BigCache

Fast, concurrent, evicting in-memory cache written to keep big number of entries without impact on performance. BigCache keeps entries on heap but omits GC for them. To achieve that operations on bytes arrays take place, therefore entries (de)serialization in front of the cache will be needed in most use cases.

Usage

Simple initialization

import "github.com/allegro/bigcache"

cache, _ := bigcache.NewBigCache(bigcache.DefaultConfig(10 * time.Minute))

cache.Set("my-unique-key", []byte("value"))

entry, _ := cache.Get("my-unique-key")
fmt.Println(string(entry))

Custom initialization

When cache load can be predicted in advance then it is better to use custom initialization because additional memory allocation can be avoided in that way.

import "github.com/allegro/bigcache"

config := bigcache.Config{
		Shards: 1024,                       // number of shards (must be a power of 2)
		LifeWindow: 10 * time.Minute,       // time after which entry can be evicted
		MaxEntriesInWindow: 1000 * 10 * 60, // rps * lifeWindow
		MaxEntrySize: 500,                  // max entry size in bytes, used only in initial memory allocation
		Verbose: true,                      // prints information about additional memory allocation
	}

cache, error := bigcache.NewBigCache(config)

cache.Set("my-unique-key", []byte("value"))

if entry, err := cache.Get("my-unique-key"); err == nil {
	fmt.Println(string(entry))
}

Benchmarks

Three caches were compared: bigcache, freecache and map. Benchmark tests were made on MacBook Pro (3 GHz Processor Intel Core i7, 16GB Memory).

Writes and reads

cd caches_bench; go test -bench=. -benchtime=10s ./...

BenchmarkMapSet-4              	10000000	      1430 ns/op
BenchmarkFreeCacheSet-4        	20000000	      1115 ns/op
BenchmarkBigCacheSet-4         	20000000	       873 ns/op
BenchmarkMapGet-4              	30000000	       558 ns/op
BenchmarkFreeCacheGet-4        	20000000	       973 ns/op
BenchmarkBigCacheGet-4         	20000000	       737 ns/op
BenchmarkBigCacheSetParallel-4 	30000000	       545 ns/op
BenchmarkFreeCacheSetParallel-4	20000000	       654 ns/op
BenchmarkBigCacheGetParallel-4 	50000000	       426 ns/op
BenchmarkFreeCacheGetParallel-4	50000000	       715 ns/op

Writes and reads in bigcache are faster than in freecache. Writes to map are the slowest.

GC pause time

cd caches_bench; go run caches_gc_overhead_comparsion.go

Number of entries:  20000000
GC pause for bigcache:  27.81671ms
GC pause for freecache:  30.218371ms
GC pause for map:  11.590772251s

Test shows how long are the GC pauses for caches filled with 20mln of entries. Bigcache and freecache have very similar GC pause time. It is clear that both reduce GC overhead in contrast to map which GC pause time took more than 10 seconds.

How it works

BigCache relays on optimization presented in 1.5 version of Go (issue-9477). This optimization states that if map without pointers in keys and values is used then GC will omit it’s content. Therefore BigCache uses map[uint64]uint32 where keys are hashed and values are offsets of entries.

Entries are kept in bytes array, to omit GC again. Bytes array size can grow to gigabytes without impact on performance because GC will only see single pointer to it.

Bigcache vs Freecache

Both caches provide the same core features but they reduce GC overhead in different ways. Bigcache relays on map[uint64]uint32, freecache implements its own mapping built on slices to reduce number of pointers.

Results from benchmark tests are presented above. One of the advantage of bigcache over freecache is that you don’t need to know the size of cache in advance, because when bigcache is full, it allocates additional memory for new entries instead of overwriting existing ones as freecache does currently.

More

Bigcache genesis is described in allegro.tech blog post: writing a very fast cache service in Go

License

BigCache is released under the Apache 2.0 license (see LICENSE)