Skip to content
About
Main content
Model

Deduplication window

Every dedup window in production is a bet that the retry arrives before you forget the key.

Window
how long keys are remembered

Trade-off: Exactness vs. approximation

How it behaves

One input on the horizontal axis, the others held fixed.

Data table
Key storage by window (100,000 keys/s, 32 bytes each)
Window WKey storage
60 s183 MB
67.3 s205 MB
75.6 s231 MB
84.8 s259 MB
95.1 s290 MB
107 s326 MB
120 s366 MB
134 s410 MB
151 s460 MB
169 s517 MB
190 s580 MB
213 s650 MB
239 s730 MB
268 s819 MB
301 s919 MB
338 s1.01 GB
379 s1.13 GB
425 s1.27 GB
477 s1.42 GB
536 s1.6 GB
601 s1.79 GB
675 s2.01 GB
757 s2.26 GB
849 s2.53 GB
953 s2.84 GB
1,070 s3.19 GB
1,200 s3.58 GB
1,350 s4.01 GB
1,510 s4.5 GB
1,700 s5.05 GB
1,900 s5.67 GB
2,140 s6.36 GB
2,400 s7.14 GB
2,690 s8.01 GB
3,020 s8.99 GB
3,390 s10.1 GB
3,800 s11.3 GB
4,260 s12.7 GB
4,780 s14.3 GB
5,370 s16 GB
6,020 s18 GB
6,760 s20.1 GB
7,590 s22.6 GB
8,510 s25.4 GB
9,550 s28.5 GB
10,700 s31.9 GB
12,000 s35.8 GB
13,500 s40.2 GB
15,100 s45.1 GB
17,000 s50.6 GB
19,100 s56.8 GB
21,400 s63.8 GB
24,000 s71.6 GB
26,900 s80.3 GB
30,200 s90.1 GB
33,900 s101 GB
38,100 s113 GB
42,700 s127 GB
47,900 s143 GB
53,800 s160 GB
60,400 s180 GB
67,700 s202 GB
76,000 s227 GB
85,300 s254 GB
95,700 s285 GB
107,000 s320 GB
121,000 s359 GB
135,000 s403 GB
152,000 s452 GB
170,000 s507 GB
191,000 s569 GB
214,000 s639 GB
241,000 s717 GB
270,000 s805 GB
303,000 s903 GB
340,000 s1,010 GB
381,000 s1.11 TB
428,000 s1.25 TB
480,000 s1.4 TB
539,000 s1.57 TB
605,000 s1.76 TB

Worked examples, recomputed

Every number the mechanisms quote for this relationship is recomputed from it, so none of them disagrees with it.

Where it appears

Each mechanism states it in its own names. The general symbol each name stands for is shown with it.