PREVIOUSLY
```
#### The magical solution: multi-DC aware MagLev
-(insert algorithm description here, in the meantime refer to `method4` in the simulation script)
+Suppose we want to select three replicas for each partition (this is what we do in our simulation and in most Garage deployments).
+We apply MagLev three times consecutively, one for each replica selection.
+The first time is pretty much the same as normal MagLev, but for the following times, when a node runs through its preference
+list to select a partition to replicate, we skip partitions for which adding this node would not bring datacenter-diversity.
+More precisely, we skip a partition in the preference list if:
+
+- the node already replicates the partition (from one of the previous rounds of MagLev)
+- the node is in a datacenter where a node already replicates the partition and there are other datacenters available
+
+Refer to `method4` in the simulation script for a formal definition.
```
##### Multi-DC aware MagLev #####
@@ -180,5 +191,5 @@ removing grog moxi : 80.18% 19.04% 0.78% 0.00%
removing grog modi : 79.69% 19.92% 0.39% 0.00%
removing grisou geant : 44.63% 52.15% 3.22% 0.00%
removing grisou gipsie : 43.55% 52.54% 3.91% 0.00%
-on average: 64.94% 33.33% 1.72% 0.01% <-- VERY GOOD
+on average: 64.94% 33.33% 1.72% 0.01% <-- VERY GOOD (VERY LOW VALUES FOR 2 AND 3 NODES)
```