diff --git a/nimbus/p2p/validate.nim b/nimbus/p2p/validate.nim
index c93151def..a7499b1e7 100644
--- a/nimbus/p2p/validate.nim
+++ b/nimbus/p2p/validate.nim
@@ -17,6 +17,7 @@ import
   ../vm_state,
   ../vm_types,
   ../vm_types2,
+  ./validate/cache,
   chronicles,
   eth/[common, rlp, trie/trie_defs],
   ethash,
@@ -29,6 +30,8 @@ import
   times
 
 export
+  cache.EpochHashCache,
+  cache.initEpochHashCache,
   results
 
 type
@@ -49,11 +52,6 @@ type
 
   Hash512 = MDigest[512]
 
-  CacheByEpoch* = OrderedTableRef[uint64, seq[Hash512]]
-
-const
-  CACHE_MAX_ITEMS = 10
-
 # ------------------------------------------------------------------------------
 # Private Helpers
 # ------------------------------------------------------------------------------
@@ -85,40 +83,7 @@ func isGenesis(header: BlockHeader): bool =
 # Private cache management functions
 # ------------------------------------------------------------------------------
 
-proc mkCacheBytes(blockNumber: uint64): seq[Hash512] =
-  mkcache(getCacheSize(blockNumber), getSeedhash(blockNumber))
-
-proc findFirstIndex(tab: CacheByEpoch): uint64 =
-  # Kludge: OrderedTable[] still misses a proper API
-  for key in tab.keys:
-    result = key
-    break
-
-
-proc getCache(cacheByEpoch: CacheByEpoch; blockNumber: uint64): seq[Hash512] =
-  # TODO: this is very inefficient
-  let epochIndex = blockNumber div EPOCH_LENGTH
-
-  # Get the cache if already generated, marking it as recently used
-  if epochIndex in cacheByEpoch:
-    let c = cacheByEpoch[epochIndex]
-    cacheByEpoch.del(epochIndex)  # pop and append at end
-    cacheByEpoch[epochIndex] = c
-    return c
-
-  # Limit memory usage for cache
-  if cacheByEpoch.len >= CACHE_MAX_ITEMS:
-    cacheByEpoch.del(cacheByEpoch.findFirstIndex)
-
-  # Generate the cache if it was not already in memory
-  # Simulate requesting mkcache by block number: multiply index by epoch length
-  var c = mkCacheBytes(epochIndex * EPOCH_LENGTH)
-  cacheByEpoch[epochIndex] = c
-
-  result = system.move(c)
-
-
-func cacheHash(x: openArray[Hash512]): Hash256 =
+func cacheHash(x: EpochHashDigest): Hash256 =
   var ctx: keccak256
   ctx.init()
 
@@ -134,10 +99,10 @@ func cacheHash(x: openArray[Hash512]): Hash256 =
 
 proc checkPOW(blockNumber: Uint256; miningHash, mixHash: Hash256;
               nonce: BlockNonce; difficulty: DifficultyInt;
-              cacheByEpoch: CacheByEpoch): Result[void,string] =
+              cacheByEpoch: var EpochHashCache): Result[void,string] =
   let
     blockNumber = blockNumber.truncate(uint64)
-    cache = cacheByEpoch.getCache(blockNumber)
+    cache = cacheByEpoch.getEpochCacheHash(blockNumber)
     size = getDataSize(blockNumber)
     miningOutput = hashimotoLight(
       size, cache, miningHash, uint64.fromBytesBE(nonce))
@@ -161,7 +126,7 @@ proc checkPOW(blockNumber: Uint256; miningHash, mixHash: Hash256;
   result = ok()
 
 
-proc validateSeal(cacheByEpoch: CacheByEpoch;
+proc validateSeal(cacheByEpoch: var EpochHashCache;
                   header: BlockHeader): Result[void,string] =
   let miningHeader = header.toMiningHeader
   let miningHash = miningHeader.hash
@@ -196,7 +161,7 @@ func validateGasLimit(gasLimit, parentGasLimit: GasInt): Result[void,string] =
   result = ok()
 
 proc validateHeader(header, parentHeader: BlockHeader; checkSealOK: bool;
-                     cacheByEpoch: CacheByEpoch): Result[void,string] =
+                     cacheByEpoch: var EpochHashCache): Result[void,string] =
   if header.extraData.len > 32:
     return err("BlockHeader.extraData larger than 32 bytes")
 
@@ -235,7 +200,7 @@ func validateUncle(currBlock, uncle, uncleParent: BlockHeader):
 
 proc validateUncles(chainDB: BaseChainDB; header: BlockHeader;
                     uncles: seq[BlockHeader]; checkSealOK: bool;
-                    cacheByEpoch: CacheByEpoch): Result[void,string] =
+                    cacheByEpoch: var EpochHashCache): Result[void,string] =
   let hasUncles = uncles.len > 0
   let shouldHaveUncles = header.ommersHash != EMPTY_UNCLE_HASH
 
@@ -343,13 +308,9 @@ proc validateTransaction*(vmState: BaseVMState, tx: Transaction,
 # Public functions, extracted from test_blockchain_json
 # ------------------------------------------------------------------------------
 
-proc newCacheByEpoch*(): CacheByEpoch =
-  newOrderedTable[uint64, seq[Hash512]]()
-
-
 proc validateKinship*(chainDB: BaseChainDB; header: BlockHeader;
                       uncles: seq[BlockHeader]; checkSealOK: bool;
-                      cacheByEpoch: CacheByEpoch): Result[void,string] =
+                      cacheByEpoch: var EpochHashCache): Result[void,string] =
   if header.isGenesis:
     if header.extraData.len > 32:
       return err("BlockHeader.extraData larger than 32 bytes")
diff --git a/nimbus/p2p/validate/cache.nim b/nimbus/p2p/validate/cache.nim
new file mode 100644
index 000000000..4f9922948
--- /dev/null
+++ b/nimbus/p2p/validate/cache.nim
@@ -0,0 +1,75 @@
+# Nimbus
+# Copyright (c) 2018 Status Research & Development GmbH
+# Licensed under either of
+#  * Apache License, version 2.0, ([LICENSE-APACHE](LICENSE-APACHE) or
+#    http://www.apache.org/licenses/LICENSE-2.0)
+#  * MIT license ([LICENSE-MIT](LICENSE-MIT) or
+#    http://opensource.org/licenses/MIT)
+# at your option. This file may not be copied, modified, or distributed except
+# according to those terms.
+
+## Hash as hash can
+## ================
+##
+## provide hash lists, indexed by epoch
+
+import
+  ethash,
+  nimcrypto,
+  tables
+
+type
+  EpochHashDigest* = seq[MDigest[512]]
+
+  EpochHashCache* = object
+    maxItems: int                              ## max number of entries
+    tab: OrderedTable[uint64,EpochHashDigest]  ## cache data table
+
+# ------------------------------------------------------------------------------
+# Private cache management functions
+# ------------------------------------------------------------------------------
+
+proc mkCacheBytes(blockNumber: uint64): seq[MDigest[512]] {.inline.} =
+  mkcache(getCacheSize(blockNumber), getSeedhash(blockNumber))
+
+# ------------------------------------------------------------------------------
+# Public functions
+# ------------------------------------------------------------------------------
+
+proc initEpochHashCache*(cache: var EpochHashCache; cacheMaxItems = 10) =
+  ## Initialise a new cache indexed by block epoch
+  cache.maxItems = cacheMaxItems
+
+  # note: Starting from Nim v0.20, tables are initialized by default and it
+  #       is not necessary to call initOrderedTable() function explicitly.
+
+
+proc getEpochCacheHash*(cache: var EpochHashCache;
+                        blockNumber: uint64): EpochHashDigest =
+  ## Return hash list, indexed by epoch of argument `blockNumber`
+  let epochIndex = blockNumber div EPOCH_LENGTH
+
+  # Get the cache if already generated, marking it as recently used
+  if epochIndex in cache.tab:
+    let value = cache.tab[epochIndex]
+    cache.tab.del(epochIndex)  # pop and append at end
+    cache.tab[epochIndex] = value
+    return value
+
+  # Limit memory usage for cache
+  if cache.maxItems <= cache.tab.len:
+    # Delete oldest entry
+    for key in cache.tab.keys:
+      # Kludge: OrderedTable[] still misses a proper API
+      cache.tab.del(key)
+      break
+
+  # Simulate requesting mkcache by block number: multiply index by epoch length
+  var data = mkCacheBytes(epochIndex * EPOCH_LENGTH)
+  cache.tab[epochIndex] = data
+
+  result = system.move(data)
+
+# ------------------------------------------------------------------------------
+# End
+# ------------------------------------------------------------------------------
diff --git a/tests/test_blockchain_json.nim b/tests/test_blockchain_json.nim
index ef14fccd0..b1bcfcfb5 100644
--- a/tests/test_blockchain_json.nim
+++ b/tests/test_blockchain_json.nim
@@ -40,7 +40,8 @@ type
     debugData    : JsonNode
     network      : string
 
-var cacheByEpoch = newCacheByEpoch()
+var cacheByEpoch: EpochHashCache
+cacheByEpoch.initEpochHashCache
 
 proc testFixture(node: JsonNode, testStatusIMPL: var TestStatus, debugMode = false, trace = false)