Merge pull request #25 from mailgun/thrawn/develop

Consistent hash improvements

CHANGELOG

@@ -4,6 +4,12 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## [2.2.1] - 2021-01-13
+### Changes
+* Now uses the much faster fnv1
+* Now md5 hashs the keys to help distribute hosts more evenly in some
+  cases.
+
 ## [2.2.0] - 2019-07-09
 ### Added
 * Added `SetLogger()` to pass in a logrus entry for logging peer errors

consistenthash/consistenthash.go

@@ -18,12 +18,15 @@ limitations under the License.
 package consistenthash
 
 import (
-	"hash/crc32"
+	"crypto/md5"
+	"fmt"
 	"sort"
 	"strconv"
+
+	"github.com/segmentio/fasthash/fnv1"
 )
 
-type Hash func(data []byte) uint32
+type Hash func(data []byte) uint64
 
 type Map struct {
 	hash     Hash
@@ -39,7 +42,7 @@ func New(replicas int, fn Hash) *Map {
 		hashMap:  make(map[int]string),
 	}
 	if m.hash == nil {
-		m.hash = crc32.ChecksumIEEE
+		m.hash = fnv1.HashBytes64
 	}
 	return m
 }
@@ -53,7 +56,7 @@ func (m *Map) IsEmpty() bool {
 func (m *Map) Add(keys ...string) {
 	for _, key := range keys {
 		for i := 0; i < m.replicas; i++ {
-			hash := int(m.hash([]byte(strconv.Itoa(i) + key)))
+			hash := int(m.hash([]byte(fmt.Sprintf("%x", md5.Sum([]byte(strconv.Itoa(i)+key))))))
 			m.keys = append(m.keys, hash)
 			m.hashMap[hash] = key
 		}

consistenthash/consistenthash_test.go

@@ -18,51 +18,45 @@ package consistenthash
 
 import (
 	"fmt"
-	"strconv"
+	"math/rand"
+	"net"
 	"testing"
+	"time"
+
+	"github.com/segmentio/fasthash/fnv1"
 )
 
 func TestHashing(t *testing.T) {
 
 	// Override the hash function to return easier to reason about values. Assumes
 	// the keys can be converted to an integer.
-	hash := New(3, func(key []byte) uint32 {
-		i, err := strconv.Atoi(string(key))
-		if err != nil {
-			panic(err)
-		}
-		return uint32(i)
-	})
+	hash := New(512, nil)
 
-	// Given the above hash function, this will give replicas with "hashes":
-	// 2, 4, 6, 12, 14, 16, 22, 24, 26
 	hash.Add("6", "4", "2")
 
 	testCases := map[string]string{
-		"2":  "2",
-		"11": "2",
-		"23": "4",
-		"27": "2",
+		"12,000":    "4",
+		"11":        "6",
+		"500,000":   "4",
+		"1,000,000": "2",
 	}
 
 	for k, v := range testCases {
-		if hash.Get(k) != v {
-			t.Errorf("Asking for %s, should have yielded %s", k, v)
+		if got := hash.Get(k); got != v {
+			t.Errorf("Asking for %s, should have yielded %s; got %s instead", k, v, got)
 		}
 	}
 
-	// Adds 8, 18, 28
 	hash.Add("8")
 
-	// 27 should now map to 8.
-	testCases["27"] = "8"
+	testCases["11"] = "8"
+	testCases["1,000,000"] = "8"
 
 	for k, v := range testCases {
-		if hash.Get(k) != v {
-			t.Errorf("Asking for %s, should have yielded %s", k, v)
+		if got := hash.Get(k); got != v {
+			t.Errorf("Asking for %s, should have yielded %s; got %s instead", k, v, got)
 		}
 	}
-
 }
 
 func TestConsistency(t *testing.T) {
@@ -77,13 +71,52 @@ func TestConsistency(t *testing.T) {
 	}
 
 	hash2.Add("Becky", "Ben", "Bobby")
+	hash1.Add("Becky", "Ben", "Bobby")
 
 	if hash1.Get("Ben") != hash2.Get("Ben") ||
 		hash1.Get("Bob") != hash2.Get("Bob") ||
 		hash1.Get("Bonny") != hash2.Get("Bonny") {
 		t.Errorf("Direct matches should always return the same entry")
 	}
+}
+
+func TestDistribution(t *testing.T) {
+	hosts := []string{"a.svc.local", "b.svc.local", "c.svc.local"}
+	rand.Seed(time.Now().Unix())
+	const cases = 10000
+
+	strings := make([]string, cases)
 
+	for i := 0; i < cases; i++ {
+		r := rand.Int31()
+		ip := net.IPv4(192, byte(r>>16), byte(r>>8), byte(r))
+		strings[i] = ip.String()
+	}
+
+	hashFuncs := map[string]Hash{
+		"fasthash/fnv1": fnv1.HashBytes64,
+	}
+
+	for name, hashFunc := range hashFuncs {
+		t.Run(name, func(t *testing.T) {
+			hash := New(512, hashFunc)
+			hostMap := map[string]int{}
+
+			for _, host := range hosts {
+				hash.Add(host)
+				hostMap[host] = 0
+			}
+
+			for i := range strings {
+				host := hash.Get(strings[i])
+				hostMap[host]++
+			}
+
+			for host, a := range hostMap {
+				t.Logf("host: %s, percent: %f", host, float64(a)/cases)
+			}
+		})
+	}
 }
 
 func BenchmarkGet8(b *testing.B)   { benchmarkGet(b, 8) }

go.mod

@@ -2,6 +2,7 @@ module github.com/mailgun/groupcache/v2
 
 require (
 	github.com/golang/protobuf v1.3.1
+	github.com/segmentio/fasthash v1.0.3
 	github.com/sirupsen/logrus v1.6.0
 )
 

go.sum

@@ -12,3 +12,5 @@ github.com/stretchr/testify v1.2.2 h1:bSDNvY7ZPG5RlJ8otE/7V6gMiyenm9RtJ7IUVIAoJ1
 github.com/stretchr/testify v1.2.2/go.mod h1:a8OnRcib4nhh0OaRAV+Yts87kKdq0PP7pXfy6kDkUVs=
 golang.org/x/sys v0.0.0-20190422165155-953cdadca894 h1:Cz4ceDQGXuKRnVBDTS23GTn/pU5OE2C0WrNTOYK1Uuc=
 golang.org/x/sys v0.0.0-20190422165155-953cdadca894/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
+github.com/segmentio/fasthash v1.0.3 h1:EI9+KE1EwvMLBWwjpRDc+fEM+prwxDYbslddQGtrmhM=
+github.com/segmentio/fasthash v1.0.3/go.mod h1:waKX8l2N8yckOgmSsXJi7x1ZfdKZ4x7KRMzBtS3oedY=