diff --git a/cmd/utils/hierarchical_coordinator.go b/cmd/utils/hierarchical_coordinator.go
index 470f360516..f427bac307 100644
--- a/cmd/utils/hierarchical_coordinator.go
+++ b/cmd/utils/hierarchical_coordinator.go
@@ -137,6 +137,7 @@ func (hc *HierarchicalCoordinator) startNode(logPath string, quaiBackend quai.Co
 	hc.p2p.Subscribe(location, &types.WorkObject{})
 	hc.p2p.Subscribe(location, common.Hash{})
 	hc.p2p.Subscribe(location, &types.Transaction{})
+	hc.p2p.Subscribe(location, &types.WorkObjectHeader{})
 
 	StartNode(stack)
 
diff --git a/consensus/blake3pow/consensus.go b/consensus/blake3pow/consensus.go
index d72d701dae..2d78881c26 100644
--- a/consensus/blake3pow/consensus.go
+++ b/consensus/blake3pow/consensus.go
@@ -24,7 +24,6 @@ import (
 
 // Blake3pow proof-of-work protocol constants.
 var (
-	maxUncles                     = 2         // Maximum number of uncles allowed in a single block
 	allowedFutureBlockTimeSeconds = int64(15) // Max seconds from current time allowed for blocks, before they're considered future blocks
 
 	ContextTimeFactor = big10
@@ -196,7 +195,7 @@ func (blake3pow *Blake3pow) VerifyUncles(chain consensus.ChainReader, block *typ
 		return nil
 	}
 	// Verify that there are at most 2 uncles included in this block
-	if len(block.Uncles()) > maxUncles {
+	if len(block.Uncles()) > params.MaxUncleCount {
 		return errTooManyUncles
 	}
 	if len(block.Uncles()) == 0 {
@@ -206,7 +205,7 @@ func (blake3pow *Blake3pow) VerifyUncles(chain consensus.ChainReader, block *typ
 	uncles, ancestors := mapset.NewSet(), make(map[common.Hash]*types.WorkObject)
 
 	number, parent := block.NumberU64(nodeCtx)-1, block.ParentHash(nodeCtx)
-	for i := 0; i < 7; i++ {
+	for i := 0; i < params.WorkSharesInclusionDepth; i++ {
 		ancestorHeader := chain.GetHeader(parent, number)
 		if ancestorHeader == nil {
 			break
@@ -244,8 +243,9 @@ func (blake3pow *Blake3pow) VerifyUncles(chain consensus.ChainReader, block *typ
 		if ancestors[uncle.ParentHash()] == nil || uncle.ParentHash() == block.ParentHash(nodeCtx) {
 			return errDanglingUncle
 		}
-		// Verify the seal and get the powHash for the given header
-		err := blake3pow.verifySeal(uncle)
+
+		// make sure that the work can be computed
+		_, err := blake3pow.ComputePowHash(uncle)
 		if err != nil {
 			return err
 		}
@@ -258,6 +258,15 @@ func (blake3pow *Blake3pow) VerifyUncles(chain consensus.ChainReader, block *typ
 			if expected.Cmp(uncle.Difficulty()) != 0 {
 				return fmt.Errorf("uncle has invalid difficulty: have %v, want %v", uncle.Difficulty(), expected)
 			}
+
+			// Verify that the block number is parent's +1
+			parentNumber := parent.Number(nodeCtx)
+			if chain.IsGenesisHash(parent.Hash()) {
+				parentNumber = big.NewInt(0)
+			}
+			if diff := new(big.Int).Sub(uncle.Number(), parentNumber); diff.Cmp(big.NewInt(1)) != 0 {
+				return consensus.ErrInvalidNumber
+			}
 		}
 	}
 	return nil
@@ -627,3 +636,7 @@ func (blake3pow *Blake3pow) NodeLocation() common.Location {
 func (blake3pow *Blake3pow) ComputePowLight(header *types.WorkObjectHeader) (common.Hash, common.Hash) {
 	panic("compute pow light doesnt exist for blake3")
 }
+
+func (blake3pow *Blake3pow) ComputePowHash(header *types.WorkObjectHeader) (common.Hash, error) {
+	return header.Hash(), nil
+}
diff --git a/consensus/blake3pow/poem.go b/consensus/blake3pow/poem.go
index 9d6fc48ebe..a986627384 100644
--- a/consensus/blake3pow/poem.go
+++ b/consensus/blake3pow/poem.go
@@ -77,6 +77,11 @@ func (blake3pow *Blake3pow) TotalLogS(chain consensus.GenesisReader, header *typ
 	if err != nil {
 		return big.NewInt(0)
 	}
+	workShareS, err := blake3pow.WorkShareLogS(header)
+	if err != nil {
+		return big.NewInt(0)
+	}
+	intrinsicS = new(big.Int).Add(intrinsicS, workShareS)
 	switch order {
 	case common.PRIME_CTX:
 		totalS := new(big.Int).Add(header.ParentEntropy(common.PRIME_CTX), header.ParentDeltaS(common.REGION_CTX))
@@ -119,6 +124,11 @@ func (blake3pow *Blake3pow) DeltaLogS(chain consensus.GenesisReader, header *typ
 	if err != nil {
 		return big.NewInt(0)
 	}
+	workShareS, err := blake3pow.WorkShareLogS(header)
+	if err != nil {
+		return big.NewInt(0)
+	}
+	intrinsicS = new(big.Int).Add(intrinsicS, workShareS)
 	switch order {
 	case common.PRIME_CTX:
 		return big.NewInt(0)
@@ -149,6 +159,43 @@ func (blake3pow *Blake3pow) UncledLogS(block *types.WorkObject) *big.Int {
 	return totalUncledLogS
 }
 
+func (blake3pow *Blake3pow) WorkShareLogS(wo *types.WorkObject) (*big.Int, error) {
+	workShares := wo.Uncles()
+	totalWsEntropy := big.NewInt(0)
+	for _, ws := range workShares {
+		powHash, err := blake3pow.ComputePowHash(ws)
+		if err != nil {
+			return big.NewInt(0), err
+		}
+		// compute the diff from the pow hash so that the work can be discounted
+		powDiff := new(big.Int).Div(common.Big2e256, new(big.Int).SetBytes(powHash.Bytes()))
+		// Two discounts need to be applied to the weight of each work share
+		// 1) Discount based on the amount of number of other possible work
+		// shares for the same entropy value
+		// 2) Discount based on the staleness of inclusion, for every block
+		// delay the weight gets reduced by the factor of 2
+
+		// Discount 1) only applies if the workshare has less weight than the
+		// work object threshold
+		wsDiff := new(big.Int).Set(powDiff)
+		woDiff := new(big.Int).Set(wo.Difficulty())
+		target := new(big.Int).Div(common.Big2e256, woDiff)
+		if new(big.Int).SetBytes(powHash.Bytes()).Cmp(target) > 0 { // powHash > target
+			c, _ := mathutil.BinaryLog(powDiff, mantBits)
+			thresoldC, _ := mathutil.BinaryLog(woDiff, mantBits)
+			wsDiff = new(big.Int).Div(wsDiff, new(big.Int).Exp(big.NewInt(2), big.NewInt(int64(thresoldC-c)), nil))
+		}
+		// Discount 2) applies to all shares regardless of the weight
+		wsDiff = new(big.Int).Div(wsDiff, big.NewInt(int64(wo.NumberU64(common.ZONE_CTX)-ws.NumberU64())))
+
+		wsTarget := new(big.Int).Div(common.Big2e256, wsDiff)
+		wsEntropy := blake3pow.IntrinsicLogS(common.BytesToHash(wsTarget.Bytes()))
+		// Add the entropy into the total entropy once the discount calculation is done
+		totalWsEntropy.Add(totalWsEntropy, wsEntropy)
+	}
+	return totalWsEntropy, nil
+}
+
 func (blake3pow *Blake3pow) UncledSubDeltaLogS(chain consensus.GenesisReader, header *types.WorkObject) *big.Int {
 	// Treating the genesis block differntly
 	if chain.IsGenesisHash(header.Hash()) {
diff --git a/consensus/blake3pow/sealer.go b/consensus/blake3pow/sealer.go
index f66663d1af..7268fd90d6 100644
--- a/consensus/blake3pow/sealer.go
+++ b/consensus/blake3pow/sealer.go
@@ -12,6 +12,8 @@ import (
 
 	"github.com/dominant-strategies/go-quai/core/types"
 	"github.com/dominant-strategies/go-quai/log"
+	"github.com/dominant-strategies/go-quai/params"
+	"modernc.org/mathutil"
 )
 
 const (
@@ -127,8 +129,12 @@ func (blake3pow *Blake3pow) Seal(header *types.WorkObject, results chan<- *types
 // seed that results in correct final header difficulty.
 func (blake3pow *Blake3pow) mine(header *types.WorkObject, id int, seed uint64, abort chan struct{}, found chan *types.WorkObject) {
 	// Extract some data from the header
+	diff := new(big.Int).Set(header.Difficulty())
+	c, _ := mathutil.BinaryLog(diff, mantBits)
+	workShareThreshold := c - params.WorkSharesThresholdDiff
+	workShareDiff := new(big.Int).Exp(big.NewInt(2), big.NewInt(int64(workShareThreshold)), nil)
 	var (
-		target = new(big.Int).Div(big2e256, header.Difficulty())
+		target = new(big.Int).Div(big2e256, workShareDiff)
 	)
 	// Start generating random nonces until we abort or find a good one
 	var (
diff --git a/consensus/consensus.go b/consensus/consensus.go
index 5be12b2b17..200c8d652c 100644
--- a/consensus/consensus.go
+++ b/consensus/consensus.go
@@ -99,6 +99,9 @@ type Engine interface {
 	// UncledLogS returns the log of the entropy reduction by uncles referenced in the block
 	UncledLogS(block *types.WorkObject) *big.Int
 
+	// WorkShareLogS returns the log of the entropy reduction by the workshare referenced in the block
+	WorkShareLogS(block *types.WorkObject) (*big.Int, error)
+
 	// UncledUncledSubDeltaLogS returns the log of the uncled entropy reduction  since the past coincident
 	UncledSubDeltaLogS(chain GenesisReader, header *types.WorkObject) *big.Int
 
@@ -151,6 +154,9 @@ type Engine interface {
 	// that a new block should have.
 	CalcDifficulty(chain ChainHeaderReader, parent *types.WorkObjectHeader) *big.Int
 
+	// ComputePowHash returns the pow hash of the workobject header
+	ComputePowHash(header *types.WorkObjectHeader) (common.Hash, error)
+
 	// IsDomCoincident returns true if this block satisfies the difficulty order
 	// of a dominant chain. If this node does not have a dominant chain (i.e.
 	// if this is a prime node), then the function will always return false.
diff --git a/consensus/progpow/consensus.go b/consensus/progpow/consensus.go
index 20120e2f78..397d4190de 100644
--- a/consensus/progpow/consensus.go
+++ b/consensus/progpow/consensus.go
@@ -25,7 +25,6 @@ import (
 
 // Progpow proof-of-work protocol constants.
 var (
-	maxUncles                     = 2         // Maximum number of uncles allowed in a single block
 	allowedFutureBlockTimeSeconds = int64(15) // Max seconds from current time allowed for blocks, before they're considered future blocks
 
 	ContextTimeFactor = big10
@@ -198,7 +197,7 @@ func (progpow *Progpow) VerifyUncles(chain consensus.ChainReader, block *types.W
 		return nil
 	}
 	// Verify that there are at most 2 uncles included in this block
-	if len(block.Uncles()) > maxUncles {
+	if len(block.Uncles()) > params.MaxUncleCount {
 		return errTooManyUncles
 	}
 	if len(block.Uncles()) == 0 {
@@ -208,7 +207,7 @@ func (progpow *Progpow) VerifyUncles(chain consensus.ChainReader, block *types.W
 	uncles, ancestors := mapset.NewSet(), make(map[common.Hash]*types.WorkObject)
 
 	number, parent := block.NumberU64(nodeCtx)-1, block.ParentHash(nodeCtx)
-	for i := 0; i < 7; i++ {
+	for i := 0; i < params.WorkSharesInclusionDepth; i++ {
 		ancestorHeader := chain.GetHeader(parent, number)
 		if ancestorHeader == nil {
 			break
@@ -246,8 +245,7 @@ func (progpow *Progpow) VerifyUncles(chain consensus.ChainReader, block *types.W
 		if ancestors[uncle.ParentHash()] == nil || uncle.ParentHash() == block.ParentHash(nodeCtx) {
 			return errDanglingUncle
 		}
-		// Verify the seal and get the powHash for the given header
-		_, err := progpow.verifySeal(uncle)
+		_, err := progpow.ComputePowHash(uncle)
 		if err != nil {
 			return err
 		}
@@ -260,6 +258,15 @@ func (progpow *Progpow) VerifyUncles(chain consensus.ChainReader, block *types.W
 			if expected.Cmp(uncle.Difficulty()) != 0 {
 				return fmt.Errorf("uncle has invalid difficulty: have %v, want %v", uncle.Difficulty(), expected)
 			}
+
+			// Verify that the block number is parent's +1
+			parentNumber := parent.Number(nodeCtx)
+			if chain.IsGenesisHash(parent.Hash()) {
+				parentNumber = big.NewInt(0)
+			}
+			if diff := new(big.Int).Sub(uncle.Number(), parentNumber); diff.Cmp(big.NewInt(1)) != 0 {
+				return consensus.ErrInvalidNumber
+			}
 		}
 	}
 	return nil
@@ -590,6 +597,20 @@ func (progpow *Progpow) verifySeal(header *types.WorkObjectHeader) (common.Hash,
 	return powHash.(common.Hash), nil
 }
 
+func (progpow *Progpow) ComputePowHash(header *types.WorkObjectHeader) (common.Hash, error) {
+	// Check progpow
+	mixHash := header.PowDigest.Load()
+	powHash := header.PowHash.Load()
+	if powHash == nil || mixHash == nil {
+		mixHash, powHash = progpow.ComputePowLight(header)
+	}
+	// Verify the calculated values against the ones provided in the header
+	if !bytes.Equal(header.MixHash().Bytes(), mixHash.(common.Hash).Bytes()) {
+		return common.Hash{}, errInvalidMixHash
+	}
+	return powHash.(common.Hash), nil
+}
+
 // Prepare implements consensus.Engine, initializing the difficulty field of a
 // header to conform to the progpow protocol. The changes are done inline.
 func (progpow *Progpow) Prepare(chain consensus.ChainHeaderReader, header *types.WorkObject, parent *types.WorkObject) error {
diff --git a/consensus/progpow/poem.go b/consensus/progpow/poem.go
index cb93439bdb..ea780bfe82 100644
--- a/consensus/progpow/poem.go
+++ b/consensus/progpow/poem.go
@@ -77,6 +77,11 @@ func (progpow *Progpow) TotalLogS(chain consensus.GenesisReader, header *types.W
 	if err != nil {
 		return big.NewInt(0)
 	}
+	workShareS, err := progpow.WorkShareLogS(header)
+	if err != nil {
+		return big.NewInt(0)
+	}
+	intrinsicS = new(big.Int).Add(intrinsicS, workShareS)
 	switch order {
 	case common.PRIME_CTX:
 		totalS := new(big.Int).Add(header.ParentEntropy(common.PRIME_CTX), header.ParentDeltaS(common.REGION_CTX))
@@ -111,10 +116,18 @@ func (progpow *Progpow) TotalLogPhS(header *types.WorkObject) *big.Int {
 }
 
 func (progpow *Progpow) DeltaLogS(chain consensus.GenesisReader, header *types.WorkObject) *big.Int {
+	if chain.IsGenesisHash(header.Hash()) {
+		return big.NewInt(0)
+	}
 	intrinsicS, order, err := progpow.CalcOrder(header)
 	if err != nil {
 		return big.NewInt(0)
 	}
+	workShareS, err := progpow.WorkShareLogS(header)
+	if err != nil {
+		return big.NewInt(0)
+	}
+	intrinsicS = new(big.Int).Add(intrinsicS, workShareS)
 	switch order {
 	case common.PRIME_CTX:
 		return big.NewInt(0)
@@ -145,6 +158,43 @@ func (progpow *Progpow) UncledLogS(block *types.WorkObject) *big.Int {
 	return totalUncledLogS
 }
 
+func (progpow *Progpow) WorkShareLogS(wo *types.WorkObject) (*big.Int, error) {
+	workShares := wo.Uncles()
+	totalWsEntropy := big.NewInt(0)
+	for _, ws := range workShares {
+		powHash, err := progpow.ComputePowHash(ws)
+		if err != nil {
+			return big.NewInt(0), err
+		}
+		// compute the diff from the pow hash so that the work can be discounted
+		powDiff := new(big.Int).Div(common.Big2e256, new(big.Int).SetBytes(powHash.Bytes()))
+		// Two discounts need to be applied to the weight of each work share
+		// 1) Discount based on the amount of number of other possible work
+		// shares for the same entropy value
+		// 2) Discount based on the staleness of inclusion, for every block
+		// delay the weight gets reduced by the factor of 2
+
+		// Discount 1) only applies if the workshare has less weight than the
+		// work object threshold
+		wsDiff := new(big.Int).Set(powDiff)
+		woDiff := new(big.Int).Set(wo.Difficulty())
+		target := new(big.Int).Div(common.Big2e256, woDiff)
+		if new(big.Int).SetBytes(powHash.Bytes()).Cmp(target) > 0 { // powHash > target
+			c, _ := mathutil.BinaryLog(powDiff, mantBits)
+			thresoldC, _ := mathutil.BinaryLog(woDiff, mantBits)
+			wsDiff = new(big.Int).Div(wsDiff, new(big.Int).Exp(big.NewInt(2), big.NewInt(int64(thresoldC-c)), nil))
+		}
+		// Discount 2) applies to all shares regardless of the weight
+		wsDiff = new(big.Int).Div(wsDiff, big.NewInt(int64(wo.NumberU64(common.ZONE_CTX)-ws.NumberU64())))
+
+		wsTarget := new(big.Int).Div(common.Big2e256, wsDiff)
+		wsEntropy := progpow.IntrinsicLogS(common.BytesToHash(wsTarget.Bytes()))
+		// Add the entropy into the total entropy once the discount calculation is done
+		totalWsEntropy.Add(totalWsEntropy, wsEntropy)
+	}
+	return totalWsEntropy, nil
+}
+
 func (progpow *Progpow) UncledSubDeltaLogS(chain consensus.GenesisReader, header *types.WorkObject) *big.Int {
 	// Treating the genesis block differntly
 	if chain.IsGenesisHash(header.Hash()) {
diff --git a/consensus/progpow/sealer.go b/consensus/progpow/sealer.go
index 2354f789a2..84d1d9b513 100644
--- a/consensus/progpow/sealer.go
+++ b/consensus/progpow/sealer.go
@@ -13,6 +13,8 @@ import (
 	"github.com/dominant-strategies/go-quai/common"
 	"github.com/dominant-strategies/go-quai/core/types"
 	"github.com/dominant-strategies/go-quai/log"
+	"github.com/dominant-strategies/go-quai/params"
+	"modernc.org/mathutil"
 )
 
 const (
@@ -128,8 +130,12 @@ func (progpow *Progpow) Seal(header *types.WorkObject, results chan<- *types.Wor
 // seed that results in correct final block difficulty.
 func (progpow *Progpow) mine(header *types.WorkObject, id int, seed uint64, abort chan struct{}, found chan *types.WorkObject) {
 	// Extract some data from the header
+	diff := new(big.Int).Set(header.Difficulty())
+	c, _ := mathutil.BinaryLog(diff, mantBits)
+	workShareThreshold := c - params.WorkSharesThresholdDiff
+	workShareDiff := new(big.Int).Exp(big.NewInt(2), big.NewInt(int64(workShareThreshold)), nil)
 	var (
-		target  = new(big.Int).Div(big2e256, header.Difficulty())
+		target  = new(big.Int).Div(big2e256, workShareDiff)
 		nodeCtx = progpow.config.NodeLocation.Context()
 	)
 	// Start generating random nonces until we abort or find a good one
diff --git a/core/core.go b/core/core.go
index c56791c7e5..6a260258c5 100644
--- a/core/core.go
+++ b/core/core.go
@@ -1037,6 +1037,10 @@ func (c *Core) SubscribePendingHeader(ch chan<- *types.WorkObject) event.Subscri
 
 func (c *Core) IsMining() bool { return c.sl.miner.Mining() }
 
+func (c *Core) SendWorkShare(workShare *types.WorkObjectHeader) error {
+	return c.sl.miner.worker.AddWorkShare(workShare)
+}
+
 //-------------------------//
 // State Processor methods //
 //-------------------------//
diff --git a/core/events.go b/core/events.go
index a5c019d2b0..076c454625 100644
--- a/core/events.go
+++ b/core/events.go
@@ -21,8 +21,7 @@ type ChainEvent struct {
 }
 
 type ChainSideEvent struct {
-	Blocks      []*types.WorkObject
-	ResetUncles bool
+	Blocks []*types.WorkObject
 }
 
 type ChainHeadEvent struct {
diff --git a/core/headerchain.go b/core/headerchain.go
index 35dbcbc775..57a71624e0 100644
--- a/core/headerchain.go
+++ b/core/headerchain.go
@@ -422,7 +422,7 @@ func (hc *HeaderChain) SetCurrentHeader(head *types.WorkObject) error {
 					blocks = append(blocks, block)
 				}
 			}
-			hc.chainSideFeed.Send(ChainSideEvent{Blocks: blocks, ResetUncles: true})
+			hc.chainSideFeed.Send(ChainSideEvent{Blocks: blocks})
 		}()
 	}
 
diff --git a/core/slice.go b/core/slice.go
index 5897a6de34..00adb75ab2 100644
--- a/core/slice.go
+++ b/core/slice.go
@@ -404,7 +404,7 @@ func (sl *Slice) Append(header *types.WorkObject, domPendingHeader *types.WorkOb
 			"location":   block.Location(),
 			"parentHash": block.ParentHash(nodeCtx),
 		}).Debug("Found uncle")
-		sl.hc.chainSideFeed.Send(ChainSideEvent{Blocks: []*types.WorkObject{block}, ResetUncles: false})
+		sl.hc.chainSideFeed.Send(ChainSideEvent{Blocks: []*types.WorkObject{block}})
 	}
 
 	// Chain head feed is only used by the Zone chains
diff --git a/core/worker.go b/core/worker.go
index 034f4ba0de..61300d2599 100644
--- a/core/worker.go
+++ b/core/worker.go
@@ -6,6 +6,7 @@ import (
 	"fmt"
 	"math/big"
 	"runtime/debug"
+	"sort"
 	"strings"
 	"sync"
 	"sync/atomic"
@@ -28,6 +29,7 @@ import (
 	"github.com/dominant-strategies/go-quai/trie"
 	lru "github.com/hashicorp/golang-lru"
 	expireLru "github.com/hnlq715/golang-lru"
+	"modernc.org/mathutil"
 )
 
 const (
@@ -56,6 +58,10 @@ const (
 
 	// c_chainSideChanSize is the size of the channel listening to uncle events
 	chainSideChanSize = 10
+
+	c_uncleCacheSize = 50
+
+	c_workShareFilterDist = 100 // the dist from the current block for the work share inclusion in the worker
 )
 
 // environment is the worker's current environment and holds all
@@ -209,9 +215,8 @@ type worker struct {
 
 	wg sync.WaitGroup
 
-	localUncles  map[common.Hash]*types.WorkObjectHeader // A set of side blocks generated locally as the possible uncle blocks.
-	remoteUncles map[common.Hash]*types.WorkObjectHeader // A set of side blocks as the possible uncle blocks.
-	uncleMu      sync.RWMutex
+	Uncles  *lru.Cache
+	uncleMu sync.RWMutex
 
 	mu       sync.RWMutex // The lock used to protect the coinbase and extra fields
 	coinbase common.Address
@@ -279,8 +284,6 @@ func newWorker(config *Config, chainConfig *params.ChainConfig, db ethdb.Databas
 		coinbase:                       config.Etherbase,
 		isLocalBlock:                   isLocalBlock,
 		workerDb:                       db,
-		localUncles:                    make(map[common.Hash]*types.WorkObjectHeader),
-		remoteUncles:                   make(map[common.Hash]*types.WorkObjectHeader),
 		chainHeadCh:                    make(chan ChainHeadEvent, chainHeadChanSize),
 		chainSideCh:                    make(chan ChainSideEvent, chainSideChanSize),
 		taskCh:                         make(chan *task),
@@ -292,6 +295,8 @@ func newWorker(config *Config, chainConfig *params.ChainConfig, db ethdb.Databas
 		fillTransactionsRollingAverage: &RollingAverage{windowSize: 100},
 		logger:                         logger,
 	}
+	// initialize a uncle cache
+	worker.Uncles, _ = lru.New(c_uncleCacheSize)
 	// Set the GasFloor of the worker to the minGasLimit
 	worker.config.GasFloor = params.MinGasLimit
 
@@ -498,34 +503,12 @@ func (w *worker) asyncStateLoop() {
 						}).Fatal("Go-Quai Panicked")
 					}
 				}()
-				if side.ResetUncles {
-					w.uncleMu.Lock()
-					w.localUncles = make(map[common.Hash]*types.WorkObjectHeader)
-					w.remoteUncles = make(map[common.Hash]*types.WorkObjectHeader)
-					w.uncleMu.Unlock()
-				}
 				for _, wo := range side.Blocks {
-
 					// Short circuit for duplicate side blocks
-					w.uncleMu.RLock()
-					if _, exists := w.localUncles[wo.Hash()]; exists {
-						w.uncleMu.RUnlock()
-						continue
-					}
-					if _, exists := w.remoteUncles[wo.Hash()]; exists {
-						w.uncleMu.RUnlock()
+					if _, exists := w.Uncles.Get(wo.Hash()); exists {
 						continue
 					}
-					w.uncleMu.RUnlock()
-					if w.isLocalBlock != nil && w.isLocalBlock(wo) {
-						w.uncleMu.Lock()
-						w.localUncles[wo.Hash()] = wo.WorkObjectHeader()
-						w.uncleMu.Unlock()
-					} else {
-						w.uncleMu.Lock()
-						w.remoteUncles[wo.Hash()] = wo.WorkObjectHeader()
-						w.uncleMu.Unlock()
-					}
+					w.Uncles.ContainsOrAdd(wo.Hash(), wo.WorkObjectHeader())
 				}
 			}()
 		case <-w.exitCh:
@@ -715,7 +698,7 @@ func (w *worker) makeEnv(parent *types.WorkObject, proposedWo *types.WorkObject,
 		etxPLimit: etxPLimit,
 	}
 	// when 08 is processed ancestors contain 07 (quick block)
-	for _, ancestor := range w.hc.GetBlocksFromHash(parent.Header().Hash(), 7) {
+	for _, ancestor := range w.hc.GetBlocksFromHash(parent.Header().Hash(), params.WorkSharesInclusionDepth) {
 		for _, uncle := range ancestor.Uncles() {
 			env.family.Add(uncle.Hash())
 		}
@@ -1221,29 +1204,42 @@ func (w *worker) prepareWork(genParams *generateParams, wo *types.WorkObject) (*
 			return nil, err
 		}
 		// Accumulate the uncles for the sealing work.
-		commitUncles := func(wos map[common.Hash]*types.WorkObjectHeader) {
-			for hash, uncle := range wos {
+		commitUncles := func(wos *lru.Cache) {
+			var uncles []*types.WorkObjectHeader
+			keys := wos.Keys()
+			for _, hash := range keys {
+				if value, exist := wos.Peek(hash); exist {
+					uncle := value.(*types.WorkObjectHeader)
+					uncles = append(uncles, uncle)
+				}
+			}
+			// sort the uncles in the decreasing order of entropy
+			sort.Slice(uncles, func(i, j int) bool {
+				powHash1, _ := w.engine.ComputePowHash(uncles[i])
+				powHash2, _ := w.engine.ComputePowHash(uncles[j])
+				return new(big.Int).SetBytes(powHash1.Bytes()).Cmp(new(big.Int).SetBytes(powHash2.Bytes())) < 0
+			})
+			for _, uncle := range uncles {
 				env.uncleMu.RLock()
-				if len(env.uncles) == 2 {
+				if len(env.uncles) == params.MaxUncleCount {
 					env.uncleMu.RUnlock()
 					break
 				}
 				env.uncleMu.RUnlock()
 				if err := w.commitUncle(env, uncle); err != nil {
 					w.logger.WithFields(log.Fields{
-						"hash":   hash,
+						"hash":   uncle.Hash(),
 						"reason": err,
 					}).Trace("Possible uncle rejected")
 				} else {
-					w.logger.WithField("hash", hash).Debug("Committing new uncle to block")
+					w.logger.WithField("hash", uncle.Hash()).Debug("Committing new uncle to block")
 				}
 			}
 		}
 		if nodeCtx == common.ZONE_CTX && w.hc.ProcessingState() {
 			w.uncleMu.RLock()
 			// Prefer to locally generated uncle
-			commitUncles(w.localUncles)
-			commitUncles(w.remoteUncles)
+			commitUncles(w.Uncles)
 			w.uncleMu.RUnlock()
 		}
 		return env, nil
@@ -1353,7 +1349,7 @@ func (w *worker) FinalizeAssemble(chain consensus.ChainHeaderReader, newWo *type
 
 	// Once the uncles list is assembled in the block
 	if nodeCtx == common.ZONE_CTX {
-		wo.Header().SetUncledS(w.engine.UncledLogS(wo))
+		wo.Header().SetUncledS(w.engine.UncledLogS(parent))
 	}
 
 	manifestHash := w.ComputeManifestHash(parent)
@@ -1420,6 +1416,36 @@ func totalFees(block *types.WorkObject, receipts []*types.Receipt) *big.Float {
 	return new(big.Float).Quo(new(big.Float).SetInt(feesWei), new(big.Float).SetInt(big.NewInt(params.Ether)))
 }
 
+func (w *worker) AddWorkShare(workShare *types.WorkObjectHeader) error {
+	// compute the work of the workshare and see if there is any error and have
+	// an option to filter the workshares
+	powHash, err := w.engine.ComputePowHash(workShare)
+	if err != nil { // make sure it has valid pow
+		return err
+	}
+
+	// Don't add the workshare into the list if its farther than the worksharefilterdist
+	if workShare.NumberU64()+c_workShareFilterDist < w.hc.CurrentHeader().NumberU64(common.ZONE_CTX) {
+		return nil
+	}
+
+	// have a filter based on the current header entropy This filter can be
+	// optimized and the node can have the option to choose the threshold it
+	// likes, for now, we are using the default threshold
+	diff := new(big.Int).Set(w.hc.CurrentHeader().Difficulty())
+	c, _ := mathutil.BinaryLog(diff, 64)
+	// subtract params.WorkSharesThresholdDiff bits from the current difficulty
+	workShareThreshold := c - params.WorkSharesThresholdDiff
+	workShareDiff := new(big.Int).Exp(big.NewInt(2), big.NewInt(int64(workShareThreshold)), nil)
+	target := new(big.Int).Div(common.Big2e256, workShareDiff)
+	if new(big.Int).SetBytes(powHash.Bytes()).Cmp(target) > 0 {
+		return nil
+	}
+
+	w.Uncles.ContainsOrAdd(workShare.Hash(), workShare)
+	return nil
+}
+
 func (w *worker) CurrentInfo(header *types.WorkObject) bool {
 	if w.headerPrints.Contains(header.Hash()) {
 		return false
diff --git a/internal/quaiapi/backend.go b/internal/quaiapi/backend.go
index fac1263d0b..43ef0a537f 100644
--- a/internal/quaiapi/backend.go
+++ b/internal/quaiapi/backend.go
@@ -99,6 +99,7 @@ type Backend interface {
 	AddGenesisPendingEtxs(block *types.WorkObject)
 	SubscribeExpansionEvent(ch chan<- core.ExpansionEvent) event.Subscription
 	WriteGenesisBlock(block *types.WorkObject, location common.Location)
+	SendWorkShare(workShare *types.WorkObjectHeader) error
 
 	// Transaction pool API
 	SendTx(ctx context.Context, signedTx *types.Transaction) error
@@ -129,6 +130,7 @@ type Backend interface {
 
 	// P2P apis
 	BroadcastBlock(block *types.WorkObject, location common.Location) error
+	BroadcastWorkShare(workShare *types.WorkObjectHeader, location common.Location) error
 }
 
 func GetAPIs(apiBackend Backend) []rpc.API {
diff --git a/internal/quaiapi/quai_api.go b/internal/quaiapi/quai_api.go
index aa92617d9c..e2c11c0ae6 100644
--- a/internal/quaiapi/quai_api.go
+++ b/internal/quaiapi/quai_api.go
@@ -675,6 +675,26 @@ func (s *PublicBlockChainQuaiAPI) ReceiveMinedHeader(ctx context.Context, raw js
 	return nil
 }
 
+func (s *PublicBlockChainQuaiAPI) ReceiveWorkShare(ctx context.Context, raw json.RawMessage) error {
+	nodeCtx := s.b.NodeCtx()
+	if nodeCtx != common.ZONE_CTX {
+		return errors.New("work shares cannot be broadcasted in non-zone chain")
+	}
+	var workShare *types.WorkObjectHeader
+	if err := json.Unmarshal(raw, &workShare); err != nil {
+		return err
+	}
+	if workShare != nil {
+		s.b.Logger().WithField("number", workShare.NumberU64()).Info("Received Work Share")
+		err := s.b.BroadcastWorkShare(workShare, s.b.NodeLocation())
+		if err != nil {
+			log.Global.WithField("err", err).Error("Error broadcasting work share")
+			return err
+		}
+	}
+	return nil
+}
+
 type tdBlock struct {
 	Header           *types.WorkObject   `json:"header"`
 	Manifest         types.BlockManifest `json:"manifest"`
diff --git a/p2p/node/api.go b/p2p/node/api.go
index 0208449f2e..f0b31dcd82 100644
--- a/p2p/node/api.go
+++ b/p2p/node/api.go
@@ -315,6 +315,7 @@ func (p *P2PNode) handleBroadcast(sourcePeer peer.ID, data interface{}, nodeLoca
 		p.cacheAdd(v.Hash(), &v, nodeLocation)
 	// TODO: send it to consensus
 	case types.Transaction:
+	case types.WorkObjectHeader:
 	default:
 		log.Global.Debugf("received unsupported block broadcast")
 		// TODO: ban the peer which sent it?
diff --git a/p2p/pb/proto_services.go b/p2p/pb/proto_services.go
index 0f20051077..25aeee883a 100644
--- a/p2p/pb/proto_services.go
+++ b/p2p/pb/proto_services.go
@@ -215,6 +215,13 @@ func ConvertAndMarshal(data interface{}) ([]byte, error) {
 		log.Global.Tracef("marshalling hash: %+v", data)
 		protoHash := data.ProtoEncode()
 		return proto.Marshal(protoHash)
+	case *types.WorkObjectHeader:
+		log.Global.Tracef("marshalling block header: %+v", data)
+		protoWoHeader, err := data.ProtoEncode()
+		if err != nil {
+			return nil, err
+		}
+		return proto.Marshal(protoWoHeader)
 	default:
 		return nil, errors.New("unsupported data type")
 	}
@@ -239,6 +246,19 @@ func UnmarshalAndConvert(data []byte, sourceLocation common.Location, dataPtr *i
 		}
 		*dataPtr = *workObject
 		return nil
+	case *types.WorkObjectHeader:
+		protoWorkObjectHeader := &types.ProtoWorkObjectHeader{}
+		err := proto.Unmarshal(data, protoWorkObjectHeader)
+		if err != nil {
+			return err
+		}
+		workObjectHeader := &types.WorkObjectHeader{}
+		err = workObjectHeader.ProtoDecode(protoWorkObjectHeader)
+		if err != nil {
+			return err
+		}
+		*dataPtr = *workObjectHeader
+		return nil
 	case *types.Header:
 		protoHeader := &types.ProtoHeader{}
 		err := proto.Unmarshal(data, protoHeader)
diff --git a/p2p/pubsubManager/utils.go b/p2p/pubsubManager/utils.go
index 5a0aa51a48..de2e6dd29e 100644
--- a/p2p/pubsubManager/utils.go
+++ b/p2p/pubsubManager/utils.go
@@ -10,10 +10,11 @@ import (
 
 const (
 	// Data types for gossipsub topics
-	C_workObjectType  = "blocks"
-	C_transactionType = "transactions"
-	C_headerType      = "headers"
-	C_hashType        = "hash"
+	C_workObjectType       = "blocks"
+	C_transactionType      = "transactions"
+	C_headerType           = "headers"
+	C_hashType             = "hash"
+	C_workObjectHeaderType = "woHeaders"
 )
 
 // gets the name of the topic for the given type of data
@@ -26,6 +27,8 @@ func TopicName(genesis common.Hash, location common.Location, data interface{})
 		return strings.Join([]string{baseTopic, C_hashType}, "/"), nil
 	case *types.Transaction:
 		return strings.Join([]string{baseTopic, C_transactionType}, "/"), nil
+	case *types.WorkObjectHeader:
+		return strings.Join([]string{baseTopic, C_workObjectHeaderType}, "/"), nil
 	default:
 		return "", ErrUnsupportedType
 	}
diff --git a/params/protocol_params.go b/params/protocol_params.go
index dffdc4ade2..7e8f2be60f 100644
--- a/params/protocol_params.go
+++ b/params/protocol_params.go
@@ -201,4 +201,7 @@ var (
 	DifficultyAdjustmentPeriod        = big.NewInt(360)                                             // This is the number of blocks over which the average has to be taken
 	DifficultyAdjustmentFactor int64  = 40                                                          // This is the factor that divides the log of the change in the difficulty
 	MinQuaiConversionAmount           = new(big.Int).Mul(big.NewInt(1), big.NewInt(GWei))           // 0.000000001 Quai
+	MaxUncleCount                     = 50
+	WorkSharesThresholdDiff           = 4 // Number of bits lower than the target that the default consensus engine uses
+	WorkSharesInclusionDepth          = 7 // Number of blocks upto which the work shares can be referenced and this is protocol enforced
 )
diff --git a/quai/api_backend.go b/quai/api_backend.go
index ded98db698..801fffbe6b 100644
--- a/quai/api_backend.go
+++ b/quai/api_backend.go
@@ -574,9 +574,17 @@ func (b *QuaiAPIBackend) SubscribeExpansionEvent(ch chan<- core.ExpansionEvent)
 	return b.quai.core.SubscribeExpansionEvent(ch)
 }
 
+func (b *QuaiAPIBackend) SendWorkShare(workShare *types.WorkObjectHeader) error {
+	return b.quai.core.SendWorkShare(workShare)
+}
+
 // ///////////////////////////
 // /////// P2P ///////////////
 // ///////////////////////////
 func (b *QuaiAPIBackend) BroadcastBlock(block *types.WorkObject, location common.Location) error {
 	return b.quai.p2p.Broadcast(location, block)
 }
+
+func (b *QuaiAPIBackend) BroadcastWorkShare(workShare *types.WorkObjectHeader, location common.Location) error {
+	return b.quai.p2p.Broadcast(location, workShare)
+}
diff --git a/quai/p2p_backend.go b/quai/p2p_backend.go
index a088481d1b..ab1f005730 100644
--- a/quai/p2p_backend.go
+++ b/quai/p2p_backend.go
@@ -103,6 +103,14 @@ func (qbe *QuaiBackend) OnNewBroadcast(sourcePeer p2p.PeerID, data interface{},
 			backend.SendRemoteTx(&tx)
 		}
 		// TODO: Handle the error here and mark the peers accordingly
+	case types.WorkObjectHeader:
+		woHeader := data.(types.WorkObjectHeader)
+		backend := *qbe.GetBackend(nodeLocation)
+		if backend == nil {
+			log.Global.Error("no backend found")
+			return false
+		}
+		backend.SendWorkShare(&woHeader)
 	}
 
 	// If it was a good broadcast, mark the peer as lively
diff --git a/quaiclient/quaiclient.go b/quaiclient/quaiclient.go
index bc84712a1d..79f0ce28a9 100644
--- a/quaiclient/quaiclient.go
+++ b/quaiclient/quaiclient.go
@@ -303,6 +303,11 @@ func (ec *Client) ReceiveMinedHeader(ctx context.Context, header *types.WorkObje
 	return ec.c.CallContext(ctx, nil, "quai_receiveMinedHeader", data)
 }
 
+func (ec *Client) ReceiveWorkShare(ctx context.Context, header *types.WorkObjectHeader) error {
+	data := header.RPCMarshalWorkObjectHeader()
+	return ec.c.CallContext(ctx, nil, "quai_receiveWorkShare", data)
+}
+
 // Filters
 
 // SubscribeFilterLogs subscribes to the results of a streaming filter query.