2025-10-14 23:12:06 +02:00
13 changed files with 1475 additions and 295 deletions
--- a/README.md
+++ b/README.md
@@ -1,5 +1,36 @@
 # Go Cart Actor

+## Migration Notes (Ring-based Ownership Transition)
+
+This release removes the legacy ConfirmOwner ownership negotiation RPC in favor of deterministic ownership via the consistent hashing ring.
+
+Summary of changes:
+- ConfirmOwner RPC removed from the ControlPlane service.
+- OwnerChangeRequest message removed (was only used by ConfirmOwner).
+- OwnerChangeAck retained solely as the response type for the Closing RPC.
+- SyncedPool now relies exclusively on the ring for ownership (no quorum negotiation).
+- Remote proxy creation includes a bounded readiness retry to reduce first-call failures.
+- New Prometheus ring metrics: 
+  - cart_ring_epoch
+  - cart_ring_hosts
+  - cart_ring_vnodes
+  - cart_ring_host_share{host}
+  - cart_ring_lookup_local_total
+  - cart_ring_lookup_remote_total
+
+Action required for consumers:
+1. Regenerate protobuf code after pulling (requires protoc-gen-go and protoc-gen-go-grpc installed).
+2. Remove any client code or automation invoking ConfirmOwner (calls will now return UNIMPLEMENTED if using stale generated stubs).
+3. Update monitoring/alerts that referenced ConfirmOwner or ownership quorum failures—use ring metrics instead.
+4. If you previously interpreted “ownership flapping” via ConfirmOwner logs, now check for:
+   - Rapid changes in ring epoch (cart_ring_epoch)
+   - Host churn (cart_ring_hosts)
+   - Imbalance in vnode distribution (cart_ring_host_share)
+
+No data migration is necessary; cart IDs and grain state are unaffected.
+
+---
+
 A distributed cart management system using the actor model pattern.

 ## Prerequisites
@@ -240,8 +271,8 @@ Responsibilities:

 1. Discovery integration (via a `Discovery` interface) adds/removes hosts.
 2. Periodic ping health checks (ControlPlane.Ping).
-3. Ownership negotiation:
-   - On first contention / unknown owner, node calls `ConfirmOwner` on peers to achieve quorum before making a local grain authoritative.
+3. Ring-based deterministic ownership:
+   - Ownership is derived directly from the consistent hashing ring (no quorum RPC or `ConfirmOwner`).
 4. Remote spawning:
   - When a remote host reports its cart ids (`GetCartIds`), the pool creates remote proxies for fast routing.

@@ -270,7 +301,6 @@ Defined in `proto/control_plane.proto`:
 | `Ping` | Liveness; increments missed ping counter if failing. |
 | `Negotiate` | Merges membership views; used after discovery events. |
 | `GetCartIds` | Enumerate locally owned carts for remote index seeding. |
-| `ConfirmOwner` | Quorum acknowledgment for ownership claim. |
 | `Closing` | Graceful shutdown notice; peers remove host & associated remote grains. |

 ### Ownership / Quorum Rules
@@ -347,7 +377,7 @@ Defined in `proto/control_plane.proto`:
 ## gRPC Interfaces

 - **CartActor**: Per-mutation unary RPCs + `GetState`. (Checkout logic intentionally excluded; handled at HTTP layer.)
- **ControlPlane**: Cluster coordination (Ping, Negotiate, ConfirmOwner, etc.).
+- **ControlPlane**: Cluster coordination (Ping, Negotiate, GetCartIds, Closing) — ownership now ring-determined (no ConfirmOwner).

 **Ports** (default / implied):
 - CartActor & ControlPlane share the same gRPC server/listener (single port, e.g. `:1337`).
@@ -396,7 +426,7 @@ Defined in `proto/control_plane.proto`:
 ```
 Client -> HTTP Handler -> SyncedPool -> (local?) -> Registry -> Grain State
                       \-> (remote?) -> RemoteGrainGRPC -> gRPC -> Remote CartActor -> Registry -> Grain
-ControlPlane: Discovery Events <-> Negotiation/Ping/ConfirmOwner <-> SyncedPool state
+ControlPlane: Discovery Events <-> Negotiation/Ping <-> SyncedPool state (ring determines ownership)
 ```

 ---
@@ -407,7 +437,7 @@ ControlPlane: Discovery Events <-> Negotiation/Ping/ConfirmOwner <-> SyncedPool
 |---------|--------------|--------|
 | New cart every request | Secure cookie over plain HTTP or not sending cookie jar | Disable Secure locally or use HTTPS & proper curl `-b` |
 | Unsupported mutation error | Missing registry handler | Add `RegisterMutation` for that proto |
-| Ownership flapping | Quorum failing due to intermittent peers | Investigate `ConfirmOwner` errors / network |
+| Ownership imbalance | Ring host distribution skew or rapid host churn | Examine `cart_ring_host_share`, `cart_ring_hosts`, and logs for host add/remove; rebalance or investigate instability |
 | Remote mutation latency | Network / serialization overhead | Consider batching or colocating hot carts |
 | Checkout returns 500 | Klarna call failed | Inspect logs; no grain state mutated |

--- a/cart_id.go
+++ b/cart_id.go
@@ -0,0 +1,327 @@
+package main
+
+import (
+	"crypto/rand"
+	"encoding/binary"
+	"errors"
+	"fmt"
+	"strings"
+)
+
+// cart_id.go
+//
+// Compact CartID implementation using 64 bits of cryptographic randomness,
+// base62 encoded (0-9 A-Z a-z). Typical length is 11 characters (since 62^11 > 2^64).
+//
+// Motivation:
+//   * Shorter identifiers for cookies / URLs than legacy padded 16-byte CartId
+//   * O(1) hashing (raw uint64) for consistent hashing ring integration
+//   * Extremely low collision probability (birthday bound negligible at scale)
+//
+// Backward Compatibility Strategy (Phased):
+//   Phase 1: Introduce CartID helpers while continuing to accept legacy CartId.
+//   Phase 2: Internally migrate maps to key by uint64 (CartID.Raw()).
+//   Phase 3: Canonicalize all inbound IDs to short base62; reissue Set-Cart-Id header.
+//
+// NOTE:
+//   The legacy type `CartId [16]byte` is still present elsewhere; helper
+//   UpgradeLegacyCartId bridges that representation to the new form without
+//   breaking deterministic mapping for existing carts.
+//
+// Security / Predictability:
+//   Uses crypto/rand for generation. If ever required, you can layer an
+//   HMAC-based derivation for additional secrecy. Current approach already
+//   provides 64 bits of entropy (brute force infeasible for practical risk).
+//
+// Future Extensions:
+//   * Time-sortable IDs: prepend a 48-bit timestamp field and encode 80 bits.
+//   * Add metrics counters for: generated_new, parsed_existing, legacy_fallback.
+//   * Add a pool of pre-generated IDs for ultra-low-latency hot paths (rarely needed).
+//
+// Public Surface Summary:
+//   NewCartID() (CartID, error)
+//   ParseCartID(string) (CartID, bool)
+//   FallbackFromString(string) CartID
+//   UpgradeLegacyCartId(CartId) CartID
+//   CanonicalizeIncoming(string) (CartID, bool /*wasGenerated*/, error)
+//
+// Encoding Details:
+//   encodeBase62 / decodeBase62 maintain a stable alphabet. DO NOT change
+//   alphabet order once IDs are in circulation, or previously issued IDs
+//   will change meaning.
+//
+// Zero Values:
+//   The zero value CartID{} has raw=0, txt="0". Treat it as valid but
+//   usually you will call NewCartID instead.
+//
+// ---------------------------------------------------------------------------
+
+const base62Alphabet = "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz"
+
+// Precomputed reverse lookup table for decode (255 = invalid).
+var base62Rev [256]byte
+
+func init() {
+	for i := range base62Rev {
+		base62Rev[i] = 0xFF
+	}
+	for i := 0; i < len(base62Alphabet); i++ {
+		base62Rev[base62Alphabet[i]] = byte(i)
+	}
+}
+
+// CartID is the compact representation of a cart identifier.
+// raw: 64-bit entropy (also used directly for consistent hashing).
+// txt: cached base62 textual form.
+type CartID struct {
+	raw uint64
+	txt string
+}
+
+// String returns the canonical base62 encoded ID.
+func (c CartID) String() string {
+	if c.txt == "" { // lazily encode if constructed manually
+		c.txt = encodeBase62(c.raw)
+	}
+	return c.txt
+}
+
+// Raw returns the 64-bit numeric value (useful for hashing / ring lookup).
+func (c CartID) Raw() uint64 {
+	return c.raw
+}
+
+// IsZero reports whether this CartID is the zero value.
+func (c CartID) IsZero() bool {
+	return c.raw == 0
+}
+
+// NewCartID generates a new cryptographically random 64-bit ID.
+func NewCartID() (CartID, error) {
+	var b [8]byte
+	if _, err := rand.Read(b[:]); err != nil {
+		return CartID{}, fmt.Errorf("NewCartID: %w", err)
+	}
+	u := binary.BigEndian.Uint64(b[:])
+	// Reject zero if you want to avoid ever producing "0" (optional).
+	if u == 0 {
+		// Extremely unlikely; recurse once.
+		return NewCartID()
+	}
+	return CartID{raw: u, txt: encodeBase62(u)}, nil
+}
+
+// MustNewCartID panics on failure (suitable for tests / initialization).
+func MustNewCartID() CartID {
+	id, err := NewCartID()
+	if err != nil {
+		panic(err)
+	}
+	return id
+}
+
+// ParseCartID attempts to parse a base62 canonical ID.
+// Returns (id, true) if fully valid; (zero, false) otherwise.
+func ParseCartID(s string) (CartID, bool) {
+	if len(s) == 0 {
+		return CartID{}, false
+	}
+	// Basic length sanity; allow a bit of headroom for future timestamp variant.
+	if len(s) > 16 {
+		return CartID{}, false
+	}
+	u, ok := decodeBase62(s)
+	if !ok {
+		return CartID{}, false
+	}
+	return CartID{raw: u, txt: s}, true
+}
+
+// FallbackFromString produces a deterministic CartID from arbitrary input
+// using a 64-bit FNV-1a hash. This allows legacy or malformed IDs to map
+// consistently into the new scheme (collision probability still low).
+func FallbackFromString(s string) CartID {
+	const (
+		offset64 = 1469598103934665603
+		prime64  = 1099511628211
+	)
+	h := uint64(offset64)
+	for i := 0; i < len(s); i++ {
+		h ^= uint64(s[i])
+		h *= prime64
+	}
+	return CartID{raw: h, txt: encodeBase62(h)}
+}
+
+// UpgradeLegacyCartId converts the old 16-byte CartId (padded) to CartID
+// by hashing its trimmed string form. Keeps stable mapping across restarts.
+func UpgradeLegacyCartId(old CartId) CartID {
+	return FallbackFromString(old.String())
+}
+
+// CanonicalizeIncoming normalizes user-provided ID strings.
+// Behavior:
+//
+//	Empty string  -> generate new ID (wasGenerated = true)
+//	Valid base62  -> parse and return (wasGenerated = false)
+//	Anything else -> fallback deterministic hash (wasGenerated = false)
+//
+// Errors only occur if crypto/rand fails during generation.
+func CanonicalizeIncoming(s string) (CartID, bool, error) {
+	if s == "" {
+		id, err := NewCartID()
+		return id, true, err
+	}
+	if cid, ok := ParseCartID(s); ok {
+		return cid, false, nil
+	}
+	// Legacy heuristic: if length == 16 and contains non-base62 chars, treat as legacy padded ID.
+	if len(s) == 16 && !isAllBase62(s) {
+		return FallbackFromString(strings.TrimRight(s, "\x00")), false, nil
+	}
+	return FallbackFromString(s), false, nil
+}
+
+// isAllBase62 returns true if every byte is in the base62 alphabet.
+func isAllBase62(s string) bool {
+	for i := 0; i < len(s); i++ {
+		if base62Rev[s[i]] == 0xFF {
+			return false
+		}
+	}
+	return true
+}
+
+// encodeBase62 turns a uint64 into base62 text.
+// Complexity: O(log_62 n) ~ at most 11 iterations for 64 bits.
+func encodeBase62(u uint64) string {
+	if u == 0 {
+		return "0"
+	}
+	// 62^11 = 743008370688 > 2^39; 62^11 > 2^64? Actually 62^11 ~= 5.18e19 < 2^64 (1.84e19)? 2^64 ≈ 1.84e19.
+	// 62^11 ≈ 5.18e19 > 2^64? Correction: 2^64 ≈ 1.844e19, so 62^11 > 2^64. Thus 11 chars suffice.
+	var buf [11]byte
+	i := len(buf)
+	for u > 0 {
+		i--
+		buf[i] = base62Alphabet[u%62]
+		u /= 62
+	}
+	return string(buf[i:])
+}
+
+// decodeBase62 converts a base62 string to uint64.
+// Returns (value, false) if any invalid character appears.
+func decodeBase62(s string) (uint64, bool) {
+	var v uint64
+	for i := 0; i < len(s); i++ {
+		c := s[i]
+		d := base62Rev[c]
+		if d == 0xFF {
+			return 0, false
+		}
+		v = v*62 + uint64(d)
+	}
+	return v, true
+}
+
+// ErrInvalidCartID can be returned by higher-level validation layers if you decide
+// to reject fallback-derived IDs (currently unused here).
+var ErrInvalidCartID = errors.New("invalid cart id")
+
+// ---------------------------------------------------------------------------
+// Legacy / Compatibility Conversion Helpers
+// ---------------------------------------------------------------------------
+
+// CartIDToLegacy converts a CartID (base62) into the legacy fixed-size CartId
+// ([16]byte) by copying the textual form (truncated or zero-padded).
+// NOTE: If the base62 string is longer than 16 (should not happen with current
+// 64-bit space), it will be truncated.
+func CartIDToLegacy(c CartID) CartId {
+	var id CartId
+	txt := c.String()
+	copy(id[:], []byte(txt))
+	return id
+}
+
+// LegacyToCartID upgrades a legacy CartId (padded) to a CartID by hashing its
+// trimmed string form (deterministic). This preserves stable mapping without
+// depending on original randomness.
+func LegacyToCartID(old CartId) CartID {
+	return UpgradeLegacyCartId(old)
+}
+
+// CartIDToKey returns the numeric key representation (uint64) for map indexing.
+func CartIDToKey(c CartID) uint64 {
+	return c.Raw()
+}
+
+// LegacyToCartKey converts a legacy CartId to the numeric key via deterministic
+// fallback hashing. (Uses the same logic as LegacyToCartID then returns raw.)
+func LegacyToCartKey(old CartId) uint64 {
+	return LegacyToCartID(old).Raw()
+}
+
+// ---------------------- Optional Helper Utilities ----------------------------
+
+// CartIDOrNew tries to parse s; if empty OR invalid returns a fresh ID.
+func CartIDOrNew(s string) (CartID, bool /*wasParsed*/, error) {
+	if cid, ok := ParseCartID(s); ok {
+		return cid, true, nil
+	}
+	id, err := NewCartID()
+	return id, false, err
+}
+
+// MustParseCartID panics if s is not a valid base62 ID (useful in tests).
+func MustParseCartID(s string) CartID {
+	if cid, ok := ParseCartID(s); ok {
+		return cid
+	}
+	panic(fmt.Sprintf("invalid CartID: %s", s))
+}
+
+// DebugString returns a verbose description (for logging / diagnostics).
+func (c CartID) DebugString() string {
+	return fmt.Sprintf("CartID(raw=%d txt=%s)", c.raw, c.String())
+}
+
+// Equal compares two CartIDs by raw value.
+func (c CartID) Equal(other CartID) bool {
+	return c.raw == other.raw
+}
+
+// CanonicalizeOrLegacy preserves legacy (non-base62) IDs without altering their
+// textual form, avoiding the previous behavior where fallback hashing replaced
+// the original string with a base62-encoded hash (which broke deterministic
+// key derivation across mixed call paths).
+//
+// Behavior:
+//   - s == ""      -> generate new CartID (generatedNew = true, wasBase62 = true)
+//   - base62 ok    -> return parsed CartID (generatedNew = false, wasBase62 = true)
+//   - otherwise    -> treat as legacy: raw = hash(s), txt = original s
+//
+// Returns:
+//
+//	cid          - CartID (txt preserved for legacy inputs)
+//	generatedNew - true only when a brand new ID was created due to empty input
+//	wasBase62    - true if the input was already canonical base62 (or generated)
+//	err          - only set if crypto/rand fails when generating a new ID
+func CanonicalizeOrLegacy(s string) (cid CartID, generatedNew bool, wasBase62 bool, err error) {
+	if s == "" {
+		id, e := NewCartID()
+		if e != nil {
+			return CartID{}, false, false, e
+		}
+		return id, true, true, nil
+	}
+	if parsed, ok := ParseCartID(s); ok {
+		return parsed, false, true, nil
+	}
+	// Legacy path: keep original text so downstream legacy-to-key hashing
+	// (which uses the visible string) yields consistent keys across code paths.
+	hashCID := FallbackFromString(s)
+	// Preserve original textual form
+	hashCID.txt = s
+	return hashCID, false, false, nil
+}
--- a/cart_id_test.go
+++ b/cart_id_test.go
@@ -0,0 +1,259 @@
+package main
+
+import (
+	"crypto/rand"
+	"encoding/binary"
+	"fmt"
+	mrand "math/rand"
+	"testing"
+)
+
+// TestEncodeDecodeBase62RoundTrip verifies encodeBase62/decodeBase62 are inverse.
+func TestEncodeDecodeBase62RoundTrip(t *testing.T) {
+	mrand.Seed(42)
+	for i := 0; i < 1000; i++ {
+		// Random 64-bit value
+		v := mrand.Uint64()
+		s := encodeBase62(v)
+		dec, ok := decodeBase62(s)
+		if !ok {
+			t.Fatalf("decodeBase62 failed for %d encoded=%s", v, s)
+		}
+		if dec != v {
+			t.Fatalf("round trip mismatch: have %d got %d (encoded=%s)", v, dec, s)
+		}
+	}
+	// Explicit zero test
+	if s := encodeBase62(0); s != "0" {
+		t.Fatalf("expected encodeBase62(0) == \"0\", got %q", s)
+	}
+	if v, ok := decodeBase62("0"); !ok || v != 0 {
+		t.Fatalf("decodeBase62(0) unexpected result v=%d ok=%v", v, ok)
+	}
+}
+
+// TestNewCartIDUniqueness generates a number of IDs and checks for duplicates.
+func TestNewCartIDUniqueness(t *testing.T) {
+	const n = 10000
+	seen := make(map[string]struct{}, n)
+	for i := 0; i < n; i++ {
+		id, err := NewCartID()
+		if err != nil {
+			t.Fatalf("NewCartID error: %v", err)
+		}
+		s := id.String()
+		if _, exists := seen[s]; exists {
+			t.Fatalf("duplicate CartID generated: %s", s)
+		}
+		seen[s] = struct{}{}
+		if id.IsZero() {
+			t.Fatalf("NewCartID returned zero value")
+		}
+	}
+}
+
+// TestParseCartIDValidation tests parsing of valid and invalid base62 strings.
+func TestParseCartIDValidation(t *testing.T) {
+	id, err := NewCartID()
+	if err != nil {
+		t.Fatalf("NewCartID error: %v", err)
+	}
+	parsed, ok := ParseCartID(id.String())
+	if !ok {
+		t.Fatalf("ParseCartID failed for valid id %s", id)
+	}
+	if parsed.raw != id.raw {
+		t.Fatalf("parsed raw mismatch: %d vs %d", parsed.raw, id.raw)
+	}
+
+	if _, ok := ParseCartID(""); ok {
+		t.Fatalf("expected empty string to be invalid")
+	}
+	// Invalid char ('-')
+	if _, ok := ParseCartID("abc-123"); ok {
+		t.Fatalf("expected invalid chars to fail parse")
+	}
+	// Overly long ( >16 )
+	if _, ok := ParseCartID("1234567890abcdefg"); ok {
+		t.Fatalf("expected overly long string to fail parse")
+	}
+}
+
+// TestFallbackDeterminism ensures fallback hashing is deterministic.
+func TestFallbackDeterminism(t *testing.T) {
+	inputs := []string{
+		"legacy-cart-1",
+		"legacy-cart-2",
+		"UPPER_lower_123",
+		"🚀unicode", // unicode bytes (will hash byte sequence)
+	}
+	for _, in := range inputs {
+		a := FallbackFromString(in)
+		b := FallbackFromString(in)
+		if a.raw != b.raw || a.String() != b.String() {
+			t.Fatalf("fallback mismatch for %q: %+v vs %+v", in, a, b)
+		}
+	}
+	// Distinct inputs should almost always differ; sample check
+	a := FallbackFromString("distinct-A")
+	b := FallbackFromString("distinct-B")
+	if a.raw == b.raw {
+		t.Fatalf("unexpected identical fallback hashes for distinct inputs")
+	}
+}
+
+// TestCanonicalizeIncomingBehavior covers main control flow branches.
+func TestCanonicalizeIncomingBehavior(t *testing.T) {
+	// Empty => new id
+	id1, generated, err := CanonicalizeIncoming("")
+	if err != nil || !generated || id1.IsZero() {
+		t.Fatalf("CanonicalizeIncoming empty failed: id=%v gen=%v err=%v", id1, generated, err)
+	}
+
+	// Valid base62 => parse; no generation
+	id2, gen2, err := CanonicalizeIncoming(id1.String())
+	if err != nil || gen2 || id2.raw != id1.raw {
+		t.Fatalf("CanonicalizeIncoming parse mismatch: id2=%v gen2=%v err=%v", id2, gen2, err)
+	}
+
+	// Legacy-like random containing invalid chars -> fallback
+	fallbackInput := "legacy\x00\x00padding"
+	id3, gen3, err := CanonicalizeIncoming(fallbackInput)
+	if err != nil || gen3 {
+		t.Fatalf("CanonicalizeIncoming fallback unexpected: id3=%v gen3=%v err=%v", id3, gen3, err)
+	}
+
+	// Deterministic fallback
+	id4, _, _ := CanonicalizeIncoming(fallbackInput)
+	if id3.raw != id4.raw {
+		t.Fatalf("fallback canonicalization not deterministic")
+	}
+}
+
+// TestUpgradeLegacyCartId ensures mapping of old CartId is stable.
+func TestUpgradeLegacyCartId(t *testing.T) {
+	var legacy CartId
+	copy(legacy[:], []byte("legacy-123456789")) // 15 bytes + padding
+	up1 := UpgradeLegacyCartId(legacy)
+	up2 := UpgradeLegacyCartId(legacy)
+	if up1.raw != up2.raw {
+		t.Fatalf("UpgradeLegacyCartId not deterministic: %v vs %v", up1, up2)
+	}
+	if up1.String() != up2.String() {
+		t.Fatalf("UpgradeLegacyCartId string mismatch: %s vs %s", up1, up2)
+	}
+}
+
+// BenchmarkNewCartID gives a rough idea of generation cost.
+func BenchmarkNewCartID(b *testing.B) {
+	for i := 0; i < b.N; i++ {
+		if _, err := NewCartID(); err != nil {
+			b.Fatalf("error: %v", err)
+		}
+	}
+}
+
+// BenchmarkEncodeBase62 measures encode speed in isolation.
+func BenchmarkEncodeBase62(b *testing.B) {
+	// Random sample of values
+	samples := make([]uint64, 1024)
+	for i := range samples {
+		var buf [8]byte
+		if _, err := rand.Read(buf[:]); err != nil {
+			b.Fatalf("rand: %v", err)
+		}
+		samples[i] = binary.BigEndian.Uint64(buf[:])
+	}
+	b.ResetTimer()
+	var sink string
+	for i := 0; i < b.N; i++ {
+		sink = encodeBase62(samples[i%len(samples)])
+	}
+	_ = sink
+}
+
+// BenchmarkDecodeBase62 measures decode speed.
+func BenchmarkDecodeBase62(b *testing.B) {
+	// Pre-encode
+	encoded := make([]string, 1024)
+	for i := range encoded {
+		encoded[i] = encodeBase62(uint64(i)<<32 | uint64(i))
+	}
+	b.ResetTimer()
+	var sum uint64
+	for i := 0; i < b.N; i++ {
+		v, ok := decodeBase62(encoded[i%len(encoded)])
+		if !ok {
+			b.Fatalf("decode failed")
+		}
+		sum ^= v
+	}
+	_ = sum
+}
+
+// TestLookupNDeterminism (ring integration smoke test) ensures LookupN
+// returns distinct hosts and stable ordering for a fixed ring.
+func TestLookupNDeterminism(t *testing.T) {
+	rb := NewRingBuilder().WithEpoch(1).WithVnodesPerHost(8).WithHosts([]string{"a", "b", "c"})
+	ring := rb.Build()
+	if ring.Empty() {
+		t.Fatalf("expected non-empty ring")
+	}
+	id := MustNewCartID()
+	owners1 := ring.LookupN(id.Raw(), 3)
+	owners2 := ring.LookupN(id.Raw(), 3)
+	if len(owners1) != len(owners2) {
+		t.Fatalf("LookupN length mismatch")
+	}
+	for i := range owners1 {
+		if owners1[i].Host != owners2[i].Host {
+			t.Fatalf("LookupN ordering instability at %d: %v vs %v", i, owners1[i], owners2[i])
+		}
+	}
+	// Distinct host constraint
+	seen := map[string]struct{}{}
+	for _, v := range owners1 {
+		if _, ok := seen[v.Host]; ok {
+			t.Fatalf("duplicate host in LookupN result: %v", owners1)
+		}
+		seen[v.Host] = struct{}{}
+	}
+}
+
+// TestRingFingerprintChanges ensures fingerprint updates with membership changes.
+func TestRingFingerprintChanges(t *testing.T) {
+	b1 := NewRingBuilder().WithEpoch(1).WithHosts([]string{"node1", "node2"})
+	r1 := b1.Build()
+	b2 := NewRingBuilder().WithEpoch(2).WithHosts([]string{"node1", "node2", "node3"})
+	r2 := b2.Build()
+	if r1.Fingerprint() == r2.Fingerprint() {
+		t.Fatalf("expected differing fingerprints after host set change")
+	}
+}
+
+// TestRingDiffHosts verifies added/removed host detection.
+func TestRingDiffHosts(t *testing.T) {
+	r1 := NewRingBuilder().WithEpoch(1).WithHosts([]string{"a", "b"}).Build()
+	r2 := NewRingBuilder().WithEpoch(2).WithHosts([]string{"b", "c"}).Build()
+	added, removed := r1.DiffHosts(r2)
+	if fmt.Sprintf("%v", added) != "[c]" {
+		t.Fatalf("expected added [c], got %v", added)
+	}
+	if fmt.Sprintf("%v", removed) != "[a]" {
+		t.Fatalf("expected removed [a], got %v", removed)
+	}
+}
+
+// TestRingLookupConsistency ensures direct Lookup and LookupID are aligned.
+func TestRingLookupConsistency(t *testing.T) {
+	ring := NewRingBuilder().WithEpoch(1).WithHosts([]string{"alpha", "beta"}).WithVnodesPerHost(4).Build()
+	id, _ := ParseCartID("1")
+	if id.IsZero() {
+		t.Fatalf("expected parsed id non-zero")
+	}
+	v1 := ring.Lookup(id.Raw())
+	v2 := ring.LookupID(id)
+	if v1.Host != v2.Host || v1.Hash != v2.Hash {
+		t.Fatalf("Lookup vs LookupID mismatch: %+v vs %+v", v1, v2)
+	}
+}
--- a/grain-pool.go
+++ b/grain-pool.go
@@ -10,6 +10,23 @@ import (
 	"github.com/prometheus/client_golang/prometheus/promauto"
 )

+// grain-pool.go
+//
+// Migration Note:
+//   This file has been migrated to use uint64 cart keys internally (derived
+//   from the new CartID base62 representation). For backward compatibility,
+//   a deprecated legacy map keyed by CartId is maintained so existing code
+//   that directly indexes pool.grains with a CartId continues to compile
+//   until the full refactor across SyncedPool / remoteIndex is completed.
+//
+//   Authoritative storage: grains (map[uint64]*CartGrain)
+//   Legacy compatibility:  grainsLegacy (map[CartId]*CartGrain) - kept in sync.
+//
+//   Once all external usages are updated to rely on helper accessors,
+//   grainsLegacy can be removed.
+//
+// ---------------------------------------------------------------------------
+
 var (
 	poolGrains = promauto.NewGauge(prometheus.GaugeOpts{
 		Name: "cart_grains_in_pool",
@@ -25,49 +42,71 @@ var (
 	})
 )

+// GrainPool interface remains legacy-compatible.
 type GrainPool interface {
 	Apply(id CartId, mutation interface{}) (*CartGrain, error)
 	Get(id CartId) (*CartGrain, error)
 }

+// Ttl keeps expiry info
 type Ttl struct {
 	Expires time.Time
 	Grain   *CartGrain
 }

+// GrainLocalPool now stores grains keyed by uint64 (CartKey).
 type GrainLocalPool struct {
 	mu       sync.RWMutex
-	grains   map[CartId]*CartGrain
+	grains   map[uint64]*CartGrain // authoritative only
 	expiry   []Ttl
 	spawn    func(id CartId) (*CartGrain, error)
 	Ttl      time.Duration
 	PoolSize int
 }

+// NewGrainLocalPool constructs a new pool.
 func NewGrainLocalPool(size int, ttl time.Duration, spawn func(id CartId) (*CartGrain, error)) *GrainLocalPool {
-
 	ret := &GrainLocalPool{
 		spawn:    spawn,
-		grains:   make(map[CartId]*CartGrain),
+		grains:   make(map[uint64]*CartGrain),
 		expiry:   make([]Ttl, 0),
 		Ttl:      ttl,
 		PoolSize: size,
 	}
-
 	cartPurge := time.NewTicker(time.Minute)
 	go func() {
-		<-cartPurge.C
+		for range cartPurge.C {
 			ret.Purge()
+		}
 	}()
 	return ret
 }

+// keyFromCartId derives the uint64 key from a legacy CartId deterministically.
+func keyFromCartId(id CartId) uint64 {
+	return LegacyToCartKey(id)
+}
+
+// storeGrain indexes a grain in both maps.
+func (p *GrainLocalPool) storeGrain(id CartId, g *CartGrain) {
+	k := keyFromCartId(id)
+	p.grains[k] = g
+}
+
+// deleteGrain removes a grain from both maps.
+func (p *GrainLocalPool) deleteGrain(id CartId) {
+	k := keyFromCartId(id)
+	delete(p.grains, k)
+}
+
+// SetAvailable pre-populates placeholder entries (legacy signature).
 func (p *GrainLocalPool) SetAvailable(availableWithLastChangeUnix map[CartId]int64) {
 	p.mu.Lock()
 	defer p.mu.Unlock()
 	for id := range availableWithLastChangeUnix {
-		if _, ok := p.grains[id]; !ok {
-			p.grains[id] = nil
+		k := keyFromCartId(id)
+		if _, ok := p.grains[k]; !ok {
+			p.grains[k] = nil
 			p.expiry = append(p.expiry, Ttl{
 				Expires: time.Now().Add(p.Ttl),
 				Grain:   nil,
@@ -76,13 +115,19 @@ func (p *GrainLocalPool) SetAvailable(availableWithLastChangeUnix map[CartId]int
 	}
 }

+// Purge removes expired grains.
 func (p *GrainLocalPool) Purge() {
 	lastChangeTime := time.Now().Add(-p.Ttl)
 	keepChanged := lastChangeTime.Unix()
+
 	p.mu.Lock()
 	defer p.mu.Unlock()
+
 	for i := 0; i < len(p.expiry); i++ {
 		item := p.expiry[i]
+		if item.Grain == nil {
+			continue
+		}
 		if item.Expires.Before(time.Now()) {
 			if item.Grain.GetLastChange() > keepChanged {
 				log.Printf("Expired item %s changed, keeping", item.Grain.GetId())
@@ -90,12 +135,12 @@ func (p *GrainLocalPool) Purge() {
 					p.expiry = append(p.expiry[:i], p.expiry[i+1:]...)
 					p.expiry = append(p.expiry, item)
 				} else {
+					// move last to end (noop)
 					p.expiry = append(p.expiry[:i], item)
 				}
-
 			} else {
 				log.Printf("Item %s expired", item.Grain.GetId())
-				delete(p.grains, item.Grain.GetId())
+				p.deleteGrain(item.Grain.GetId())
 				if i < len(p.expiry)-1 {
 					p.expiry = append(p.expiry[:i], p.expiry[i+1:]...)
 				} else {
@@ -108,40 +153,69 @@ func (p *GrainLocalPool) Purge() {
 	}
 }

+// GetGrains returns a legacy view of grains (copy) for compatibility.
 func (p *GrainLocalPool) GetGrains() map[CartId]*CartGrain {
-	return p.grains
+	p.mu.RLock()
+	defer p.mu.RUnlock()
+	out := make(map[CartId]*CartGrain, len(p.grains))
+	for _, g := range p.grains {
+		if g != nil {
+			out[g.GetId()] = g
+		}
+	}
+	return out
 }

-func (p *GrainLocalPool) GetGrain(id CartId) (*CartGrain, error) {
-	var err error
-	// p.mu.RLock()
-	// defer p.mu.RUnlock()
-	grain, ok := p.grains[id]
-	grainLookups.Inc()
-	if grain == nil || !ok {
-		if len(p.grains) >= p.PoolSize {
-			if p.expiry[0].Expires.Before(time.Now()) {
-				delete(p.grains, p.expiry[0].Grain.GetId())
-				p.expiry = p.expiry[1:]
-			} else {
-				return nil, fmt.Errorf("pool is full")
-			}
-		}
-		grain, err = p.spawn(id)
-		p.mu.Lock()
-		p.grains[id] = grain
-		p.mu.Unlock()
-	}
-	go func() {
-		l := float64(len(p.grains))
+// statsUpdate updates Prometheus gauges asynchronously.
+func (p *GrainLocalPool) statsUpdate() {
+	go func(size int) {
+		l := float64(size)
 		ps := float64(p.PoolSize)
 		poolUsage.Set(l / ps)
 		poolGrains.Set(l)
 		poolSize.Set(ps)
-	}()
+	}(len(p.grains))
+}
+
+// GetGrain retrieves or spawns a grain (legacy id signature).
+func (p *GrainLocalPool) GetGrain(id CartId) (*CartGrain, error) {
+	grainLookups.Inc()
+	k := keyFromCartId(id)
+
+	p.mu.RLock()
+	grain, ok := p.grains[k]
+	p.mu.RUnlock()
+
+	var err error
+	if grain == nil || !ok {
+		p.mu.Lock()
+		// Re-check under write lock
+		grain, ok = p.grains[k]
+		if grain == nil || !ok {
+			// Capacity check
+			if len(p.grains) >= p.PoolSize && len(p.expiry) > 0 {
+				if p.expiry[0].Expires.Before(time.Now()) && p.expiry[0].Grain != nil {
+					oldId := p.expiry[0].Grain.GetId()
+					p.deleteGrain(oldId)
+					p.expiry = p.expiry[1:]
+				} else {
+					p.mu.Unlock()
+					return nil, fmt.Errorf("pool is full")
+				}
+			}
+			grain, err = p.spawn(id)
+			if err == nil {
+				p.storeGrain(id, grain)
+			}
+		}
+		p.mu.Unlock()
+		p.statsUpdate()
+	}
+
 	return grain, err
 }

+// Apply applies a mutation (legacy compatibility).
 func (p *GrainLocalPool) Apply(id CartId, mutation interface{}) (*CartGrain, error) {
 	grain, err := p.GetGrain(id)
 	if err != nil || grain == nil {
@@ -150,6 +224,21 @@ func (p *GrainLocalPool) Apply(id CartId, mutation interface{}) (*CartGrain, err
 	return grain.Apply(mutation, false)
 }

+// Get returns current state (legacy wrapper).
 func (p *GrainLocalPool) Get(id CartId) (*CartGrain, error) {
 	return p.GetGrain(id)
 }
+
+// DebugGrainCount returns counts for debugging.
+func (p *GrainLocalPool) DebugGrainCount() (authoritative int) {
+	p.mu.RLock()
+	defer p.mu.RUnlock()
+	return len(p.grains)
+}
+
+// UnsafePointerToLegacyMap exposes the legacy map pointer (for transitional
+// tests that still poke the field directly). DO NOT rely on this long-term.
+func (p *GrainLocalPool) UnsafePointerToLegacyMap() uintptr {
+	// Legacy map removed; retained only to satisfy any transitional callers.
+	return 0
+}
--- a/grpc_server.go
+++ b/grpc_server.go
@@ -32,7 +32,18 @@ func NewCartActorGRPCServer(pool GrainPool, syncedPool *SyncedPool) *cartActorGR

 // applyMutation routes a single cart mutation to the target grain (used by per-mutation RPC handlers).
 func (s *cartActorGRPCServer) applyMutation(cartID string, mutation interface{}) *messages.CartMutationReply {
-	grain, err := s.pool.Apply(ToCartId(cartID), mutation)
+	// Canonicalize or preserve legacy id (do NOT hash-rewrite legacy textual ids)
+	cid, _, wasBase62, cerr := CanonicalizeOrLegacy(cartID)
+	if cerr != nil {
+		return &messages.CartMutationReply{
+			StatusCode:      500,
+			Result:          &messages.CartMutationReply_Error{Error: fmt.Sprintf("cart_id canonicalization failed: %v", cerr)},
+			ServerTimestamp: time.Now().Unix(),
+		}
+	}
+	_ = wasBase62 // placeholder; future: propagate canonical id in reply metadata
+	legacy := CartIDToLegacy(cid)
+	grain, err := s.pool.Apply(legacy, mutation)
 	if err != nil {
 		return &messages.CartMutationReply{
 			StatusCode:      500,
@@ -159,9 +170,17 @@ func (s *cartActorGRPCServer) GetState(ctx context.Context, req *messages.StateR
 			Result:     &messages.StateReply_Error{Error: "cart_id is required"},
 		}, nil
 	}
-	cartID := ToCartId(req.GetCartId())
+	// Canonicalize / upgrade incoming cart id (preserve legacy strings)
+	cid, _, _, cerr := CanonicalizeOrLegacy(req.GetCartId())
+	if cerr != nil {
+		return &messages.StateReply{
+			StatusCode: 500,
+			Result:     &messages.StateReply_Error{Error: fmt.Sprintf("cart_id canonicalization failed: %v", cerr)},
+		}, nil
+	}
+	legacy := CartIDToLegacy(cid)

-	grain, err := s.pool.Get(cartID)
+	grain, err := s.pool.Get(legacy)
 	if err != nil {
 		return &messages.StateReply{
 			StatusCode: 500,
@@ -212,47 +231,18 @@ func (s *cartActorGRPCServer) Negotiate(ctx context.Context, req *messages.Negot

 // ControlPlane: GetCartIds (locally owned carts only)
 func (s *cartActorGRPCServer) GetCartIds(ctx context.Context, _ *messages.Empty) (*messages.CartIdsReply, error) {
-	ids := make([]string, 0, len(s.syncedPool.local.grains))
 	s.syncedPool.local.mu.RLock()
-	for id, g := range s.syncedPool.local.grains {
-		if g != nil {
-			ids = append(ids, id.String())
+	ids := make([]string, 0, len(s.syncedPool.local.grains))
+	for _, g := range s.syncedPool.local.grains {
+		if g == nil {
+			continue
 		}
+		ids = append(ids, g.GetId().String())
 	}
 	s.syncedPool.local.mu.RUnlock()
 	return &messages.CartIdsReply{CartIds: ids}, nil
 }

-// ControlPlane: ConfirmOwner (simple always-accept implementation)
-// Future enhancement: add fencing / versioning & validate current holder.
-func (s *cartActorGRPCServer) ConfirmOwner(ctx context.Context, req *messages.OwnerChangeRequest) (*messages.OwnerChangeAck, error) {
-	if req.GetCartId() == "" || req.GetNewHost() == "" {
-		return &messages.OwnerChangeAck{
-			Accepted: false,
-			Message:  "cart_id and new_host required",
-		}, nil
-	}
-	// If we are *not* the new host and currently have a local grain, we:
-	//   1. Drop any local grain (relinquish ownership)
-	//   2. Spawn (or refresh) a remote proxy pointing to the new owner so
-	//      subsequent mutations from this node route correctly.
-	if req.GetNewHost() != s.syncedPool.Hostname {
-		cid := ToCartId(req.GetCartId())
-		// Drop local ownership if present.
-		s.syncedPool.local.mu.Lock()
-		delete(s.syncedPool.local.grains, cid)
-		s.syncedPool.local.mu.Unlock()
-
-		// Ensure a remote proxy exists for the new owner. SpawnRemoteGrain will
-		// no-op if host unknown and attempt AddRemote asynchronously.
-		s.syncedPool.SpawnRemoteGrain(cid, req.GetNewHost())
-	}
-	return &messages.OwnerChangeAck{
-		Accepted: true,
-		Message:  "accepted",
-	}, nil
-}
-
 // ControlPlane: Closing (peer shutdown notification)
 func (s *cartActorGRPCServer) Closing(ctx context.Context, req *messages.ClosingNotice) (*messages.OwnerChangeAck, error) {
 	if req.GetHost() != "" {
--- a/multi_node_ownership_test.go
+++ b/multi_node_ownership_test.go
@@ -81,13 +81,17 @@ func TestMultiNodeOwnershipNegotiation(t *testing.T) {
 	link(syncedA, hostB, addrB)
 	link(syncedB, hostA, addrA)

+	// Rebuild rings after manual cross-link so deterministic ownership works immediately.
+	syncedA.ForceRingRefresh()
+	syncedB.ForceRingRefresh()
+
 	// Allow brief stabilization (control plane pings / no real negotiation needed here).
 	time.Sleep(200 * time.Millisecond)

 	// Create a deterministic cart id for test readability.
 	cartID := ToCartId(fmt.Sprintf("cart-%d", time.Now().UnixNano()))

-	// Mutation payload (local ownership claim expected on nodeA).
+	// Mutation payload (ring-determined ownership; no assumption about which node owns).
 	addItem := &messages.AddItem{
 		ItemId:   1,
 		Quantity: 1,
@@ -101,32 +105,54 @@ func TestMultiNodeOwnershipNegotiation(t *testing.T) {
 		Country:  "se",
 	}

-	// Apply mutation on nodeA (should create local grain + claim ownership).
-	if _, err := syncedA.Apply(cartID, addItem); err != nil {
-		t.Fatalf("nodeA Apply addItem error: %v", err)
+	// Determine ring owner and set primary / secondary references.
+	ownerHost := syncedA.DebugOwnerHost(cartID)
+	var ownerSynced, otherSynced *SyncedPool
+	var ownerPool, otherPool *GrainLocalPool
+	switch ownerHost {
+	case hostA:
+		ownerSynced, ownerPool = syncedA, poolA
+		otherSynced, otherPool = syncedB, poolB
+	case hostB:
+		ownerSynced, ownerPool = syncedB, poolB
+		otherSynced, otherPool = syncedA, poolA
+	default:
+		t.Fatalf("unexpected ring owner %s (expected %s or %s)", ownerHost, hostA, hostB)
 	}

-	// Validate nodeA local pool has the grain.
-	if _, ok := poolA.grains[cartID]; !ok {
-		t.Fatalf("nodeA expected local grain ownership but grain missing")
+	// Apply mutation on the ring-designated owner.
+	if _, err := ownerSynced.Apply(cartID, addItem); err != nil {
+		t.Fatalf("owner %s Apply addItem error: %v", ownerHost, err)
 	}

-	// Attempt to mutate same cart from nodeB (should route remotely, not create local duplication).
+	// Validate owner pool has the grain and the other does not.
+	if _, ok := ownerPool.GetGrains()[cartID]; !ok {
+		t.Fatalf("expected owner %s to have local grain", ownerHost)
+	}
+	if _, ok := otherPool.GetGrains()[cartID]; ok {
+		t.Fatalf("non-owner unexpectedly holds local grain")
+	}
+
+	// Prepare change mutation to be applied from the non-owner (should route remotely).
 	change := &messages.ChangeQuantity{
-		Id:       1, // line id is 1 after first AddItem
+		Id:       1, // line id after first AddItem
 		Quantity: 2,
 	}
-
-	// Apply on nodeB; if ownership logic works, this will call remote RPC and succeed without creating a local grain.
-	if _, err := syncedB.Apply(cartID, change); err != nil {
-		t.Fatalf("nodeB remote Apply changeQuantity error: %v", err)
+	// Apply remotely via the non-owner.
+	if _, err := otherSynced.Apply(cartID, change); err != nil {
+		t.Fatalf("non-owner remote Apply changeQuantity error: %v", err)
 	}

-	// NodeB should NOT have a local grain (ownership), but may or may not have a remote proxy
-	// entry in remoteIndex depending on internal propagation. We assert it does NOT hold local.
-	if _, local := poolB.grains[cartID]; local {
+	// Remote re-mutation already performed via otherSynced; removed duplicate block.
+
+	// NodeB local grain assertion:
+	// Only assert absence if nodeB is NOT the ring-designated owner. If nodeB is the owner,
+	// it is expected to have a local grain (previous generic ownership assertions already ran).
+	if ownerHost != hostB {
+		if _, local := poolB.GetGrains()[cartID]; local {
 			t.Fatalf("nodeB unexpectedly created local grain (ownership duplication)")
 		}
+	}

 	// Fetch state from nodeB to ensure we see updated quantity (2).
 	grainStateB, err := syncedB.Get(cartID)
--- a/multi_node_three_test.go
+++ b/multi_node_three_test.go
@@ -10,21 +10,21 @@ import (
 	"google.golang.org/grpc"
 )

-// TestThreeNodeMajorityOwnership exercises the revised majority quorum semantics
-// with a 3-node cluster (A,B,C). After the quorum refactor, a 3-node cluster
-// (all=2 remotes) now requires only floor((all+1)/2) = 1 remote acceptance
-// instead of unanimity. Since our current ConfirmOwner implementation always
-// accepts, we mainly validate:
+// TestThreeNodeMajorityOwnership validates ring-determined ownership and routing
+// in a 3-node cluster (A,B,C) using the consistent hashing ring (no quorum RPC).
+// The previous ConfirmOwner / quorum semantics have been removed; ownership is
+// deterministic and derived from the ring.
 //
-//  1. Ownership is established on the first node that mutates (nodeA).
+// It validates:
+//  1. The ring selects exactly one primary owner for a new cart.
 //  2. Other nodes (B,C) do NOT create local grains for the cart.
-//  3. Remote proxies are installed on B and C (so they can route mutations).
-//  4. A remote mutation from nodeB updates state visible from nodeC.
+//  3. Remote proxies are installed lazily so remote mutations can route.
+//  4. A remote mutation from one non-owner updates state visible on another.
+//  5. Authoritative state on the owner matches remote observations.
+//  6. (Future) This scaffolds replication tests when RF>1 is enabled.
 //
-// NOTE: ConfirmOwner currently always accepts, so we cannot directly observe
-// a reduced acceptance threshold here without introducing a test hook that
-// can force a rejection. This test still validates that multi-node routing
-// works under the new quorum rule for N=3 (where previously unanimity was required).
+// (Legacy comments about ConfirmOwner acceptance thresholds have been removed.)
+// (Function name retained for historical continuity.)
 func TestThreeNodeMajorityOwnership(t *testing.T) {
 	const (
 		addrA = "127.0.0.1:18181"
@@ -102,6 +102,11 @@ func TestThreeNodeMajorityOwnership(t *testing.T) {
 	link(syncedC, hostA, addrA)
 	link(syncedC, hostB, addrB)

+	// Rebuild rings after manual linking so ownership resolution is immediate.
+	syncedA.ForceRingRefresh()
+	syncedB.ForceRingRefresh()
+	syncedC.ForceRingRefresh()
+
 	// Allow brief stabilization
 	time.Sleep(200 * time.Millisecond)

@@ -121,62 +126,71 @@ func TestThreeNodeMajorityOwnership(t *testing.T) {
 		Country:  "se",
 	}

-	// Apply on nodeA (ownership should establish here)
-	if _, err := syncedA.Apply(cartID, addItem); err != nil {
-		t.Fatalf("nodeA Apply addItem error: %v", err)
+	// Determine ring-designated owner (may be any of the three hosts)
+	ownerPre := syncedA.DebugOwnerHost(cartID)
+	if ownerPre != hostA && ownerPre != hostB && ownerPre != hostC {
+		t.Fatalf("ring returned unexpected owner %s (not in set {%s,%s,%s})", ownerPre, hostA, hostB, hostC)
+	}
+	var ownerSynced *SyncedPool
+	var ownerPool *GrainLocalPool
+	switch ownerPre {
+	case hostA:
+		ownerSynced, ownerPool = syncedA, poolA
+	case hostB:
+		ownerSynced, ownerPool = syncedB, poolB
+	case hostC:
+		ownerSynced, ownerPool = syncedC, poolC
+	}
+	// Pick two distinct non-owner nodes for remote mutation assertions
+	var remote1Synced, remote2Synced *SyncedPool
+	switch ownerPre {
+	case hostA:
+		remote1Synced, remote2Synced = syncedB, syncedC
+	case hostB:
+		remote1Synced, remote2Synced = syncedA, syncedC
+	case hostC:
+		remote1Synced, remote2Synced = syncedA, syncedB
 	}

-	// Small wait for ConfirmOwner RPC propagation & remote proxy spawn
+	// Apply on the ring-designated owner
+	if _, err := ownerSynced.Apply(cartID, addItem); err != nil {
+		t.Fatalf("owner %s Apply addItem error: %v", ownerPre, err)
+	}
+
+	// Small wait for remote proxy spawn (ring ownership already deterministic)
 	time.Sleep(150 * time.Millisecond)

 	// Assert only nodeA has local grain
 	localCount := 0
-	if _, ok := poolA.grains[cartID]; ok {
+	if _, ok := poolA.GetGrains()[cartID]; ok {
 		localCount++
 	}
-	if _, ok := poolB.grains[cartID]; ok {
+	if _, ok := poolB.GetGrains()[cartID]; ok {
 		localCount++
 	}
-	if _, ok := poolC.grains[cartID]; ok {
+	if _, ok := poolC.GetGrains()[cartID]; ok {
 		localCount++
 	}
 	if localCount != 1 {
 		t.Fatalf("expected exactly 1 local grain, got %d", localCount)
 	}
-	if _, ok := poolA.grains[cartID]; !ok {
-		t.Fatalf("expected nodeA to own cart locally")
+	if _, ok := ownerPool.GetGrains()[cartID]; !ok {
+		t.Fatalf("expected owner %s to hold local grain", ownerPre)
 	}

-	// Verify nodeB and nodeC have remote proxies (best-effort; if not present yet, wait briefly)
-	waitForRemote := func(sp *SyncedPool, label string) {
-		deadline := time.Now().Add(500 * time.Millisecond)
-		for {
-			sp.mu.RLock()
-			_, remoteOk := sp.remoteIndex[cartID]
-			sp.mu.RUnlock()
-			if remoteOk {
-				return
-			}
-			if time.Now().After(deadline) {
-				t.Fatalf("%s expected remote proxy for cart not found (timeout)", label)
-			}
-			time.Sleep(25 * time.Millisecond)
-		}
-	}
-	waitForRemote(syncedB, "nodeB")
-	waitForRemote(syncedC, "nodeC")
+	// Remote proxies may not pre-exist; first remote mutation will trigger SpawnRemoteGrain lazily.

-	// Issue remote mutation from nodeB -> ChangeQuantity (increase)
+	// Issue remote mutation from one non-owner -> ChangeQuantity (increase)
 	change := &messages.ChangeQuantity{
 		Id:       1,
 		Quantity: 3,
 	}
-	if _, err := syncedB.Apply(cartID, change); err != nil {
-		t.Fatalf("nodeB remote Apply changeQuantity error: %v", err)
+	if _, err := remote1Synced.Apply(cartID, change); err != nil {
+		t.Fatalf("remote mutation (remote1) changeQuantity error: %v", err)
 	}

 	// Validate updated state visible via nodeC
-	stateC, err := syncedC.Get(cartID)
+	stateC, err := remote2Synced.Get(cartID)
 	if err != nil {
 		t.Fatalf("nodeC Get error: %v", err)
 	}
--- a/pool-server.go
+++ b/pool-server.go
@@ -264,30 +264,64 @@ func (s *PoolServer) HandleCheckout(w http.ResponseWriter, r *http.Request, id C
 }

 func NewCartId() CartId {
+	// Deprecated: legacy random/time based cart id generator.
+	// Retained for compatibility; new code should prefer canonical CartID path.
+	cid, err := NewCartID()
+	if err != nil {
+		// Fallback to legacy method only if crypto/rand fails
 		id := time.Now().UnixNano() + rand.Int63()
-
 		return ToCartId(fmt.Sprintf("%d", id))
 	}
+	return CartIDToLegacy(cid)
+}

 func CookieCartIdHandler(fn func(w http.ResponseWriter, r *http.Request, cartId CartId) error) func(w http.ResponseWriter, r *http.Request) error {
 	return func(w http.ResponseWriter, r *http.Request) error {
-		var cartId CartId
-		cartIdCookie := r.CookiesNamed("cartid")
-		if cartIdCookie == nil || len(cartIdCookie) == 0 {
-			cartId = NewCartId()
+		// Extract / normalize cookie (preserve legacy textual IDs without rewriting).
+		var legacy CartId
+		cookies := r.CookiesNamed("cartid")
+		if len(cookies) == 0 {
+			// No cookie -> generate new canonical base62 id.
+			cid, generated, _, err := CanonicalizeOrLegacy("")
+			if err != nil {
+				return fmt.Errorf("failed to generate cart id: %w", err)
+			}
+			legacy = CartIDToLegacy(cid)
+			if generated {
 				http.SetCookie(w, &http.Cookie{
 					Name:     "cartid",
-				Value:    cartId.String(),
+					Value:    cid.String(),
 					Secure:   r.TLS != nil,
 					HttpOnly: true,
 					Path:     "/",
 					Expires:  time.Now().AddDate(0, 0, 14),
 					SameSite: http.SameSiteLaxMode,
 				})
-		} else {
-			cartId = ToCartId(cartIdCookie[0].Value)
+				w.Header().Set("Set-Cart-Id", cid.String())
 			}
-		return fn(w, r, cartId)
+		} else {
+			raw := cookies[0].Value
+			cid, generated, wasBase62, err := CanonicalizeOrLegacy(raw)
+			if err != nil {
+				return fmt.Errorf("failed to canonicalize cart id: %w", err)
+			}
+			legacy = CartIDToLegacy(cid)
+			// Only set a new cookie if we actually generated a brand-new ID (empty input).
+			// For legacy (non-base62) ids we preserve the original text and do not overwrite.
+			if generated && wasBase62 {
+				http.SetCookie(w, &http.Cookie{
+					Name:     "cartid",
+					Value:    cid.String(),
+					Secure:   r.TLS != nil,
+					HttpOnly: true,
+					Path:     "/",
+					Expires:  time.Now().AddDate(0, 0, 14),
+					SameSite: http.SameSiteLaxMode,
+				})
+				w.Header().Set("Set-Cart-Id", cid.String())
+			}
+		}
+		return fn(w, r, legacy)
 	}
 }

@@ -308,8 +342,18 @@ func (s *PoolServer) RemoveCartCookie(w http.ResponseWriter, r *http.Request, ca

 func CartIdHandler(fn func(w http.ResponseWriter, r *http.Request, cartId CartId) error) func(w http.ResponseWriter, r *http.Request) error {
 	return func(w http.ResponseWriter, r *http.Request) error {
-		cartId := ToCartId(r.PathValue("id"))
-		return fn(w, r, cartId)
+		raw := r.PathValue("id")
+		cid, generated, wasBase62, err := CanonicalizeOrLegacy(raw)
+		if err != nil {
+			return fmt.Errorf("invalid cart id: %w", err)
+		}
+		legacy := CartIDToLegacy(cid)
+		// Only emit Set-Cart-Id header if we produced a brand-new canonical id
+		// AND it is base62 (avoid rewriting legacy textual identifiers).
+		if generated && wasBase62 {
+			w.Header().Set("Set-Cart-Id", cid.String())
+		}
+		return fn(w, r, legacy)
 	}
 }

--- a/proto/control_plane.pb.go
+++ b/proto/control_plane.pb.go
@@ -420,13 +420,12 @@ const file_control_plane_proto_rawDesc = "" +
 	"\baccepted\x18\x01 \x01(\bR\baccepted\x12\x18\n" +
 	"\amessage\x18\x02 \x01(\tR\amessage\"#\n" +
 	"\rClosingNotice\x12\x12\n" +
-	"\x04host\x18\x01 \x01(\tR\x04host2\xbc\x02\n" +
+	"\x04host\x18\x01 \x01(\tR\x04host2\xf4\x01\n" +
 	"\fControlPlane\x12,\n" +
 	"\x04Ping\x12\x0f.messages.Empty\x1a\x13.messages.PingReply\x12A\n" +
 	"\tNegotiate\x12\x1a.messages.NegotiateRequest\x1a\x18.messages.NegotiateReply\x125\n" +
 	"\n" +
-	"GetCartIds\x12\x0f.messages.Empty\x1a\x16.messages.CartIdsReply\x12F\n" +
-	"\fConfirmOwner\x12\x1c.messages.OwnerChangeRequest\x1a\x18.messages.OwnerChangeAck\x12<\n" +
+	"GetCartIds\x12\x0f.messages.Empty\x1a\x16.messages.CartIdsReply\x12<\n" +
 	"\aClosing\x12\x17.messages.ClosingNotice\x1a\x18.messages.OwnerChangeAckB.Z,git.tornberg.me/go-cart-actor/proto;messagesb\x06proto3"

 var (
@@ -456,15 +455,13 @@ var file_control_plane_proto_depIdxs = []int32{
 	0, // 0: messages.ControlPlane.Ping:input_type -> messages.Empty
 	2, // 1: messages.ControlPlane.Negotiate:input_type -> messages.NegotiateRequest
 	0, // 2: messages.ControlPlane.GetCartIds:input_type -> messages.Empty
-	5, // 3: messages.ControlPlane.ConfirmOwner:input_type -> messages.OwnerChangeRequest
-	7, // 4: messages.ControlPlane.Closing:input_type -> messages.ClosingNotice
-	1, // 5: messages.ControlPlane.Ping:output_type -> messages.PingReply
-	3, // 6: messages.ControlPlane.Negotiate:output_type -> messages.NegotiateReply
-	4, // 7: messages.ControlPlane.GetCartIds:output_type -> messages.CartIdsReply
-	6, // 8: messages.ControlPlane.ConfirmOwner:output_type -> messages.OwnerChangeAck
-	6, // 9: messages.ControlPlane.Closing:output_type -> messages.OwnerChangeAck
-	5, // [5:10] is the sub-list for method output_type
-	0, // [0:5] is the sub-list for method input_type
+	7, // 3: messages.ControlPlane.Closing:input_type -> messages.ClosingNotice
+	1, // 4: messages.ControlPlane.Ping:output_type -> messages.PingReply
+	3, // 5: messages.ControlPlane.Negotiate:output_type -> messages.NegotiateReply
+	4, // 6: messages.ControlPlane.GetCartIds:output_type -> messages.CartIdsReply
+	6, // 7: messages.ControlPlane.Closing:output_type -> messages.OwnerChangeAck
+	4, // [4:8] is the sub-list for method output_type
+	0, // [0:4] is the sub-list for method input_type
 	0, // [0:0] is the sub-list for extension type_name
 	0, // [0:0] is the sub-list for extension extendee
 	0, // [0:0] is the sub-list for field type_name
--- a/proto/control_plane.proto
+++ b/proto/control_plane.proto
@@ -11,7 +11,7 @@ option go_package = "git.tornberg.me/go-cart-actor/proto;messages";
 // Responsibilities:
 //   - Liveness (Ping)
 //   - Membership negotiation (Negotiate)
-//   - Cart ownership change propagation (ConfirmOwner)
+//   - Deterministic ring-based ownership (ConfirmOwner RPC removed)
 //   - Cart ID listing for remote grain spawning (GetCartIds)
 //   - Graceful shutdown notifications (Closing)
 // No authentication / TLS is defined initially (can be added later).
@@ -41,13 +41,7 @@ message CartIdsReply {
  repeated string cart_ids = 1;
 }

-// OwnerChangeRequest notifies peers that ownership of a cart moved (or is moving) to new_host.
-message OwnerChangeRequest {
-  string cart_id = 1;
-  string new_host = 2;
-}
-
-// OwnerChangeAck indicates acceptance or rejection of an ownership change.
+// OwnerChangeAck retained as response type for Closing RPC (ConfirmOwner removed).
 message OwnerChangeAck {
  bool accepted = 1;
  string message = 2;
@@ -69,8 +63,7 @@ service ControlPlane {
  // GetCartIds lists currently owned cart IDs on this node.
  rpc GetCartIds(Empty) returns (CartIdsReply);

-  // ConfirmOwner announces/asks peers to acknowledge ownership transfer.
-  rpc ConfirmOwner(OwnerChangeRequest) returns (OwnerChangeAck);
+  // ConfirmOwner RPC removed (was legacy ownership acknowledgement; ring-based ownership now authoritative)

  // Closing announces graceful shutdown so peers can proactively adjust.
  rpc Closing(ClosingNotice) returns (OwnerChangeAck);
--- a/proto/control_plane_grpc.pb.go
+++ b/proto/control_plane_grpc.pb.go
@@ -22,7 +22,6 @@ const (
 	ControlPlane_Ping_FullMethodName       = "/messages.ControlPlane/Ping"
 	ControlPlane_Negotiate_FullMethodName  = "/messages.ControlPlane/Negotiate"
 	ControlPlane_GetCartIds_FullMethodName = "/messages.ControlPlane/GetCartIds"
-	ControlPlane_ConfirmOwner_FullMethodName = "/messages.ControlPlane/ConfirmOwner"
 	ControlPlane_Closing_FullMethodName    = "/messages.ControlPlane/Closing"
 )

@@ -38,8 +37,6 @@ type ControlPlaneClient interface {
 	Negotiate(ctx context.Context, in *NegotiateRequest, opts ...grpc.CallOption) (*NegotiateReply, error)
 	// GetCartIds lists currently owned cart IDs on this node.
 	GetCartIds(ctx context.Context, in *Empty, opts ...grpc.CallOption) (*CartIdsReply, error)
-	// ConfirmOwner announces/asks peers to acknowledge ownership transfer.
-	ConfirmOwner(ctx context.Context, in *OwnerChangeRequest, opts ...grpc.CallOption) (*OwnerChangeAck, error)
 	// Closing announces graceful shutdown so peers can proactively adjust.
 	Closing(ctx context.Context, in *ClosingNotice, opts ...grpc.CallOption) (*OwnerChangeAck, error)
 }
@@ -82,16 +79,6 @@ func (c *controlPlaneClient) GetCartIds(ctx context.Context, in *Empty, opts ...
 	return out, nil
 }

-func (c *controlPlaneClient) ConfirmOwner(ctx context.Context, in *OwnerChangeRequest, opts ...grpc.CallOption) (*OwnerChangeAck, error) {
-	cOpts := append([]grpc.CallOption{grpc.StaticMethod()}, opts...)
-	out := new(OwnerChangeAck)
-	err := c.cc.Invoke(ctx, ControlPlane_ConfirmOwner_FullMethodName, in, out, cOpts...)
-	if err != nil {
-		return nil, err
-	}
-	return out, nil
-}
-
 func (c *controlPlaneClient) Closing(ctx context.Context, in *ClosingNotice, opts ...grpc.CallOption) (*OwnerChangeAck, error) {
 	cOpts := append([]grpc.CallOption{grpc.StaticMethod()}, opts...)
 	out := new(OwnerChangeAck)
@@ -114,8 +101,6 @@ type ControlPlaneServer interface {
 	Negotiate(context.Context, *NegotiateRequest) (*NegotiateReply, error)
 	// GetCartIds lists currently owned cart IDs on this node.
 	GetCartIds(context.Context, *Empty) (*CartIdsReply, error)
-	// ConfirmOwner announces/asks peers to acknowledge ownership transfer.
-	ConfirmOwner(context.Context, *OwnerChangeRequest) (*OwnerChangeAck, error)
 	// Closing announces graceful shutdown so peers can proactively adjust.
 	Closing(context.Context, *ClosingNotice) (*OwnerChangeAck, error)
 	mustEmbedUnimplementedControlPlaneServer()
@@ -137,9 +122,6 @@ func (UnimplementedControlPlaneServer) Negotiate(context.Context, *NegotiateRequ
 func (UnimplementedControlPlaneServer) GetCartIds(context.Context, *Empty) (*CartIdsReply, error) {
 	return nil, status.Errorf(codes.Unimplemented, "method GetCartIds not implemented")
 }
-func (UnimplementedControlPlaneServer) ConfirmOwner(context.Context, *OwnerChangeRequest) (*OwnerChangeAck, error) {
-	return nil, status.Errorf(codes.Unimplemented, "method ConfirmOwner not implemented")
-}
 func (UnimplementedControlPlaneServer) Closing(context.Context, *ClosingNotice) (*OwnerChangeAck, error) {
 	return nil, status.Errorf(codes.Unimplemented, "method Closing not implemented")
 }
@@ -218,24 +200,6 @@ func _ControlPlane_GetCartIds_Handler(srv interface{}, ctx context.Context, dec
 	return interceptor(ctx, in, info, handler)
 }

-func _ControlPlane_ConfirmOwner_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
-	in := new(OwnerChangeRequest)
-	if err := dec(in); err != nil {
-		return nil, err
-	}
-	if interceptor == nil {
-		return srv.(ControlPlaneServer).ConfirmOwner(ctx, in)
-	}
-	info := &grpc.UnaryServerInfo{
-		Server:     srv,
-		FullMethod: ControlPlane_ConfirmOwner_FullMethodName,
-	}
-	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
-		return srv.(ControlPlaneServer).ConfirmOwner(ctx, req.(*OwnerChangeRequest))
-	}
-	return interceptor(ctx, in, info, handler)
-}
-
 func _ControlPlane_Closing_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
 	in := new(ClosingNotice)
 	if err := dec(in); err != nil {
@@ -273,10 +237,6 @@ var ControlPlane_ServiceDesc = grpc.ServiceDesc{
 			MethodName: "GetCartIds",
 			Handler:    _ControlPlane_GetCartIds_Handler,
 		},
-		{
-			MethodName: "ConfirmOwner",
-			Handler:    _ControlPlane_ConfirmOwner_Handler,
-		},
 		{
 			MethodName: "Closing",
 			Handler:    _ControlPlane_Closing_Handler,
--- a/ring.go
+++ b/ring.go
@@ -0,0 +1,344 @@
+package main
+
+import (
+	"encoding/binary"
+	"fmt"
+	"hash/fnv"
+	"sort"
+	"strings"
+	"sync"
+)
+
+// ring.go
+//
+// Consistent hashing ring skeleton for future integration.
+// --------------------------------------------------------
+// This file introduces a minimal, allocation‑light consistent hashing structure
+// intended to replace per-cart ownership negotiation. It focuses on:
+//   * Deterministic lookup: O(log V) via binary search
+//   * Even(ish) distribution using virtual nodes (vnodes)
+//   * Epoch / fingerprint tracking to detect membership drift
+//
+// NOT YET WIRED:
+//   * SyncedPool integration (ownerForCart, lazy migration)
+//   * Replication factor > 1
+//   * Persistent state migration
+//
+// Safe to import now; unused until explicit integration code is added.
+//
+// Design Notes
+// ------------
+// - Hosts contribute `vnodesPerHost` virtual nodes. Higher counts smooth
+//   distribution at cost of memory (V = hosts * vnodesPerHost).
+// - Hash of vnode = FNV1a64(host + "#" + index). For improved quality you
+//   can swap in xxhash or siphash later without changing API (but doing so
+//   will reshuffle ownership).
+// - Cart ownership lookup uses either cartID.Raw() when provided (uniform
+//   64-bit space) or falls back to hashing string forms (legacy).
+// - Epoch is monotonically increasing; consumers can fence stale results.
+//
+// Future Extensions
+// -----------------
+// - Weighted hosts (proportionally more vnodes).
+// - Replication: LookupN(h, n) to return primary + replicas.
+// - Streaming / diff-based ring updates (gossip).
+// - Hash function injection for deterministic test scenarios.
+//
+// ---------------------------------------------------------------------------
+
+// Vnode represents a single virtual node position on the ring.
+type Vnode struct {
+	Hash  uint64 // position on the ring
+	Host  string // physical host owning this vnode
+	Index int    // per-host vnode index (0..vnodesPerHost-1)
+}
+
+// Ring is an immutable consistent hash ring snapshot.
+type Ring struct {
+	Epoch       uint64
+	Vnodes      []Vnode // sorted by Hash
+	hosts       []string
+	fingerprint uint64 // membership fingerprint (order-independent)
+}
+
+// RingBuilder accumulates parameters to construct a Ring.
+type RingBuilder struct {
+	epoch         uint64
+	vnodesPerHost int
+	hosts         []string
+}
+
+// NewRingBuilder creates a builder with defaults.
+func NewRingBuilder() *RingBuilder {
+	return &RingBuilder{
+		vnodesPerHost: 64, // a reasonable default for small clusters
+	}
+}
+
+func (b *RingBuilder) WithEpoch(e uint64) *RingBuilder {
+	b.epoch = e
+	return b
+}
+
+func (b *RingBuilder) WithVnodesPerHost(n int) *RingBuilder {
+	if n > 0 {
+		b.vnodesPerHost = n
+	}
+	return b
+}
+
+func (b *RingBuilder) WithHosts(hosts []string) *RingBuilder {
+	uniq := make(map[string]struct{}, len(hosts))
+	out := make([]string, 0, len(hosts))
+	for _, h := range hosts {
+		h = strings.TrimSpace(h)
+		if h == "" {
+			continue
+		}
+		if _, ok := uniq[h]; ok {
+			continue
+		}
+		uniq[h] = struct{}{}
+		out = append(out, h)
+	}
+	sort.Strings(out)
+	b.hosts = out
+	return b
+}
+
+func (b *RingBuilder) Build() *Ring {
+	if len(b.hosts) == 0 {
+		return &Ring{
+			Epoch:       b.epoch,
+			Vnodes:      nil,
+			hosts:       nil,
+			fingerprint: 0,
+		}
+	}
+
+	totalVnodes := len(b.hosts) * b.vnodesPerHost
+	vnodes := make([]Vnode, 0, totalVnodes)
+
+	for _, host := range b.hosts {
+		for i := 0; i < b.vnodesPerHost; i++ {
+			h := hashVnode(host, i)
+			vnodes = append(vnodes, Vnode{
+				Hash:  h,
+				Host:  host,
+				Index: i,
+			})
+		}
+	}
+	sort.Slice(vnodes, func(i, j int) bool {
+		if vnodes[i].Hash == vnodes[j].Hash {
+			// Tie-break deterministically by host then index to avoid instability
+			if vnodes[i].Host == vnodes[j].Host {
+				return vnodes[i].Index < vnodes[j].Index
+			}
+			return vnodes[i].Host < vnodes[j].Host
+		}
+		return vnodes[i].Hash < vnodes[j].Hash
+	})
+
+	fp := fingerprintHosts(b.hosts)
+
+	return &Ring{
+		Epoch:       b.epoch,
+		Vnodes:      vnodes,
+		hosts:       append([]string(nil), b.hosts...),
+		fingerprint: fp,
+	}
+}
+
+// Hosts returns a copy of the host list (sorted).
+func (r *Ring) Hosts() []string {
+	if len(r.hosts) == 0 {
+		return nil
+	}
+	cp := make([]string, len(r.hosts))
+	copy(cp, r.hosts)
+	return cp
+}
+
+// Fingerprint returns a hash representing the unordered membership set.
+func (r *Ring) Fingerprint() uint64 {
+	return r.fingerprint
+}
+
+// Empty indicates ring has no vnodes.
+func (r *Ring) Empty() bool {
+	return len(r.Vnodes) == 0
+}
+
+// Lookup returns the vnode owning a given hash value.
+func (r *Ring) Lookup(h uint64) Vnode {
+	if len(r.Vnodes) == 0 {
+		return Vnode{}
+	}
+	// Binary search: first position with Hash >= h
+	i := sort.Search(len(r.Vnodes), func(i int) bool {
+		return r.Vnodes[i].Hash >= h
+	})
+	if i == len(r.Vnodes) {
+		return r.Vnodes[0]
+	}
+	return r.Vnodes[i]
+}
+
+// LookupID selects owner vnode for a CartID (fast path).
+func (r *Ring) LookupID(id CartID) Vnode {
+	return r.Lookup(id.Raw())
+}
+
+// LookupString hashes an arbitrary string and looks up owner.
+func (r *Ring) LookupString(s string) Vnode {
+	return r.Lookup(hashKeyString(s))
+}
+
+// LookupN returns up to n distinct host vnodes in ring order
+// starting from the primary owner of hash h (for replication).
+func (r *Ring) LookupN(h uint64, n int) []Vnode {
+	if n <= 0 || len(r.Vnodes) == 0 {
+		return nil
+	}
+	if n > len(r.hosts) {
+		n = len(r.hosts)
+	}
+	owners := make([]Vnode, 0, n)
+	seen := make(map[string]struct{}, n)
+
+	start := r.Lookup(h)
+
+	// Find index of start (can binary search again or linear scan; since we
+	// already have start.Hash we do another search for clarity)
+	i := sort.Search(len(r.Vnodes), func(i int) bool {
+		return r.Vnodes[i].Hash >= start.Hash
+	})
+	if i == len(r.Vnodes) {
+		i = 0
+	}
+
+	for idx := 0; len(owners) < n && idx < len(r.Vnodes); idx++ {
+		v := r.Vnodes[(i+idx)%len(r.Vnodes)]
+		if _, ok := seen[v.Host]; ok {
+			continue
+		}
+		seen[v.Host] = struct{}{}
+		owners = append(owners, v)
+	}
+	return owners
+}
+
+// DiffHosts compares this ring's membership to another.
+func (r *Ring) DiffHosts(other *Ring) (added []string, removed []string) {
+	if other == nil {
+		return r.Hosts(), nil
+	}
+	cur := make(map[string]struct{}, len(r.hosts))
+	for _, h := range r.hosts {
+		cur[h] = struct{}{}
+	}
+	oth := make(map[string]struct{}, len(other.hosts))
+	for _, h := range other.hosts {
+		oth[h] = struct{}{}
+	}
+	for h := range cur {
+		if _, ok := oth[h]; !ok {
+			removed = append(removed, h)
+		}
+	}
+	for h := range oth {
+		if _, ok := cur[h]; !ok {
+			added = append(added, h)
+		}
+	}
+	sort.Strings(added)
+	sort.Strings(removed)
+	return
+}
+
+// ---------------------------- Hash Functions ---------------------------------
+
+func hashVnode(host string, idx int) uint64 {
+	h := fnv.New64a()
+	_, _ = h.Write([]byte(host))
+	_, _ = h.Write([]byte{'#'})
+	var buf [8]byte
+	binary.BigEndian.PutUint64(buf[:], uint64(idx))
+	_, _ = h.Write(buf[:])
+	return h.Sum64()
+}
+
+// hashKeyString provides a stable hash for arbitrary string keys (legacy IDs).
+func hashKeyString(s string) uint64 {
+	h := fnv.New64a()
+	_, _ = h.Write([]byte(s))
+	return h.Sum64()
+}
+
+// fingerprintHosts produces an order-insensitive hash over the host set.
+func fingerprintHosts(hosts []string) uint64 {
+	if len(hosts) == 0 {
+		return 0
+	}
+	h := fnv.New64a()
+	for _, host := range hosts {
+		_, _ = h.Write([]byte(host))
+		_, _ = h.Write([]byte{0})
+	}
+	return h.Sum64()
+}
+
+// --------------------------- Thread-Safe Wrapper -----------------------------
+//
+// RingRef offers atomic swap + read semantics. SyncedPool can embed or hold
+// one of these to manage live ring updates safely.
+
+type RingRef struct {
+	mu   sync.RWMutex
+	ring *Ring
+}
+
+func NewRingRef(r *Ring) *RingRef {
+	return &RingRef{ring: r}
+}
+
+func (rr *RingRef) Get() *Ring {
+	rr.mu.RLock()
+	r := rr.ring
+	rr.mu.RUnlock()
+	return r
+}
+
+func (rr *RingRef) Set(r *Ring) {
+	rr.mu.Lock()
+	rr.ring = r
+	rr.mu.Unlock()
+}
+
+func (rr *RingRef) LookupID(id CartID) Vnode {
+	r := rr.Get()
+	if r == nil {
+		return Vnode{}
+	}
+	return r.LookupID(id)
+}
+
+// ----------------------------- Debug Utilities -------------------------------
+
+func (r *Ring) String() string {
+	var b strings.Builder
+	fmt.Fprintf(&b, "Ring{epoch=%d vnodes=%d hosts=%d}\n", r.Epoch, len(r.Vnodes), len(r.hosts))
+	limit := len(r.Vnodes)
+	if limit > 16 {
+		limit = 16
+	}
+	for i := 0; i < limit; i++ {
+		v := r.Vnodes[i]
+		fmt.Fprintf(&b, "  %02d hash=%016x host=%s idx=%d\n", i, v.Hash, v.Host, v.Index)
+	}
+	if len(r.Vnodes) > limit {
+		fmt.Fprintf(&b, "  ... (%d more)\n", len(r.Vnodes)-limit)
+	}
+	return b.String()
+}
--- a/synced-pool.go
+++ b/synced-pool.go
@@ -15,14 +15,15 @@ import (
 )

 // SyncedPool coordinates cart grain ownership across nodes using gRPC control plane
-// and cart actor services. Legacy frame / TCP code has been removed.
+// and cart actor services.
 //
 // Responsibilities:
 //   - Local grain access (delegates to GrainLocalPool)
 //   - Remote grain proxy management (RemoteGrainGRPC)
 //   - Cluster membership (AddRemote via discovery + negotiation)
-//   - Ownership acquisition (quorum via ConfirmOwner RPC)
 //   - Health/ping monitoring & remote removal
+//   - Ring based deterministic ownership (no runtime negotiation)
+//   - (Scaffolding) replication factor awareness via ring.LookupN
 //
 // Thread-safety: public methods that mutate internal maps lock p.mu (RWMutex).
 type SyncedPool struct {
@@ -40,7 +41,12 @@ type SyncedPool struct {
 	// Discovery handler for re-adding hosts after failures
 	discardedHostHandler *DiscardedHostHandler

-	// Metrics / instrumentation dependencies already declared globally
+	// Consistent hashing ring (immutable snapshot reference)
+	ringRef *RingRef
+
+	// Configuration
+	vnodesPerHost     int
+	replicationFactor int // RF (>=1). Currently only primary is active; replicas are scaffolding.
 }

 // RemoteHostGRPC tracks a remote host's clients & health.
@@ -71,8 +77,34 @@ var (
 	})
 	remoteLookupCount = promauto.NewCounter(prometheus.CounterOpts{
 		Name: "cart_remote_lookup_total",
-		Help: "The total number of remote lookups",
+		Help: "The total number of remote lookups (legacy counter)",
 	})
+
+	// Ring / ownership metrics
+	ringEpoch = promauto.NewGauge(prometheus.GaugeOpts{
+		Name: "cart_ring_epoch",
+		Help: "Current consistent hashing ring epoch (fingerprint-based pseudo-epoch)",
+	})
+	ringHosts = promauto.NewGauge(prometheus.GaugeOpts{
+		Name: "cart_ring_hosts",
+		Help: "Number of hosts currently in the ring",
+	})
+	ringVnodes = promauto.NewGauge(prometheus.GaugeOpts{
+		Name: "cart_ring_vnodes",
+		Help: "Number of virtual nodes in the ring",
+	})
+	ringLookupLocal = promauto.NewCounter(prometheus.CounterOpts{
+		Name: "cart_ring_lookup_local_total",
+		Help: "Ring ownership lookups resolved to the local host",
+	})
+	ringLookupRemote = promauto.NewCounter(prometheus.CounterOpts{
+		Name: "cart_ring_lookup_remote_total",
+		Help: "Ring ownership lookups resolved to a remote host",
+	})
+	ringHostShare = promauto.NewGaugeVec(prometheus.GaugeOpts{
+		Name: "cart_ring_host_share",
+		Help: "Fractional share of ring vnodes per host",
+	}, []string{"host"})
 )

 func NewSyncedPool(local *GrainLocalPool, hostname string, discovery Discovery) (*SyncedPool, error) {
@@ -82,8 +114,12 @@ func NewSyncedPool(local *GrainLocalPool, hostname string, discovery Discovery)
 		remoteHosts:          make(map[string]*RemoteHostGRPC),
 		remoteIndex:          make(map[CartId]Grain),
 		discardedHostHandler: NewDiscardedHostHandler(1338),
+		vnodesPerHost:        64, // default smoothing factor; adjust if needed
+		replicationFactor:    1,  // RF scaffold; >1 not yet activating replicas
 	}
 	p.discardedHostHandler.SetReconnectHandler(p.AddRemote)
+	// Initialize empty ring (will be rebuilt after first AddRemote or discovery event)
+	p.rebuildRing()

 	if discovery != nil {
 		go func() {
@@ -175,6 +211,8 @@ func (p *SyncedPool) AddRemote(host string) {
 	p.remoteHosts[host] = remote
 	p.mu.Unlock()
 	connectedRemotes.Set(float64(p.RemoteCount()))
+	// Rebuild consistent hashing ring including this new host
+	p.rebuildRing()

 	log.Printf("Connected to remote host %s", host)

@@ -222,6 +260,8 @@ func (p *SyncedPool) RemoveHost(host string) {
 		remote.Conn.Close()
 	}
 	connectedRemotes.Set(float64(p.RemoteCount()))
+	// Rebuild ring after host removal
+	p.rebuildRing()
 }

 // RemoteCount returns number of tracked remote hosts.
@@ -302,6 +342,8 @@ func (p *SyncedPool) Negotiate() {
 	}
 	p.mu.RUnlock()

+	changed := false
+
 	for _, r := range remotes {
 		ctx, cancel := context.WithTimeout(context.Background(), 2*time.Second)
 		reply, err := r.ControlClient.Negotiate(ctx, &proto.NegotiateRequest{KnownHosts: hosts})
@@ -313,12 +355,18 @@ func (p *SyncedPool) Negotiate() {
 		for _, h := range reply.Hosts {
 			if !p.IsKnown(h) {
 				p.AddRemote(h)
-			}
+				changed = true
 			}
 		}
 	}

-// ------------------------- Grain Management ----------------------------------
+	// If new hosts were discovered during negotiation, rebuild the ring once at the end.
+	if changed {
+		p.rebuildRing()
+	}
+}
+
+// ------------------------- Grain / Ring Ownership ----------------------------

 // RemoveRemoteGrain removes a remote grain mapping.
 func (p *SyncedPool) RemoveRemoteGrain(id CartId) {
@@ -333,9 +381,9 @@ func (p *SyncedPool) SpawnRemoteGrain(id CartId, host string) {
 		return
 	}
 	p.mu.Lock()
-	// If local grain exists, remove it (ownership changed)
-	if g, ok := p.local.grains[id]; ok && g != nil {
-		delete(p.local.grains, id)
+	// If local grain exists (legacy key), remove from local map (ownership moved).
+	if g, ok := p.local.grains[LegacyToCartKey(id)]; ok && g != nil {
+		delete(p.local.grains, LegacyToCartKey(id))
 	}
 	remoteHost, ok := p.remoteHosts[host]
 	if !ok {
@@ -362,78 +410,115 @@ func (p *SyncedPool) GetHealthyRemotes() []*RemoteHostGRPC {
 	return ret
 }

-// RequestOwnership attempts to become owner of a cart, requiring quorum.
-// On success local grain is (or will be) created; peers spawn remote proxies.
-func (p *SyncedPool) RequestOwnership(id CartId) error {
-	ok := 0
-	all := 0
-	remotes := p.GetHealthyRemotes()
-	log.Printf("RequestOwnership start id=%s host=%s healthyRemotes=%d", id, p.Hostname, len(remotes))
-	for _, r := range remotes {
-		log.Printf("RequestOwnership sending ConfirmOwner to host=%s id=%s", r.Host, id)
-		ctx, cancel := context.WithTimeout(context.Background(), 800*time.Millisecond)
-		reply, err := r.ControlClient.ConfirmOwner(ctx, &proto.OwnerChangeRequest{
-			CartId:  id.String(),
-			NewHost: p.Hostname,
-		})
-		cancel()
-		all++
-		if err != nil || reply == nil || !reply.Accepted {
-			log.Printf("RequestOwnership negative/failed response from host=%s id=%s err=%v reply=%v", r.Host, id, err, reply)
-			continue
+// rebuildRing reconstructs the consistent hashing ring from current host set
+// and updates ring-related metrics.
+func (p *SyncedPool) rebuildRing() {
+	p.mu.RLock()
+	hosts := make([]string, 0, len(p.remoteHosts)+1)
+	hosts = append(hosts, p.Hostname)
+	for h := range p.remoteHosts {
+		hosts = append(hosts, h)
 	}
-		ok++
-		log.Printf("RequestOwnership accept from host=%s id=%s (ok=%d all=%d)", r.Host, id, ok, all)
+	p.mu.RUnlock()
+
+	epochSeed := fingerprintHosts(hosts)
+	builder := NewRingBuilder().
+		WithHosts(hosts).
+		WithEpoch(epochSeed).
+		WithVnodesPerHost(p.vnodesPerHost)
+	r := builder.Build()
+	if p.ringRef == nil {
+		p.ringRef = NewRingRef(r)
+	} else {
+		p.ringRef.Set(r)
 	}

-	// Quorum rule (majority semantics):
-	//  - Let N = all remotes + 1 (self)
-	//  - We require ok + 1 (implicit self vote) >= floor(N/2)+1
-	//    => ok >= floor(N/2)
-	//  - Examples:
-	//      N=2 (all=1): threshold=1 (need 1 remote)
-	//      N=3 (all=2): threshold=1 (need 1 remote; previously required 2)
-	//      N=4 (all=3): threshold=2
-	//      N=5 (all=4): threshold=2
-	//  - This change allows faster ownership under partial remote availability in small clusters.
-	log.Printf("RequestOwnership quorum evaluation id=%s host=%s ok=%d all=%d", id, p.Hostname, ok, all)
-	threshold := (all + 1) / 2 // floor(N/2)
-	if ok < threshold {
-		p.removeLocalGrain(id)
-		log.Printf("RequestOwnership failed quorum id=%s host=%s ok=%d all=%d threshold=%d", id, p.Hostname, ok, all, threshold)
-		return fmt.Errorf("quorum not reached (ok=%d all=%d threshold=%d)", ok, all, threshold)
+	// Metrics
+	ringEpoch.Set(float64(r.Epoch))
+	ringHosts.Set(float64(len(r.Hosts())))
+	ringVnodes.Set(float64(len(r.Vnodes)))
+	ringHostShare.Reset()
+	if len(r.Vnodes) > 0 {
+		perHost := make(map[string]int)
+		for _, v := range r.Vnodes {
+			perHost[v.Host]++
 		}
-	grainSyncCount.Inc()
-	log.Printf("RequestOwnership success id=%s host=%s ok=%d all=%d threshold=%d", id, p.Hostname, ok, all, threshold)
-	return nil
+		total := float64(len(r.Vnodes))
+		for h, c := range perHost {
+			ringHostShare.WithLabelValues(h).Set(float64(c) / total)
+		}
+	}
+}
+
+// ForceRingRefresh exposes a manual ring rebuild hook (primarily for tests).
+func (p *SyncedPool) ForceRingRefresh() {
+	p.rebuildRing()
+}
+
+// ownersFor returns the ordered list of primary + replica owners for a cart id
+// (length min(replicationFactor, #hosts)). Currently only the first (primary)
+// is used. This scaffolds future replication work.
+func (p *SyncedPool) ownersFor(id CartId) []string {
+	if p.ringRef == nil || p.replicationFactor <= 0 {
+		return []string{p.Hostname}
+	}
+	r := p.ringRef.Get()
+	if r == nil || r.Empty() {
+		return []string{p.Hostname}
+	}
+	vnodes := r.LookupN(hashKeyString(id.String()), p.replicationFactor)
+	out := make([]string, 0, len(vnodes))
+	seen := make(map[string]struct{}, len(vnodes))
+	for _, v := range vnodes {
+		if _, ok := seen[v.Host]; ok {
+			continue
+		}
+		seen[v.Host] = struct{}{}
+		out = append(out, v.Host)
+	}
+	if len(out) == 0 {
+		out = append(out, p.Hostname)
+	}
+	return out
+}
+
+// ownerHostFor returns the primary owner host for a given id.
+func (p *SyncedPool) ownerHostFor(id CartId) string {
+	return p.ownersFor(id)[0]
+}
+
+// DebugOwnerHost exposes (for tests) the currently computed primary owner host.
+func (p *SyncedPool) DebugOwnerHost(id CartId) string {
+	return p.ownerHostFor(id)
 }

 func (p *SyncedPool) removeLocalGrain(id CartId) {
 	p.mu.Lock()
-	delete(p.local.grains, id)
+	delete(p.local.grains, LegacyToCartKey(id))
 	p.mu.Unlock()
 }

-// getGrain returns a local or remote grain. If absent, it synchronously attempts
-// to acquire ownership before spawning a local grain to eliminate the race where
-// the first mutation applies before peers have installed remote proxies.
+// getGrain returns a local or remote grain. For remote ownership it performs a
+// bounded readiness wait (small retries) to reduce first-call failures while
+// the remote connection & proxy are initializing.
 func (p *SyncedPool) getGrain(id CartId) (Grain, error) {
-	p.mu.RLock()
-	localGrain, isLocal := p.local.grains[id]
-	remoteGrain, isRemote := p.remoteIndex[id]
-	p.mu.RUnlock()
-
-	if isLocal && localGrain != nil {
-		return localGrain, nil
+	owner := p.ownerHostFor(id)
+	if owner == p.Hostname {
+		ringLookupLocal.Inc()
+		grain, err := p.local.GetGrain(id)
+		if err != nil {
+			return nil, err
 		}
-	if isRemote {
-		remoteLookupCount.Inc()
-		return remoteGrain, nil
+		return grain, nil
+	}
+	ringLookupRemote.Inc()
+
+	// Kick off remote dial if we don't yet know the owner.
+	if !p.IsKnown(owner) {
+		go p.AddRemote(owner)
 	}

-	// Synchronously attempt to claim ownership. If this fails (quorum not reached)
-	// we re-check for a newly appeared remote proxy (another node became owner).
-	if err := p.RequestOwnership(id); err != nil {
+	// Fast path existing proxy
 	p.mu.RLock()
 	if rg, ok := p.remoteIndex[id]; ok {
 		p.mu.RUnlock()
@@ -441,18 +526,39 @@ func (p *SyncedPool) getGrain(id CartId) (Grain, error) {
 		return rg, nil
 	}
 	p.mu.RUnlock()
-		return nil, err
+
+	const (
+		attempts    = 5
+		sleepPerTry = 40 * time.Millisecond
+	)
+
+	for attempt := 0; attempt < attempts; attempt++ {
+		// Try to spawn (idempotent if host already known)
+		if p.IsKnown(owner) {
+			p.SpawnRemoteGrain(id, owner)
+		}
+		// Check again
+		p.mu.RLock()
+		if rg, ok := p.remoteIndex[id]; ok {
+			p.mu.RUnlock()
+			remoteLookupCount.Inc()
+			return rg, nil
+		}
+		p.mu.RUnlock()
+
+		// Last attempt? break to return error.
+		if attempt == attempts-1 {
+			break
+		}
+		time.Sleep(sleepPerTry)
 	}

-	// Ownership acquired; now lazily spawn the local grain.
-	grain, err := p.local.GetGrain(id)
-	if err != nil {
-		return nil, err
-	}
-	return grain, nil
+	return nil, fmt.Errorf("remote owner %s not yet available for cart %s (after %d attempts)", owner, id.String(), attempts)
 }

 // Apply applies a single mutation to a grain (local or remote).
+// Replication (RF>1) scaffolding: future enhancement will fan-out mutations
+// to replica owners (best-effort) and reconcile quorum on read.
 func (p *SyncedPool) Apply(id CartId, mutation interface{}) (*CartGrain, error) {
 	grain, err := p.getGrain(id)
 	if err != nil {
@@ -462,6 +568,7 @@ func (p *SyncedPool) Apply(id CartId, mutation interface{}) (*CartGrain, error)
 }

 // Get returns current state of a grain (local or remote).
+// Future replication hook: Read-repair or quorum read can be added here.
 func (p *SyncedPool) Get(id CartId) (*CartGrain, error) {
 	grain, err := p.getGrain(id)
 	if err != nil {