diff --git a/ci/scripts/prepare-stack-data-directory.sh b/ci/scripts/prepare-stack-data-directory.sh
index d89605555990dabd2c385462e3c27c2f13190888..4081d13fdf8bf55d8f588995926e06d57af0c858 100755
--- a/ci/scripts/prepare-stack-data-directory.sh
+++ b/ci/scripts/prepare-stack-data-directory.sh
@@ -7,7 +7,9 @@ HAF_MOUNTPOINT="${1:?"Please pass a valid path for the stack's data directory to
 echo "Creating HAF's mountpoint at ${HAF_MOUNTPOINT}..."
 
 mkdir -p "${HAF_MOUNTPOINT}/blockchain"
-mkdir -p "${HAF_MOUNTPOINT}/shared_memory/haf_wal"
+mkdir -p "${HAF_MOUNTPOINT}/state/shared_memory"
+mkdir -p "${HAF_MOUNTPOINT}/state/haf_wal"
+mkdir -p "${HAF_MOUNTPOINT}/state/rocksdb"
 mkdir -p "${HAF_MOUNTPOINT}/logs/caddy"
 mkdir -p "${HAF_MOUNTPOINT}/logs/pgbadger"
 mkdir -p "${HAF_MOUNTPOINT}/logs/postgresql"
diff --git a/clear_pool.sh b/clear_pool.sh
index f6268c1b042a21e31aca4e1692b2fafcd3fda472..f7510c02f33e1bab399002ab8b874e13efa6711f 100755
--- a/clear_pool.sh
+++ b/clear_pool.sh
@@ -2,4 +2,4 @@
 . ./.env
 sudo rm -rf ${HAF_LOG_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/logs}/postgresql/*
 sudo rm -rf ${HAF_DATA_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}}/haf_db_store/*
-rm -rf ${HAF_SHM_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/shared_memory}/shared_memory.bin ${HAF_SHM_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/shared_memory}/haf_wal/* ${HAF_SHM_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/shared_memory}/comments-rocksdb-storage
+rm -rf ${HAF_SHM_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/state/shared_memory}/shared_memory.bin ${HAF_WAL_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/state/haf_wal}/* ${TOP_LEVEL_DATASET_MOUNTPOINT}/state/rocksdb/*
diff --git a/create_zfs_datasets.sh b/create_zfs_datasets.sh
index c06cf8110d0fae6b32f02aab5e6515ebac658879..f07c8a2464293fc79c0cd4f39c319762a02f09d5 100755
--- a/create_zfs_datasets.sh
+++ b/create_zfs_datasets.sh
@@ -88,14 +88,32 @@ zfs_uncompressed_options="-o compression=off"
 # several sites recommend 8k blocks for optimizing postgres on zfs, but we have found that it
 # kills compression ratios for haf_block_log, so we've decided to leave it at the default 128k
 zfs_postgres_options="" # or "-o recordsize=8k -o recordsize=16k", consider also "-o logbias=throughput"
+zfs_shared_memory_options="-o primarycache=metadata -o recordsize=64k -o logbias=throughput -o sync=disabled"
+# RocksDB optimizations: smaller recordsize for better random I/O, disabled compression (RocksDB handles it),
+# and settings to optimize for write-heavy workloads with compaction
+zfs_rocksdb_options="-o recordsize=16k -o primarycache=all -o logbias=throughput -o sync=standard -o redundant_metadata=most"
+# Blockchain optimizations: larger recordsize for sequential access, metadata caching, and enhanced prefetch
+zfs_blockchain_options="-o recordsize=1M -o prefetch=1 -o logbias=latency"
+
 zfs create $zfs_common_options $zfs_compressed_options "${ZPOOL}/${TOP_LEVEL_DATASET}"
 
 # create an uncompressed dataset for the blockchain.  Blocks in it are already compressed, so won't compress further.
-zfs create $zfs_common_options $zfs_uncompressed_options "${ZPOOL}/${TOP_LEVEL_DATASET}/blockchain"
+# Optimized with larger recordsize for sequential access, metadata-only caching, and enhanced prefetch for large files
+zfs create $zfs_common_options $zfs_uncompressed_options $zfs_blockchain_options "${ZPOOL}/${TOP_LEVEL_DATASET}/blockchain"
+
+# create a dataset for state-related data
+zfs create $zfs_common_options $zfs_compressed_options "${ZPOOL}/${TOP_LEVEL_DATASET}/state"
 
 # create an uncompressed dataset for the shared_memory.bin file and WAL.
-# AFAIK we haven't done studies on whether compression helps shared_memory.bin.
-zfs create $zfs_common_options $zfs_uncompressed_options "${ZPOOL}/${TOP_LEVEL_DATASET}/shared_memory"
+# Optimized for memory-mapped files with reduced latency and tuned for random access patterns
+zfs create $zfs_common_options $zfs_uncompressed_options $zfs_shared_memory_options "${ZPOOL}/${TOP_LEVEL_DATASET}/state/shared_memory"
+
+# create an optimized dataset for RocksDB with settings tuned for LSM-tree databases:
+# - smaller recordsize (16k) for better random I/O performance
+# - full caching for both data and metadata to improve read performance
+# - logbias=throughput to optimize for write-heavy workloads with compaction
+# - redundant_metadata=most for better resilience during power loss
+zfs create $zfs_common_options $zfs_uncompressed_options $zfs_rocksdb_options "${ZPOOL}/${TOP_LEVEL_DATASET}/state/rocksdb"
 
 # create an unmountable dataset to serve as the parent for pgdata & tablespaces
 zfs create $zfs_common_options $zfs_compressed_options -o canmount=off "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store"
@@ -116,7 +134,7 @@ zfs create $zfs_common_options $zfs_compressed_options $zfs_postgres_options -o
 zfs create $zfs_common_options $zfs_compressed_options -o canmount=on "${ZPOOL}/${TOP_LEVEL_DATASET}/logs"
 
 # needs to exist to be bind-mounted, no real reason to make it a dataset of its own though
-mkdir -p "$TOP_LEVEL_DATASET_MOUNTPOINT/shared_memory/haf_wal"
+mkdir -p "$TOP_LEVEL_DATASET_MOUNTPOINT/state/haf_wal"
 
 # 1000:100 is hived:users inside the container
 chown -R 1000:100 "$TOP_LEVEL_DATASET_MOUNTPOINT"
diff --git a/haf_base.yaml b/haf_base.yaml
index 1b86cd666227f7645cebde4d4235f60f39a95f8f..6d7688c46da9563c2c9f9520a686c8998a71219b 100644
--- a/haf_base.yaml
+++ b/haf_base.yaml
@@ -4,6 +4,11 @@ services:
       - core
       - block-explorer-uninstall
     image: ${HAF_IMAGE:-${HIVE_API_NODE_REGISTRY:-registry.hive.blog}/haf}:${HAF_VERSION:-${HIVE_API_NODE_VERSION}}
+    privileged: true
+    cap_add: [SYS_ADMIN]
+    security_opt:
+       - apparmor=unconfined
+       - label=disable
     networks:
       - haf-network
     tty: true
@@ -102,11 +107,18 @@ services:
         source: ${HAF_LOG_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/logs}/postgresql
         target: /home/hived/postgresql_logs/
       - type: bind
-        source: ${HAF_SHM_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/shared_memory}/
-        target: /home/hived/shm_dir
+        source: ${HAF_SHM_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/state/shared_memory}/
+        target: /home/hived/state/shm_dir
       - type: bind
-        source: ${HAF_WAL_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/shared_memory/haf_wal}
-        target: /home/hived/wal_dir
+        source: ${HAF_WAL_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/state/haf_wal}
+        target: /home/hived/state/wal_dir
+      - type: bind
+        source: ${HAF_WAL_DIRECTORY:-${TOP_LEVEL_DATASET_MOUNTPOINT}/state/rocksdb}
+        target: /home/hived/state/rocksdb
+      - type: bind
+        source: /proc/sys/vm
+        target: /host-proc/sys/vm
+        read_only: false
     shm_size: 8gb
     healthcheck:
       test: ["CMD-SHELL","curl -f 127.0.0.1:8091"]
diff --git a/make_ramdisk.sh b/make_ramdisk.sh
index cbcd3b51164f0c91ce01dd6bdfef2366546e3949..8f574913b718a9cd043b0b0b739c4d5dd187bed3 100755
--- a/make_ramdisk.sh
+++ b/make_ramdisk.sh
@@ -1,4 +1,4 @@
 #another braindead script for me
 sudo mkdir -p /mnt/haf_shared_mem
-sudo mount -t tmpfs -o size=26G ramfs /mnt/haf_shared_mem
+sudo mount -t tmpfs -o size=12G ramfs /mnt/haf_shared_mem
 sudo chmod 777 /mnt/haf_shared_mem
diff --git a/rollback_zfs_datasets.sh b/rollback_zfs_datasets.sh
index f2360fc80f76c1ffb318db3d536bc808c63b5ec7..d9dedafc5c23ea391eb2176257c136a4dc48232c 100755
--- a/rollback_zfs_datasets.sh
+++ b/rollback_zfs_datasets.sh
@@ -94,7 +94,9 @@ check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/haf_db_store/pgdat
 check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/haf_db_store/tablespace"
 check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/logs"
 check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/blockchain"
-check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/shared_memory"
+check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/state/shared_memory"
+check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/state/rocksdb"
+check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/state"
 check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}"
 echo "All datasets appear unmountable"
 
@@ -128,7 +130,9 @@ unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store/pgdata"
 unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store/tablespace"
 unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/logs"
 unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/blockchain"
-unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/shared_memory"
+unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/state/shared_memory"
+unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/state/rocksdb"
+unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/state"
 unmount "${ZPOOL}/${TOP_LEVEL_DATASET}"
 
 rollback() {
@@ -138,7 +142,9 @@ rollback() {
 }
 
 rollback "${ZPOOL}/${TOP_LEVEL_DATASET}"
-rollback "${ZPOOL}/${TOP_LEVEL_DATASET}/shared_memory"
+rollback "${ZPOOL}/${TOP_LEVEL_DATASET}/state"
+rollback "${ZPOOL}/${TOP_LEVEL_DATASET}/state/shared_memory"
+rollback "${ZPOOL}/${TOP_LEVEL_DATASET}/state/rocksdb"
 rollback "${ZPOOL}/${TOP_LEVEL_DATASET}/blockchain"
 rollback "${ZPOOL}/${TOP_LEVEL_DATASET}/logs"
 rollback "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store/tablespace"
@@ -152,7 +158,9 @@ remount() {
 }
 
 remount "${ZPOOL}/${TOP_LEVEL_DATASET}"
-remount "${ZPOOL}/${TOP_LEVEL_DATASET}/shared_memory"
+remount "${ZPOOL}/${TOP_LEVEL_DATASET}/state"
+remount "${ZPOOL}/${TOP_LEVEL_DATASET}/state/shared_memory"
+remount "${ZPOOL}/${TOP_LEVEL_DATASET}/state/rocksdb"
 remount "${ZPOOL}/${TOP_LEVEL_DATASET}/blockchain"
 remount "${ZPOOL}/${TOP_LEVEL_DATASET}/logs"
 remount "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store/tablespace"
@@ -160,7 +168,9 @@ remount "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store/pgdata"
 remount "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store/pgdata/pg_wal"
 
 zfs list "${ZPOOL}/${TOP_LEVEL_DATASET}" \
-         "${ZPOOL}/${TOP_LEVEL_DATASET}/shared_memory" \
+         "${ZPOOL}/${TOP_LEVEL_DATASET}/state" \
+         "${ZPOOL}/${TOP_LEVEL_DATASET}/state/shared_memory" \
+         "${ZPOOL}/${TOP_LEVEL_DATASET}/state/rocksdb" \
          "${ZPOOL}/${TOP_LEVEL_DATASET}/blockchain" \
          "${ZPOOL}/${TOP_LEVEL_DATASET}/logs" \
          "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store/tablespace" \
diff --git a/snapshot_zfs_datasets.sh b/snapshot_zfs_datasets.sh
index 8626b8d6d1260ce21724191962111f4d7ef8ce3c..e06a98a1201bf65a6b29c75eb63fa7785bce4e9a 100755
--- a/snapshot_zfs_datasets.sh
+++ b/snapshot_zfs_datasets.sh
@@ -132,7 +132,9 @@ check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/haf_db_store/pgdat
 check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/haf_db_store/tablespace"
 check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/logs"
 check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/blockchain"
-check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/shared_memory"
+check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/state/shared_memory"
+check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/state/rocksdb"
+check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}/state"
 check_dataset_is_unmountable "${TOP_LEVEL_DATASET_MOUNTPOINT}"
 if [ ! -z "${SWAP_LOGS_DATASET}" ]; then
   check_dataset_is_unmountable "${SWAP_LOGS_DATASET}"
@@ -141,12 +143,12 @@ fi
 echo "All datasets appear unmountable"
 
 if [ "$SNAPSHOT_NAME" != "empty" ]; then
-  if [ ! -e "${TOP_LEVEL_DATASET_MOUNTPOINT}/shared_memory/shared_memory.bin" ]; then
-    echo "Warning: No shared memory file found in the shared_memory directory"
+  if [ ! -e "${TOP_LEVEL_DATASET_MOUNTPOINT}/state/shared_memory/shared_memory.bin" ]; then
+    echo "Warning: No shared memory file found in the state/shared_memory directory"
     exit 1
   fi
 
-  last_shared_memory_write=$(stat -c %Y "${TOP_LEVEL_DATASET_MOUNTPOINT}/shared_memory/shared_memory.bin")
+  last_shared_memory_write=$(stat -c %Y "${TOP_LEVEL_DATASET_MOUNTPOINT}/state/shared_memory/shared_memory.bin")
   last_blockchain_write=$(find "${TOP_LEVEL_DATASET_MOUNTPOINT}/blockchain" -type f -printf '%T@\n' | sort -n | tail -1 | cut -d. -f1)
 
   if [ -z "$last_blockchain_write" ]; then
@@ -213,7 +215,9 @@ unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store/pgdata"
 unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store/tablespace"
 unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/logs"
 unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/blockchain"
-unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/shared_memory"
+unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/state/shared_memory"
+unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/state/rocksdb"
+unmount "${ZPOOL}/${TOP_LEVEL_DATASET}/state"
 unmount "${ZPOOL}/${TOP_LEVEL_DATASET}"
 
 if [ ! -z "${SWAP_LOGS_DATASET}" ]; then
@@ -241,7 +245,9 @@ remount() {
 }
 
 remount "${ZPOOL}/${TOP_LEVEL_DATASET}"
-remount "${ZPOOL}/${TOP_LEVEL_DATASET}/shared_memory"
+remount "${ZPOOL}/${TOP_LEVEL_DATASET}/state"
+remount "${ZPOOL}/${TOP_LEVEL_DATASET}/state/shared_memory"
+remount "${ZPOOL}/${TOP_LEVEL_DATASET}/state/rocksdb"
 remount "${ZPOOL}/${TOP_LEVEL_DATASET}/blockchain"
 remount "${ZPOOL}/${TOP_LEVEL_DATASET}/logs"
 remount "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store/tablespace"
@@ -259,7 +265,9 @@ fi
 
 
 zfs list "${ZPOOL}/${TOP_LEVEL_DATASET}@${SNAPSHOT_NAME}" \
-         "${ZPOOL}/${TOP_LEVEL_DATASET}/shared_memory@${SNAPSHOT_NAME}" \
+         "${ZPOOL}/${TOP_LEVEL_DATASET}/state@${SNAPSHOT_NAME}" \
+         "${ZPOOL}/${TOP_LEVEL_DATASET}/state/shared_memory@${SNAPSHOT_NAME}" \
+         "${ZPOOL}/${TOP_LEVEL_DATASET}/state/rocksdb@${SNAPSHOT_NAME}" \
          "${ZPOOL}/${TOP_LEVEL_DATASET}/blockchain@${SNAPSHOT_NAME}" \
          "${SWAP_LOGS_DATASET:-${ZPOOL}/${TOP_LEVEL_DATASET}/logs}@${SNAPSHOT_NAME}" \
          "${ZPOOL}/${TOP_LEVEL_DATASET}/haf_db_store/tablespace@${SNAPSHOT_NAME}" \