Data Structures for Big Data:

Bloom Filter

Vinicius Vielmo Cogo

Smalltalks, DI, FC/UL. October 16, 2014.

2 / 30

is relative

is not defined by a specific number of TB, PB, EB

is when it becomes big for you

is when your solutions become inefficient/impractical

Data Structures for Big Data

Traditional DSs are subject to the same problems

e.g., lists, trees

(e.g., YARN, NoSQL)

or

(e.g., index, metadata)

reached the point of thinking in new DSs for BD

3 / 30

Outline

Bloom Filter

Use Cases

Implementations

Other Filters

Other Data Structures for Big Data

4 / 30

Bloom Filter

Membership testing

Does my collection contain this element?

5 / 30

Bloom Filter

City

Coimbra

Leiria

6 / 30

Bloom Filter

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Index i

bf[i]

http://billmill.org/bloomfilter-tutorial/ 7 / 30

Bloom Filter

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Index i

bf[i]

City

Coimbra

Leiria

Hash Function

Fnv

Murmur

8 / 30

Bloom Filter

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Index i

bf[i]

City

Coimbra

Leiria

Hash Function

Fnv

Murmur

i=4

i=7

9 / 30

Bloom Filter

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 0 0 1 0 0 1 0 0 0 0 0 0 0

Index i

bf[i]

City

Coimbra

Leiria

Hash Function

Fnv

Murmur

i=4

i=7

10 / 30

Bloom Filter

Index i

bf[i]

City

Coimbra

Leiria

Hash Function

Fnv

Murmur

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 0 0 1 0 0 1 0 0 0 0 0 0 0

11 / 30

Bloom Filter

Index i

bf[i]

City

Coimbra

Leiria

Hash Function

Fnv

Murmur

i=2

i=9

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 0 0 1 0 0 1 0 0 0 0 0 0 0

12 / 30

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 1 0 1 0 0 1 0 1 0 0 0 0 0

Bloom Filter

Index i

bf[i]

City

Coimbra

Leiria

Hash Function

Fnv

Murmur

i=2

i=9

13 / 30

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 1 0 1 0 0 1 0 1 0 0 0 0 0

Bloom Filter

Index i

bf[i]

City

Coimbra

Leiria

Hash Function

Fnv

Murmur

14 / 30

Bloom Filter

City

Braga

Guarda

Coimbra

Lisboa

15 / 30

Result: false

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 1 0 1 0 0 1 0 1 0 0 0 0 0

Bloom Filter

Index i

bf[i]

City

Braga

Guarda

Coimbra

Lisboa

Hash Function

Fnv

Murmur

i=10

i=14

16 / 30

Result: false

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 1 0 1 0 0 1 0 1 0 0 0 0 0

Bloom Filter

Index i

bf[i]

City

Braga

Guarda

Coimbra

Lisboa

Hash Function

Fnv

Murmur

i=2

i=12

17 / 30

Result: true

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 1 0 1 0 0 1 0 1 0 0 0 0 0

Bloom Filter

Index i

bf[i]

City

Braga

Guarda

Coimbra

Lisboa

Hash Function

Fnv

Murmur

i=4

i=7

18 / 30

Result: true (but it is a false positive)

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 1 0 1 0 0 1 0 1 0 0 0 0 0

Bloom Filter

Index i

bf[i]

City

Braga

Guarda

Coimbra

Lisboa

Hash Function

Fnv

Murmur

i=7

i=9

19 / 30

Bloom Filter

DS proposed by Burton Howard Bloom in 1970

Design principles

Space-efficient

Smaller than the original dataset

Time-efficient

Low latency R/W

O(k), which is much smaller than O(n)

High throughput

Probabilistic

E.g., myCollection.mightContain(myObject)

False positives happen (but in a configurable way)

20 / 30

= Optimal number of hash functions

Hash Function

Fnv

Murmur

Important variables

Bloom Filter

= Expected collection size

= Bitmap size

= False positive rate (e.g., 0.0001% or 1 in 1M)

City

Coimbra

Leiria

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14

0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

21 / 30

Important variables

Bloom Filter

22 / 30

Users define two of them (normally n and any other)

The other two are calculated with those equations

Interesting relations:

Bigger collection ( ) Larger bitmap ( )

Bigger collection ( ) More false positives ( )

Larger bitmap ( Less false positives ( )

Larger bitmap ( ) Less hash functions ( )

Less hash functions ( )

Bloom Filter

23 / 30

Bloom filter size vs. False positive rate

Bloom Filter

24 / 30

Use Cases

Reducing unnecessary disk reads

Client BloomFilter Dataset

RAM Hard Disk

1

2

3

F

T

T F

T

1?

2?

3?

necessary

read(2)

unnecessary

read(3)

No

2

No

F

25 / 30

Use Cases

Google BigTable, Apache Cassandra and HBase

Reducing disk lookups

Google Chrome

Lookup a list of known malicious URLs

Bitcoin

Get only the transactions relevant to your wallet

Others

In my Ph.D. work

Lookup a list of known privacy-sensitive DNA

sequences 26 / 30

Implementations

-libraries https://code.google.com/p/guava-libraries/

Orestes-Bloomfilter https://github.com/Baqend/Orestes-Bloomfilter

java-bloomfilter https://github.com/magnuss/java-bloomfilter

java-longfastbloomfilter https://code.google.com/p/java-longfastbloomfilter/

27 / 30

Other Filters

Counting Bloom filters Allow deletions (use a 4-bit counter instead of 1 bit)

Buffered Bloom filters Sub-filters in SSD with buffered R/W exploring bit locality

Quotient and Cascade filters Uses an SSD, instead of the main memory, for scalability

28 / 30

Other DSs (and techniques) for Big Data

Locality-sensitive hashing (LSH) Hashing similar elements into the same bucket with high probability

HyperLogLog for computing cardinality Counting the number of distinct elements in a collection

Log Structured Merge (LSM) trees Indexed access to files with high insert volume and background batch synchronization

29 / 30

Thank you!

Vinicius Vielmo Cogo

Smalltalks, DI, FC/UL. October 16, 2014.