diff --git a/node/Bond.cpp b/node/Bond.cpp
index 70d0e5c5a..3f65c0fe4 100644
--- a/node/Bond.cpp
+++ b/node/Bond.cpp
@@ -618,7 +618,7 @@ void Bond::sendPATH_NEGOTIATION_REQUEST(void *tPtr, const SharedPtr<Path> &path)
 	Packet outp(_peer->_id.address(),RR->identity.address(),Packet::VERB_PATH_NEGOTIATION_REQUEST);
 	outp.append<int16_t>(_localUtility);
 	if (path->address()) {
-		outp.armor(_peer->key(),false);
+		outp.armor(_peer->key(),false,_peer->aesKeysIfSupported());
 		RR->node->putPacket(tPtr,path->localSocket(),path->address(),outp.data(),outp.size());
 	}
 }
@@ -639,7 +639,7 @@ void Bond::sendACK(void *tPtr, const SharedPtr<Path> &path,const int64_t localSo
 	//RR->t->bondStateMessage(NULL, traceMsg);
 	outp.append<uint32_t>(bytesToAck);
 	if (atAddress) {
-		outp.armor(_peer->key(),false);
+		outp.armor(_peer->key(),false,_peer->aesKeysIfSupported());
 		RR->node->putPacket(tPtr,localSocket,atAddress,outp.data(),outp.size());
 	} else {
 		RR->sw->send(tPtr,outp,false);
@@ -662,7 +662,7 @@ void Bond::sendQOS_MEASUREMENT(void *tPtr,const SharedPtr<Path> &path,const int6
 	int16_t len = generateQoSPacket(path, _now,qosData);
 	outp.append(qosData,len);
 	if (atAddress) {
-		outp.armor(_peer->key(),false);
+		outp.armor(_peer->key(),false,_peer->aesKeysIfSupported());
 		RR->node->putPacket(tPtr,localSocket,atAddress,outp.data(),outp.size());
 	} else {
 		RR->sw->send(tPtr,outp,false);
diff --git a/node/C25519.cpp b/node/C25519.cpp
index d7c7011fd..aec34835e 100644
--- a/node/C25519.cpp
+++ b/node/C25519.cpp
@@ -2439,7 +2439,7 @@ static inline void get_hram(unsigned char *hram, const unsigned char *sm, const
 	for (i = 32;i < 64;++i)    playground[i] = pk[i-32];
 	for (i = 64;i < smlen;++i) playground[i] = sm[i];
 
-	ZeroTier::SHA512::hash(hram,playground,(unsigned int)smlen);
+	ZeroTier::SHA512(hram,playground,(unsigned int)smlen);
 }
 
 //////////////////////////////////////////////////////////////////////////////
@@ -2459,11 +2459,11 @@ void C25519::agree(const C25519::Private &mine,const C25519::Public &their,void
 	unsigned char digest[64];
 
 	crypto_scalarmult(rawkey,mine.data,their.data);
-	SHA512::hash(digest,rawkey,32);
+	SHA512(digest,rawkey,32);
 	for(unsigned int i=0,k=0;i<keylen;) {
 		if (k == 64) {
 			k = 0;
-			SHA512::hash(digest,digest,64);
+			SHA512(digest,digest,64);
 		}
 		((unsigned char *)keybuf)[i++] = digest[k++];
 	}
@@ -2472,7 +2472,7 @@ void C25519::agree(const C25519::Private &mine,const C25519::Public &their,void
 void C25519::sign(const C25519::Private &myPrivate,const C25519::Public &myPublic,const void *msg,unsigned int len,void *signature)
 {
 	unsigned char digest[64]; // we sign the first 32 bytes of SHA-512(msg)
-	SHA512::hash(digest,msg,len);
+	SHA512(digest,msg,len);
 
 #ifdef ZT_USE_FAST_X64_ED25519
 	ed25519_amd64_asm_sign(myPrivate.data + 32,myPublic.data + 32,digest,(unsigned char *)signature);
@@ -2486,7 +2486,7 @@ void C25519::sign(const C25519::Private &myPrivate,const C25519::Public &myPubli
 	unsigned char hram[crypto_hash_sha512_BYTES];
 	unsigned char *sig = (unsigned char *)signature;
 
-	SHA512::hash(extsk,myPrivate.data + 32,32);
+	SHA512(extsk,myPrivate.data + 32,32);
 	extsk[0] &= 248;
 	extsk[31] &= 127;
 	extsk[31] |= 64;
@@ -2496,7 +2496,7 @@ void C25519::sign(const C25519::Private &myPrivate,const C25519::Public &myPubli
 	for(unsigned int i=0;i<32;i++)
 		sig[64 + i] = digest[i];
 
-	SHA512::hash(hmg,sig + 32,64);
+	SHA512(hmg,sig + 32,64);
 
 	/* Computation of R */
 	sc25519_from64bytes(&sck, hmg);
@@ -2525,7 +2525,7 @@ bool C25519::verify(const C25519::Public &their,const void *msg,unsigned int len
 {
 	const unsigned char *const sig = (const unsigned char *)signature;
 	unsigned char digest[64]; // we sign the first 32 bytes of SHA-512(msg)
-	SHA512::hash(digest,msg,len);
+	SHA512(digest,msg,len);
 	if (!Utils::secureEq(sig + 64,digest,32))
 		return false;
 
@@ -2565,7 +2565,7 @@ void C25519::_calcPubED(C25519::Pair &kp)
 
 	// Second 32 bytes of pub and priv are the keys for ed25519
 	// signing and verification.
-	SHA512::hash(extsk,kp.priv.data + 32,32);
+	SHA512(extsk,kp.priv.data + 32,32);
 	extsk[0] &= 248;
 	extsk[31] &= 127;
 	extsk[31] |= 64;
diff --git a/node/Constants.hpp b/node/Constants.hpp
index d9aebe65b..5d6e735c4 100644
--- a/node/Constants.hpp
+++ b/node/Constants.hpp
@@ -174,6 +174,8 @@
  */
 #define ZT_ADDRESS_LENGTH_HEX 10
 
+#define ZT_SYMMETRIC_KEY_SIZE 48
+
 /**
  * Addresses beginning with this byte are reserved for the joy of in-band signaling
  */
diff --git a/node/Identity.cpp b/node/Identity.cpp
index 79ec31453..248a9f241 100644
--- a/node/Identity.cpp
+++ b/node/Identity.cpp
@@ -34,7 +34,7 @@ namespace ZeroTier {
 static inline void _computeMemoryHardHash(const void *publicKey,unsigned int publicKeyBytes,void *digest,void *genmem)
 {
 	// Digest publicKey[] to obtain initial digest
-	SHA512::hash(digest,publicKey,publicKeyBytes);
+	SHA512(digest,publicKey,publicKeyBytes);
 
 	// Initialize genmem[] using Salsa20 in a CBC-like configuration since
 	// ordinary Salsa20 is randomly seek-able. This is good for a cipher
diff --git a/node/Identity.hpp b/node/Identity.hpp
index a754b6727..fc7ac55fa 100644
--- a/node/Identity.hpp
+++ b/node/Identity.hpp
@@ -118,7 +118,7 @@ public:
 	inline bool sha512PrivateKey(void *sha) const
 	{
 		if (_privateKey) {
-			SHA512::hash(sha,_privateKey->data,ZT_C25519_PRIVATE_KEY_LEN);
+			SHA512(sha,_privateKey->data,ZT_C25519_PRIVATE_KEY_LEN);
 			return true;
 		}
 		return false;
diff --git a/node/IncomingPacket.cpp b/node/IncomingPacket.cpp
index 94f7f6301..9ff6edd58 100644
--- a/node/IncomingPacket.cpp
+++ b/node/IncomingPacket.cpp
@@ -294,7 +294,7 @@ bool IncomingPacket::_doHELLO(const RuntimeEnvironment *RR,void *tPtr,const bool
 						outp.append((uint8_t)Packet::VERB_HELLO);
 						outp.append((uint64_t)pid);
 						outp.append((uint8_t)Packet::ERROR_IDENTITY_COLLISION);
-						outp.armor(key,true);
+						outp.armor(key,true,peer->aesKeysIfSupported());
 						_path->send(RR,tPtr,outp.data(),outp.size(),RR->node->now());
 					} else {
 						RR->t->incomingPacketMessageAuthenticationFailure(tPtr,_path,pid,fromAddress,hops(),"invalid MAC");
@@ -444,8 +444,8 @@ bool IncomingPacket::_doHELLO(const RuntimeEnvironment *RR,void *tPtr,const bool
 	}
 	outp.setAt<uint16_t>(worldUpdateSizeAt,(uint16_t)(outp.size() - (worldUpdateSizeAt + 2)));
 
+	outp.armor(peer->key(),true,peer->aesKeysIfSupported());
 	peer->recordOutgoingPacket(_path,outp.packetId(),outp.payloadLength(),outp.verb(),ZT_QOS_NO_FLOW,now);
-	outp.armor(peer->key(),true);
 	_path->send(RR,tPtr,outp.data(),outp.size(),now);
 
 	peer->setRemoteVersion(protoVersion,vMajor,vMinor,vRevision); // important for this to go first so received() knows the version
@@ -593,7 +593,7 @@ bool IncomingPacket::_doWHOIS(const RuntimeEnvironment *RR,void *tPtr,const Shar
 	}
 
 	if (count > 0) {
-		outp.armor(peer->key(),true);
+		outp.armor(peer->key(),true,peer->aesKeysIfSupported());
 		_path->send(RR,tPtr,outp.data(),outp.size(),RR->node->now());
 	}
 
@@ -816,8 +816,8 @@ bool IncomingPacket::_doEXT_FRAME(const RuntimeEnvironment *RR,void *tPtr,const
 			outp.append((uint64_t)packetId());
 			outp.append((uint64_t)nwid);
 			const int64_t now = RR->node->now();
+			outp.armor(peer->key(),true,peer->aesKeysIfSupported());
 			peer->recordOutgoingPacket(_path,outp.packetId(),outp.payloadLength(),outp.verb(),ZT_QOS_NO_FLOW,now);
-			outp.armor(peer->key(),true);
 			_path->send(RR,tPtr,outp.data(),outp.size(),RR->node->now());
 		}
 
@@ -842,8 +842,8 @@ bool IncomingPacket::_doECHO(const RuntimeEnvironment *RR,void *tPtr,const Share
 	outp.append((uint64_t)pid);
 	if (size() > ZT_PACKET_IDX_PAYLOAD)
 		outp.append(reinterpret_cast<const unsigned char *>(data()) + ZT_PACKET_IDX_PAYLOAD,size() - ZT_PACKET_IDX_PAYLOAD);
+	outp.armor(peer->key(),true,peer->aesKeysIfSupported());
 	peer->recordOutgoingPacket(_path,outp.packetId(),outp.payloadLength(),outp.verb(),ZT_QOS_NO_FLOW,now);
-	outp.armor(peer->key(),true);
 	_path->send(RR,tPtr,outp.data(),outp.size(),RR->node->now());
 
 	peer->received(tPtr,_path,hops(),pid,payloadLength(),Packet::VERB_ECHO,0,Packet::VERB_NOP,false,0,ZT_QOS_NO_FLOW);
@@ -1016,7 +1016,7 @@ bool IncomingPacket::_doNETWORK_CONFIG_REQUEST(const RuntimeEnvironment *RR,void
 		outp.append(requestPacketId);
 		outp.append((unsigned char)Packet::ERROR_UNSUPPORTED_OPERATION);
 		outp.append(nwid);
-		outp.armor(peer->key(),true);
+		outp.armor(peer->key(),true,peer->aesKeysIfSupported());
 		_path->send(RR,tPtr,outp.data(),outp.size(),RR->node->now());
 	}
 
@@ -1037,8 +1037,8 @@ bool IncomingPacket::_doNETWORK_CONFIG(const RuntimeEnvironment *RR,void *tPtr,c
 			outp.append((uint64_t)network->id());
 			outp.append((uint64_t)configUpdateId);
 			const int64_t now = RR->node->now();
+			outp.armor(peer->key(),true,peer->aesKeysIfSupported());
 			peer->recordOutgoingPacket(_path,outp.packetId(),outp.payloadLength(),outp.verb(),ZT_QOS_NO_FLOW,now);
-			outp.armor(peer->key(),true);
 			_path->send(RR,tPtr,outp.data(),outp.size(),RR->node->now());
 		}
 	}
@@ -1086,8 +1086,8 @@ bool IncomingPacket::_doMULTICAST_GATHER(const RuntimeEnvironment *RR,void *tPtr
 		outp.append((uint32_t)mg.adi());
 		const unsigned int gatheredLocally = RR->mc->gather(peer->address(),nwid,mg,outp,gatherLimit);
 		if (gatheredLocally > 0) {
+			outp.armor(peer->key(),true,peer->aesKeysIfSupported());
 			peer->recordOutgoingPacket(_path,outp.packetId(),outp.payloadLength(),outp.verb(),ZT_QOS_NO_FLOW,now);
-			outp.armor(peer->key(),true);
 			_path->send(RR,tPtr,outp.data(),outp.size(),now);
 		}
 	}
@@ -1185,8 +1185,8 @@ bool IncomingPacket::_doMULTICAST_FRAME(const RuntimeEnvironment *RR,void *tPtr,
 			outp.append((unsigned char)0x02); // flag 0x02 = contains gather results
 			if (RR->mc->gather(peer->address(),nwid,to,outp,gatherLimit)) {
 				const int64_t now = RR->node->now();
+				outp.armor(peer->key(),true,peer->aesKeysIfSupported());
 				peer->recordOutgoingPacket(_path,outp.packetId(),outp.payloadLength(),outp.verb(),ZT_QOS_NO_FLOW,now);
-				outp.armor(peer->key(),true);
 				_path->send(RR,tPtr,outp.data(),outp.size(),RR->node->now());
 			}
 		}
@@ -1327,7 +1327,7 @@ void IncomingPacket::_sendErrorNeedCredentials(const RuntimeEnvironment *RR,void
 	outp.append(packetId());
 	outp.append((uint8_t)Packet::ERROR_NEED_MEMBERSHIP_CERTIFICATE);
 	outp.append(nwid);
-	outp.armor(peer->key(),true);
+	outp.armor(peer->key(),true,peer->aesKeysIfSupported());
 	_path->send(RR,tPtr,outp.data(),outp.size(),RR->node->now());
 }
 
diff --git a/node/Multicaster.cpp b/node/Multicaster.cpp
index 6b9bdd374..3b48b799e 100644
--- a/node/Multicaster.cpp
+++ b/node/Multicaster.cpp
@@ -194,7 +194,7 @@ void Multicaster::send(
 					outp.append((uint16_t)etherType);
 					outp.append(data,len);
 					if (!network->config().disableCompression()) outp.compress();
-					outp.armor(bestMulticastReplicator->key(),true);
+					outp.armor(bestMulticastReplicator->key(),true,bestMulticastReplicator->aesKeysIfSupported());
 					bestMulticastReplicatorPath->send(RR,tPtr,outp.data(),outp.size(),now);
 					return;
 				}
diff --git a/node/Packet.cpp b/node/Packet.cpp
index 395c0e192..04d8f92e5 100644
--- a/node/Packet.cpp
+++ b/node/Packet.cpp
@@ -876,39 +876,61 @@ static inline int LZ4_decompress_safe(const char* source, char* dest, int compre
 
 const unsigned char Packet::ZERO_KEY[32] = { 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0 };
 
-void Packet::armor(const void *key,bool encryptPayload)
+void Packet::armor(const void *key,bool encryptPayload,const AES aesKeys[2])
 {
-	uint8_t mangledKey[32];
 	uint8_t *const data = reinterpret_cast<uint8_t *>(unsafeData());
+	if ((aesKeys) && (encryptPayload)) {
+		setCipher(ZT_PROTO_CIPHER_SUITE__AES_GMAC_SIV);
 
-	// Set flag now, since it affects key mangle function
-	setCipher(encryptPayload ? ZT_PROTO_CIPHER_SUITE__C25519_POLY1305_SALSA2012 : ZT_PROTO_CIPHER_SUITE__C25519_POLY1305_NONE);
-
-	_salsa20MangleKey((const unsigned char *)key,mangledKey);
-
-	if (ZT_HAS_FAST_CRYPTO()) {
-		const unsigned int encryptLen = (encryptPayload) ? (size() - ZT_PACKET_IDX_VERB) : 0;
-		uint64_t keyStream[(ZT_PROTO_MAX_PACKET_LENGTH + 64 + 8) / 8];
-		ZT_FAST_SINGLE_PASS_SALSA2012(keyStream,encryptLen + 64,(data + ZT_PACKET_IDX_IV),mangledKey);
-		Salsa20::memxor(data + ZT_PACKET_IDX_VERB,reinterpret_cast<const uint8_t *>(keyStream + 8),encryptLen);
-		uint64_t mac[2];
-		Poly1305::compute(mac,data + ZT_PACKET_IDX_VERB,size() - ZT_PACKET_IDX_VERB,keyStream);
-#ifdef ZT_NO_TYPE_PUNNING
-		memcpy(data + ZT_PACKET_IDX_MAC,mac,8);
-#else
-		(*reinterpret_cast<uint64_t *>(data + ZT_PACKET_IDX_MAC)) = mac[0];
-#endif
-	} else {
-		Salsa20 s20(mangledKey,data + ZT_PACKET_IDX_IV);
-		uint64_t macKey[4];
-		s20.crypt12(ZERO_KEY,macKey,sizeof(macKey));
 		uint8_t *const payload = data + ZT_PACKET_IDX_VERB;
 		const unsigned int payloadLen = size() - ZT_PACKET_IDX_VERB;
-		if (encryptPayload)
-			s20.crypt12(payload,payload,payloadLen);
-		uint64_t mac[2];
-		Poly1305::compute(mac,payload,payloadLen,macKey);
-		memcpy(data + ZT_PACKET_IDX_MAC,mac,8);
+
+		AES::GMACSIVEncryptor enc(aesKeys[0],aesKeys[1]);
+		enc.init(Utils::loadMachineEndian<uint64_t>(data + ZT_PACKET_IDX_IV),payload);
+		enc.aad(data + ZT_PACKET_IDX_DEST,11);
+		enc.update1(payload,payloadLen);
+		enc.finish1();
+		enc.update2(payload,payloadLen);
+		const uint64_t *const tag = enc.finish2();
+
+#ifdef ZT_NO_UNALIGNED_ACCESS
+		Utils::copy<8>(data,tag);
+		Utils::copy<8>(data + ZT_PACKET_IDX_MAC,tag + 1);
+#else
+		*reinterpret_cast<uint64_t *>(data) = tag[0];
+		*reinterpret_cast<uint64_t *>(data + ZT_PACKET_IDX_MAC) = tag[1];
+#endif
+	} else {
+		uint8_t mangledKey[32];
+		setCipher(encryptPayload ? ZT_PROTO_CIPHER_SUITE__C25519_POLY1305_SALSA2012 : ZT_PROTO_CIPHER_SUITE__C25519_POLY1305_NONE);
+		_salsa20MangleKey((const unsigned char *)key,mangledKey);
+		if (ZT_HAS_FAST_CRYPTO()) {
+			const unsigned int encryptLen = (encryptPayload) ? (size() - ZT_PACKET_IDX_VERB) : 0;
+			uint64_t keyStream[(ZT_PROTO_MAX_PACKET_LENGTH + 64 + 8) / 8];
+			ZT_FAST_SINGLE_PASS_SALSA2012(keyStream,encryptLen + 64,(data + ZT_PACKET_IDX_IV),mangledKey);
+			Salsa20::memxor(data + ZT_PACKET_IDX_VERB,reinterpret_cast<const uint8_t *>(keyStream + 8),encryptLen);
+			uint64_t mac[2];
+			Poly1305::compute(mac,data + ZT_PACKET_IDX_VERB,size() - ZT_PACKET_IDX_VERB,keyStream);
+#ifdef ZT_NO_TYPE_PUNNING
+			memcpy(data + ZT_PACKET_IDX_MAC,mac,8);
+#else
+			(*reinterpret_cast<uint64_t *>(data + ZT_PACKET_IDX_MAC)) = mac[0];
+#endif
+		} else {
+			Salsa20 s20(mangledKey,data + ZT_PACKET_IDX_IV);
+
+			uint64_t macKey[4];
+			s20.crypt12(ZERO_KEY,macKey,sizeof(macKey));
+
+			uint8_t *const payload = data + ZT_PACKET_IDX_VERB;
+			const unsigned int payloadLen = size() - ZT_PACKET_IDX_VERB;
+			if (encryptPayload)
+				s20.crypt12(payload,payload,payloadLen);
+			uint64_t mac[2];
+
+			Poly1305::compute(mac,payload,payloadLen,macKey);
+			memcpy(data + ZT_PACKET_IDX_MAC,mac,8);
+		}
 	}
 }
 
diff --git a/node/Packet.hpp b/node/Packet.hpp
index 1ffca5a5c..9089e5264 100644
--- a/node/Packet.hpp
+++ b/node/Packet.hpp
@@ -26,6 +26,7 @@
 #include "Address.hpp"
 #include "Poly1305.hpp"
 #include "Salsa20.hpp"
+#include "AES.hpp"
 #include "Utils.hpp"
 #include "Buffer.hpp"
 
@@ -55,10 +56,12 @@
  *    + Tags and Capabilities
  *    + Inline push of CertificateOfMembership deprecated
  * 9  - 1.2.0 ... 1.2.14
- * 10 - 1.4.0 ... CURRENT
- *    + Multipath capability and load balancing (tentative)
+ * 10 - 1.4.0 ... 1.4.6
+ * 11 - 1.4.8 ... end of 1.4 series
+ *    + Multipath capability and load balancing (beta)
+ *    + AES-GMAC-SIV backported for faster peer-to-peer crypto
  */
-#define ZT_PROTO_VERSION 10
+#define ZT_PROTO_VERSION 11
 
 /**
  * Minimum supported protocol version
@@ -96,6 +99,21 @@
  */
 #define ZT_PROTO_CIPHER_SUITE__C25519_POLY1305_SALSA2012 1
 
+/**
+ * AES-GMAC-SIV backported from 2.x
+ */
+#define ZT_PROTO_CIPHER_SUITE__AES_GMAC_SIV 3
+
+/**
+ * AES-GMAC-SIV first of two keys
+ */
+#define ZT_KBKDF_LABEL_AES_GMAC_SIV_K0 '0'
+
+/**
+ * AES-GMAC-SIV second of two keys
+ */
+#define ZT_KBKDF_LABEL_AES_GMAC_SIV_K1 '1'
+
 /**
  * Cipher suite: NONE
  *
@@ -1295,8 +1313,9 @@ public:
 	 *
 	 * @param key 32-byte key
 	 * @param encryptPayload If true, encrypt packet payload, else just MAC
+	 * @param aes Use new AES-GMAC-SIV constrution
 	 */
-	void armor(const void *key,bool encryptPayload);
+	void armor(const void *key,bool encryptPayload,const AES aesKeys[2]);
 
 	/**
 	 * Verify and (if encrypted) decrypt packet
diff --git a/node/Peer.cpp b/node/Peer.cpp
index bf39c7d14..b2b04c17f 100644
--- a/node/Peer.cpp
+++ b/node/Peer.cpp
@@ -61,6 +61,13 @@ Peer::Peer(const RuntimeEnvironment *renv,const Identity &myIdentity,const Ident
 	if (!myIdentity.agree(peerIdentity,_key,ZT_PEER_SECRET_KEY_LENGTH)) {
 		throw ZT_EXCEPTION_INVALID_ARGUMENT;
 	}
+
+	uint8_t ktmp[32];
+	KBKDFHMACSHA384(_key,ZT_KBKDF_LABEL_AES_GMAC_SIV_K0,0,0,ktmp);
+	_aesKeys[0].init(ktmp);
+	KBKDFHMACSHA384(_key,ZT_KBKDF_LABEL_AES_GMAC_SIV_K1,0,0,ktmp);
+	_aesKeys[0].init(ktmp);
+	Utils::burn(ktmp, 32);
 }
 
 void Peer::received(
@@ -209,7 +216,7 @@ void Peer::received(
 					if (count) {
 						outp->setAt(ZT_PACKET_IDX_PAYLOAD,(uint16_t)count);
 						outp->compress();
-						outp->armor(_key,true);
+						outp->armor(_key,true,aesKeysIfSupported());
 						path->send(RR,tPtr,outp->data(),outp->size(),now);
 					}
 					delete outp;
@@ -347,7 +354,7 @@ void Peer::introduce(void *const tPtr,const int64_t now,const SharedPtr<Peer> &o
 					outp.append((uint8_t)4);
 					outp.append(other->_paths[theirs].p->address().rawIpData(),4);
 				}
-				outp.armor(_key,true);
+				outp.armor(_key,true,aesKeysIfSupported());
 				_paths[mine].p->send(RR,tPtr,outp.data(),outp.size(),now);
 			} else {
 				Packet outp(other->_id.address(),RR->identity.address(),Packet::VERB_RENDEZVOUS);
@@ -361,7 +368,7 @@ void Peer::introduce(void *const tPtr,const int64_t now,const SharedPtr<Peer> &o
 					outp.append((uint8_t)4);
 					outp.append(_paths[mine].p->address().rawIpData(),4);
 				}
-				outp.armor(other->_key,true);
+				outp.armor(other->_key,true,aesKeysIfSupported());
 				other->_paths[theirs].p->send(RR,tPtr,outp.data(),outp.size(),now);
 			}
 			++alt;
@@ -402,12 +409,12 @@ void Peer::sendHELLO(void *tPtr,const int64_t localSocket,const InetAddress &atA
 
 	outp.cryptField(_key,startCryptedPortionAt,outp.size() - startCryptedPortionAt);
 
-	RR->node->expectReplyTo(outp.packetId());
-
 	if (atAddress) {
-		outp.armor(_key,false); // false == don't encrypt full payload, but add MAC
+		outp.armor(_key,false,aesKeysIfSupported()); // false == don't encrypt full payload, but add MAC
+		RR->node->expectReplyTo(outp.packetId());
 		RR->node->putPacket(tPtr,localSocket,atAddress,outp.data(),outp.size());
 	} else {
+		RR->node->expectReplyTo(outp.packetId());
 		RR->sw->send(tPtr,outp,false); // false == don't encrypt full payload, but add MAC
 	}
 }
@@ -416,8 +423,8 @@ void Peer::attemptToContactAt(void *tPtr,const int64_t localSocket,const InetAdd
 {
 	if ( (!sendFullHello) && (_vProto >= 5) && (!((_vMajor == 1)&&(_vMinor == 1)&&(_vRevision == 0))) ) {
 		Packet outp(_id.address(),RR->identity.address(),Packet::VERB_ECHO);
+		outp.armor(_key,true,aesKeysIfSupported());
 		RR->node->expectReplyTo(outp.packetId());
-		outp.armor(_key,true);
 		RR->node->putPacket(tPtr,localSocket,atAddress,outp.data(),outp.size());
 	} else {
 		sendHELLO(tPtr,localSocket,atAddress,now);
diff --git a/node/Peer.hpp b/node/Peer.hpp
index 80a2264e3..4c8865c65 100644
--- a/node/Peer.hpp
+++ b/node/Peer.hpp
@@ -33,6 +33,7 @@
 #include "Mutex.hpp"
 #include "Bond.hpp"
 #include "BondController.hpp"
+#include "AES.hpp"
 
 #define ZT_PEER_MAX_SERIALIZED_STATE_SIZE (sizeof(Peer) + 32 + (sizeof(Path) * 2))
 
@@ -532,6 +533,9 @@ public:
 	 */
 	inline int8_t bondingPolicy() { return _bondingPolicy; }
 
+	const AES *aesKeysIfSupported() const
+	{ return (_vProto >= 10) ? _aesKeys : (const AES *)0; }
+
 private:
 	struct _PeerPath
 	{
@@ -542,6 +546,7 @@ private:
 	};
 
 	uint8_t _key[ZT_PEER_SECRET_KEY_LENGTH];
+	AES _aesKeys[2];
 
 	const RuntimeEnvironment *RR;
 
diff --git a/node/SHA512.cpp b/node/SHA512.cpp
index d3c938afb..ee28ed8bc 100644
--- a/node/SHA512.cpp
+++ b/node/SHA512.cpp
@@ -1,367 +1,274 @@
-// Code taken from NaCl by D. J. Bernstein and others
-// Public domain
-
-/*
-20080913
-D. J. Bernstein
-Public domain.
-*/
-
-#include <stdint.h>
-#include <stdlib.h>
-#include <string.h>
+// This code is public domain, taken from a PD crypto source file on GitHub.
 
 #include "SHA512.hpp"
 #include "Utils.hpp"
 
-#ifdef __APPLE__
-#include <CommonCrypto/CommonDigest.h>
-#define ZT_HAVE_NATIVE_SHA512
 namespace ZeroTier {
-void SHA512::hash(void *digest,const void *data,unsigned int len)
-{
-	CC_SHA512_CTX ctx;
-	CC_SHA512_Init(&ctx);
-	CC_SHA512_Update(&ctx,data,len);
-	CC_SHA512_Final(reinterpret_cast<unsigned char *>(digest),&ctx);
-}
-}
-#endif
-
-#ifdef ZT_USE_LIBCRYPTO
-#include <openssl/sha.h>
-#define ZT_HAVE_NATIVE_SHA512
-namespace ZeroTier {
-void SHA512::hash(void *digest,const void *data,unsigned int len)
-{
-	SHA512_CTX ctx;
-	SHA512_Init(&ctx);
-	SHA512_Update(&ctx,data,len);
-	SHA512_Final(reinterpret_cast<unsigned char *>(digest),&ctx);
-}
-}
-#endif
 
 #ifndef ZT_HAVE_NATIVE_SHA512
 
-namespace ZeroTier {
+namespace {
 
-#define uint64 uint64_t
-
-#ifdef ZT_NO_TYPE_PUNNING
-
-static uint64 load_bigendian(const unsigned char *x)
-{
-	return
-			(uint64) (x[7]) \
-	| (((uint64) (x[6])) << 8) \
-	| (((uint64) (x[5])) << 16) \
-	| (((uint64) (x[4])) << 24) \
-	| (((uint64) (x[3])) << 32) \
-	| (((uint64) (x[2])) << 40) \
-	| (((uint64) (x[1])) << 48) \
-	| (((uint64) (x[0])) << 56)
-	;
-}
-
-static void store_bigendian(unsigned char *x,uint64 u)
-{
-	x[7] = u; u >>= 8;
-	x[6] = u; u >>= 8;
-	x[5] = u; u >>= 8;
-	x[4] = u; u >>= 8;
-	x[3] = u; u >>= 8;
-	x[2] = u; u >>= 8;
-	x[1] = u; u >>= 8;
-	x[0] = u;
-}
-
-#else // !ZT_NO_TYPE_PUNNING
-
-#define load_bigendian(x) Utils::ntoh(*((const uint64_t *)(x)))
-#define store_bigendian(x,u) (*((uint64_t *)(x)) = Utils::hton((u)))
-
-#endif // ZT_NO_TYPE_PUNNING
-
-#define SHR(x,c) ((x) >> (c))
-#define ROTR(x,c) (((x) >> (c)) | ((x) << (64 - (c))))
-
-#define Ch(x,y,z) ((x & y) ^ (~x & z))
-#define Maj(x,y,z) ((x & y) ^ (x & z) ^ (y & z))
-#define Sigma0(x) (ROTR(x,28) ^ ROTR(x,34) ^ ROTR(x,39))
-#define Sigma1(x) (ROTR(x,14) ^ ROTR(x,18) ^ ROTR(x,41))
-#define sigma0(x) (ROTR(x, 1) ^ ROTR(x, 8) ^ SHR(x,7))
-#define sigma1(x) (ROTR(x,19) ^ ROTR(x,61) ^ SHR(x,6))
-
-#define M(w0,w14,w9,w1) w0 = sigma1(w14) + w9 + sigma0(w1) + w0;
-
-#define EXPAND \
-	M(w0 ,w14,w9 ,w1 ) \
-	M(w1 ,w15,w10,w2 ) \
-	M(w2 ,w0 ,w11,w3 ) \
-	M(w3 ,w1 ,w12,w4 ) \
-	M(w4 ,w2 ,w13,w5 ) \
-	M(w5 ,w3 ,w14,w6 ) \
-	M(w6 ,w4 ,w15,w7 ) \
-	M(w7 ,w5 ,w0 ,w8 ) \
-	M(w8 ,w6 ,w1 ,w9 ) \
-	M(w9 ,w7 ,w2 ,w10) \
-	M(w10,w8 ,w3 ,w11) \
-	M(w11,w9 ,w4 ,w12) \
-	M(w12,w10,w5 ,w13) \
-	M(w13,w11,w6 ,w14) \
-	M(w14,w12,w7 ,w15) \
-	M(w15,w13,w8 ,w0 )
-
-#define F(w,k) \
-	T1 = h + Sigma1(e) + Ch(e,f,g) + k + w; \
-	T2 = Sigma0(a) + Maj(a,b,c); \
-	h = g; \
-	g = f; \
-	f = e; \
-	e = d + T1; \
-	d = c; \
-	c = b; \
-	b = a; \
-	a = T1 + T2;
-
-static inline int crypto_hashblocks(unsigned char *statebytes,const unsigned char *in,unsigned long long inlen)
-{
-	uint64 state[8];
-	uint64 a;
-	uint64 b;
-	uint64 c;
-	uint64 d;
-	uint64 e;
-	uint64 f;
-	uint64 g;
-	uint64 h;
-	uint64 T1;
-	uint64 T2;
-
-	a = load_bigendian(statebytes +  0); state[0] = a;
-	b = load_bigendian(statebytes +  8); state[1] = b;
-	c = load_bigendian(statebytes + 16); state[2] = c;
-	d = load_bigendian(statebytes + 24); state[3] = d;
-	e = load_bigendian(statebytes + 32); state[4] = e;
-	f = load_bigendian(statebytes + 40); state[5] = f;
-	g = load_bigendian(statebytes + 48); state[6] = g;
-	h = load_bigendian(statebytes + 56); state[7] = h;
-
-	while (inlen >= 128) {
-		uint64 w0  = load_bigendian(in +   0);
-		uint64 w1  = load_bigendian(in +   8);
-		uint64 w2  = load_bigendian(in +  16);
-		uint64 w3  = load_bigendian(in +  24);
-		uint64 w4  = load_bigendian(in +  32);
-		uint64 w5  = load_bigendian(in +  40);
-		uint64 w6  = load_bigendian(in +  48);
-		uint64 w7  = load_bigendian(in +  56);
-		uint64 w8  = load_bigendian(in +  64);
-		uint64 w9  = load_bigendian(in +  72);
-		uint64 w10 = load_bigendian(in +  80);
-		uint64 w11 = load_bigendian(in +  88);
-		uint64 w12 = load_bigendian(in +  96);
-		uint64 w13 = load_bigendian(in + 104);
-		uint64 w14 = load_bigendian(in + 112);
-		uint64 w15 = load_bigendian(in + 120);
-
-		F(w0 ,0x428a2f98d728ae22ULL)
-		F(w1 ,0x7137449123ef65cdULL)
-		F(w2 ,0xb5c0fbcfec4d3b2fULL)
-		F(w3 ,0xe9b5dba58189dbbcULL)
-		F(w4 ,0x3956c25bf348b538ULL)
-		F(w5 ,0x59f111f1b605d019ULL)
-		F(w6 ,0x923f82a4af194f9bULL)
-		F(w7 ,0xab1c5ed5da6d8118ULL)
-		F(w8 ,0xd807aa98a3030242ULL)
-		F(w9 ,0x12835b0145706fbeULL)
-		F(w10,0x243185be4ee4b28cULL)
-		F(w11,0x550c7dc3d5ffb4e2ULL)
-		F(w12,0x72be5d74f27b896fULL)
-		F(w13,0x80deb1fe3b1696b1ULL)
-		F(w14,0x9bdc06a725c71235ULL)
-		F(w15,0xc19bf174cf692694ULL)
-
-		EXPAND
-
-		F(w0 ,0xe49b69c19ef14ad2ULL)
-		F(w1 ,0xefbe4786384f25e3ULL)
-		F(w2 ,0x0fc19dc68b8cd5b5ULL)
-		F(w3 ,0x240ca1cc77ac9c65ULL)
-		F(w4 ,0x2de92c6f592b0275ULL)
-		F(w5 ,0x4a7484aa6ea6e483ULL)
-		F(w6 ,0x5cb0a9dcbd41fbd4ULL)
-		F(w7 ,0x76f988da831153b5ULL)
-		F(w8 ,0x983e5152ee66dfabULL)
-		F(w9 ,0xa831c66d2db43210ULL)
-		F(w10,0xb00327c898fb213fULL)
-		F(w11,0xbf597fc7beef0ee4ULL)
-		F(w12,0xc6e00bf33da88fc2ULL)
-		F(w13,0xd5a79147930aa725ULL)
-		F(w14,0x06ca6351e003826fULL)
-		F(w15,0x142929670a0e6e70ULL)
-
-		EXPAND
-
-		F(w0 ,0x27b70a8546d22ffcULL)
-		F(w1 ,0x2e1b21385c26c926ULL)
-		F(w2 ,0x4d2c6dfc5ac42aedULL)
-		F(w3 ,0x53380d139d95b3dfULL)
-		F(w4 ,0x650a73548baf63deULL)
-		F(w5 ,0x766a0abb3c77b2a8ULL)
-		F(w6 ,0x81c2c92e47edaee6ULL)
-		F(w7 ,0x92722c851482353bULL)
-		F(w8 ,0xa2bfe8a14cf10364ULL)
-		F(w9 ,0xa81a664bbc423001ULL)
-		F(w10,0xc24b8b70d0f89791ULL)
-		F(w11,0xc76c51a30654be30ULL)
-		F(w12,0xd192e819d6ef5218ULL)
-		F(w13,0xd69906245565a910ULL)
-		F(w14,0xf40e35855771202aULL)
-		F(w15,0x106aa07032bbd1b8ULL)
-
-		EXPAND
-
-		F(w0 ,0x19a4c116b8d2d0c8ULL)
-		F(w1 ,0x1e376c085141ab53ULL)
-		F(w2 ,0x2748774cdf8eeb99ULL)
-		F(w3 ,0x34b0bcb5e19b48a8ULL)
-		F(w4 ,0x391c0cb3c5c95a63ULL)
-		F(w5 ,0x4ed8aa4ae3418acbULL)
-		F(w6 ,0x5b9cca4f7763e373ULL)
-		F(w7 ,0x682e6ff3d6b2b8a3ULL)
-		F(w8 ,0x748f82ee5defb2fcULL)
-		F(w9 ,0x78a5636f43172f60ULL)
-		F(w10,0x84c87814a1f0ab72ULL)
-		F(w11,0x8cc702081a6439ecULL)
-		F(w12,0x90befffa23631e28ULL)
-		F(w13,0xa4506cebde82bde9ULL)
-		F(w14,0xbef9a3f7b2c67915ULL)
-		F(w15,0xc67178f2e372532bULL)
-
-		EXPAND
-
-		F(w0 ,0xca273eceea26619cULL)
-		F(w1 ,0xd186b8c721c0c207ULL)
-		F(w2 ,0xeada7dd6cde0eb1eULL)
-		F(w3 ,0xf57d4f7fee6ed178ULL)
-		F(w4 ,0x06f067aa72176fbaULL)
-		F(w5 ,0x0a637dc5a2c898a6ULL)
-		F(w6 ,0x113f9804bef90daeULL)
-		F(w7 ,0x1b710b35131c471bULL)
-		F(w8 ,0x28db77f523047d84ULL)
-		F(w9 ,0x32caab7b40c72493ULL)
-		F(w10,0x3c9ebe0a15c9bebcULL)
-		F(w11,0x431d67c49c100d4cULL)
-		F(w12,0x4cc5d4becb3e42b6ULL)
-		F(w13,0x597f299cfc657e2aULL)
-		F(w14,0x5fcb6fab3ad6faecULL)
-		F(w15,0x6c44198c4a475817ULL)
-
-		a += state[0];
-		b += state[1];
-		c += state[2];
-		d += state[3];
-		e += state[4];
-		f += state[5];
-		g += state[6];
-		h += state[7];
-
-		state[0] = a;
-		state[1] = b;
-		state[2] = c;
-		state[3] = d;
-		state[4] = e;
-		state[5] = f;
-		state[6] = g;
-		state[7] = h;
-
-		in += 128;
-		inlen -= 128;
-	}
-
-	store_bigendian(statebytes +  0,state[0]);
-	store_bigendian(statebytes +  8,state[1]);
-	store_bigendian(statebytes + 16,state[2]);
-	store_bigendian(statebytes + 24,state[3]);
-	store_bigendian(statebytes + 32,state[4]);
-	store_bigendian(statebytes + 40,state[5]);
-	store_bigendian(statebytes + 48,state[6]);
-	store_bigendian(statebytes + 56,state[7]);
-
-	return 0;
-}
-
-#define blocks crypto_hashblocks
-
-static const unsigned char iv[64] = {
-	0x6a,0x09,0xe6,0x67,0xf3,0xbc,0xc9,0x08,
-	0xbb,0x67,0xae,0x85,0x84,0xca,0xa7,0x3b,
-	0x3c,0x6e,0xf3,0x72,0xfe,0x94,0xf8,0x2b,
-	0xa5,0x4f,0xf5,0x3a,0x5f,0x1d,0x36,0xf1,
-	0x51,0x0e,0x52,0x7f,0xad,0xe6,0x82,0xd1,
-	0x9b,0x05,0x68,0x8c,0x2b,0x3e,0x6c,0x1f,
-	0x1f,0x83,0xd9,0xab,0xfb,0x41,0xbd,0x6b,
-	0x5b,0xe0,0xcd,0x19,0x13,0x7e,0x21,0x79
+struct sha512_state {
+	uint64_t length,state[8];
+	unsigned long curlen;
+	uint8_t buf[128];
 };
 
-void SHA512::hash(void *digest,const void *data,unsigned int len)
+static const uint64_t K[80] = {
+	0x428a2f98d728ae22ULL,0x7137449123ef65cdULL,0xb5c0fbcfec4d3b2fULL,0xe9b5dba58189dbbcULL,
+	0x3956c25bf348b538ULL,0x59f111f1b605d019ULL,0x923f82a4af194f9bULL,0xab1c5ed5da6d8118ULL,
+	0xd807aa98a3030242ULL,0x12835b0145706fbeULL,0x243185be4ee4b28cULL,0x550c7dc3d5ffb4e2ULL,
+	0x72be5d74f27b896fULL,0x80deb1fe3b1696b1ULL,0x9bdc06a725c71235ULL,0xc19bf174cf692694ULL,
+	0xe49b69c19ef14ad2ULL,0xefbe4786384f25e3ULL,0x0fc19dc68b8cd5b5ULL,0x240ca1cc77ac9c65ULL,
+	0x2de92c6f592b0275ULL,0x4a7484aa6ea6e483ULL,0x5cb0a9dcbd41fbd4ULL,0x76f988da831153b5ULL,
+	0x983e5152ee66dfabULL,0xa831c66d2db43210ULL,0xb00327c898fb213fULL,0xbf597fc7beef0ee4ULL,
+	0xc6e00bf33da88fc2ULL,0xd5a79147930aa725ULL,0x06ca6351e003826fULL,0x142929670a0e6e70ULL,
+	0x27b70a8546d22ffcULL,0x2e1b21385c26c926ULL,0x4d2c6dfc5ac42aedULL,0x53380d139d95b3dfULL,
+	0x650a73548baf63deULL,0x766a0abb3c77b2a8ULL,0x81c2c92e47edaee6ULL,0x92722c851482353bULL,
+	0xa2bfe8a14cf10364ULL,0xa81a664bbc423001ULL,0xc24b8b70d0f89791ULL,0xc76c51a30654be30ULL,
+	0xd192e819d6ef5218ULL,0xd69906245565a910ULL,0xf40e35855771202aULL,0x106aa07032bbd1b8ULL,
+	0x19a4c116b8d2d0c8ULL,0x1e376c085141ab53ULL,0x2748774cdf8eeb99ULL,0x34b0bcb5e19b48a8ULL,
+	0x391c0cb3c5c95a63ULL,0x4ed8aa4ae3418acbULL,0x5b9cca4f7763e373ULL,0x682e6ff3d6b2b8a3ULL,
+	0x748f82ee5defb2fcULL,0x78a5636f43172f60ULL,0x84c87814a1f0ab72ULL,0x8cc702081a6439ecULL,
+	0x90befffa23631e28ULL,0xa4506cebde82bde9ULL,0xbef9a3f7b2c67915ULL,0xc67178f2e372532bULL,
+	0xca273eceea26619cULL,0xd186b8c721c0c207ULL,0xeada7dd6cde0eb1eULL,0xf57d4f7fee6ed178ULL,
+	0x06f067aa72176fbaULL,0x0a637dc5a2c898a6ULL,0x113f9804bef90daeULL,0x1b710b35131c471bULL,
+	0x28db77f523047d84ULL,0x32caab7b40c72493ULL,0x3c9ebe0a15c9bebcULL,0x431d67c49c100d4cULL,
+	0x4cc5d4becb3e42b6ULL,0x597f299cfc657e2aULL,0x5fcb6fab3ad6faecULL,0x6c44198c4a475817ULL
+};
+
+#define STORE64H(x, y) Utils::storeBigEndian<uint64_t>(y,x)
+#define LOAD64H(x, y) x = Utils::loadBigEndian<uint64_t>(y)
+#define ROL64c(x,y) (((x)<<(y)) | ((x)>>(64-(y))))
+#define ROR64c(x,y) (((x)>>(y)) | ((x)<<(64-(y))))
+#define Ch(x,y,z)       (z ^ (x & (y ^ z)))
+#define Maj(x,y,z)      (((x | y) & z) | (x & y))
+#define S(x, n)         ROR64c(x, n)
+#define R(x, n)         ((x)>>(n))
+#define Sigma0(x)       (S(x, 28) ^ S(x, 34) ^ S(x, 39))
+#define Sigma1(x)       (S(x, 14) ^ S(x, 18) ^ S(x, 41))
+#define Gamma0(x)       (S(x, 1) ^ S(x, 8) ^ R(x, 7))
+#define Gamma1(x)       (S(x, 19) ^ S(x, 61) ^ R(x, 6))
+
+static ZT_INLINE void sha512_compress(sha512_state *const md,uint8_t *const buf)
 {
-	unsigned char h[64];
-	unsigned char padded[256];
+	uint64_t S[8], W[80], t0, t1;
 	int i;
-	uint64_t bytes = len;
 
-	const unsigned char *in = (const unsigned char *)data;
-	unsigned int inlen = len;
+	for (i = 0; i < 8; i++)
+		S[i] = md->state[i];
+	for (i = 0; i < 16; i++)
+		LOAD64H(W[i], buf + (8*i));
+	for (i = 16; i < 80; i++)
+		W[i] = Gamma1(W[i - 2]) + W[i - 7] + Gamma0(W[i - 15]) + W[i - 16];
 
-	for (i = 0;i < 64;++i) h[i] = iv[i];
+#define RND(a,b,c,d,e,f,g,h,i) \
+	t0 = h + Sigma1(e) + Ch(e, f, g) + K[i] + W[i]; \
+	t1 = Sigma0(a) + Maj(a, b, c); \
+	d += t0; \
+	h  = t0 + t1;
 
-	blocks(h,in,inlen);
-	in += inlen;
-	inlen &= 127;
-	in -= inlen;
-
-	for (i = 0;i < (int)inlen;++i) padded[i] = in[i];
-	padded[inlen] = 0x80;
-
-	if (inlen < 112) {
-		for (i = inlen + 1;i < 119;++i) padded[i] = 0;
-		padded[119] = (unsigned char)((bytes >> 61) & 0xff);
-		padded[120] = (unsigned char)((bytes >> 53) & 0xff);
-		padded[121] = (unsigned char)((bytes >> 45) & 0xff);
-		padded[122] = (unsigned char)((bytes >> 37) & 0xff);
-		padded[123] = (unsigned char)((bytes >> 29) & 0xff);
-		padded[124] = (unsigned char)((bytes >> 21) & 0xff);
-		padded[125] = (unsigned char)((bytes >> 13) & 0xff);
-		padded[126] = (unsigned char)((bytes >> 5) & 0xff);
-		padded[127] = (unsigned char)((bytes << 3) & 0xff);
-		blocks(h,padded,128);
-	} else {
-		for (i = inlen + 1;i < 247;++i) padded[i] = 0;
-		padded[247] = (unsigned char)((bytes >> 61) & 0xff);
-		padded[248] = (unsigned char)((bytes >> 53) & 0xff);
-		padded[249] = (unsigned char)((bytes >> 45) & 0xff);
-		padded[250] = (unsigned char)((bytes >> 37) & 0xff);
-		padded[251] = (unsigned char)((bytes >> 29) & 0xff);
-		padded[252] = (unsigned char)((bytes >> 21) & 0xff);
-		padded[253] = (unsigned char)((bytes >> 13) & 0xff);
-		padded[254] = (unsigned char)((bytes >> 5) & 0xff);
-		padded[255] = (unsigned char)((bytes << 3) & 0xff);
-		blocks(h,padded,256);
+	for (i = 0; i < 80; i += 8) {
+		RND(S[0],S[1],S[2],S[3],S[4],S[5],S[6],S[7],i+0);
+		RND(S[7],S[0],S[1],S[2],S[3],S[4],S[5],S[6],i+1);
+		RND(S[6],S[7],S[0],S[1],S[2],S[3],S[4],S[5],i+2);
+		RND(S[5],S[6],S[7],S[0],S[1],S[2],S[3],S[4],i+3);
+		RND(S[4],S[5],S[6],S[7],S[0],S[1],S[2],S[3],i+4);
+		RND(S[3],S[4],S[5],S[6],S[7],S[0],S[1],S[2],i+5);
+		RND(S[2],S[3],S[4],S[5],S[6],S[7],S[0],S[1],i+6);
+		RND(S[1],S[2],S[3],S[4],S[5],S[6],S[7],S[0],i+7);
 	}
 
-	for (i = 0;i < 64;++i) ((unsigned char *)digest)[i] = h[i];
+	for (i = 0; i < 8; i++)
+		md->state[i] = md->state[i] + S[i];
 }
 
-} // namespace ZeroTier
+static ZT_INLINE void sha384_init(sha512_state *const md)
+{
+	md->curlen = 0;
+	md->length = 0;
+	md->state[0] = 0xcbbb9d5dc1059ed8ULL;
+	md->state[1] = 0x629a292a367cd507ULL;
+	md->state[2] = 0x9159015a3070dd17ULL;
+	md->state[3] = 0x152fecd8f70e5939ULL;
+	md->state[4] = 0x67332667ffc00b31ULL;
+	md->state[5] = 0x8eb44a8768581511ULL;
+	md->state[6] = 0xdb0c2e0d64f98fa7ULL;
+	md->state[7] = 0x47b5481dbefa4fa4ULL;
+}
+
+static ZT_INLINE void sha512_init(sha512_state *const md)
+{
+	md->curlen = 0;
+	md->length = 0;
+	md->state[0] = 0x6a09e667f3bcc908ULL;
+	md->state[1] = 0xbb67ae8584caa73bULL;
+	md->state[2] = 0x3c6ef372fe94f82bULL;
+	md->state[3] = 0xa54ff53a5f1d36f1ULL;
+	md->state[4] = 0x510e527fade682d1ULL;
+	md->state[5] = 0x9b05688c2b3e6c1fULL;
+	md->state[6] = 0x1f83d9abfb41bd6bULL;
+	md->state[7] = 0x5be0cd19137e2179ULL;
+}
+
+static void sha512_process(sha512_state *const md,const uint8_t *in,unsigned long inlen)
+{
+	while (inlen > 0) {
+		if (md->curlen == 0 && inlen >= 128) {
+			sha512_compress(md,(uint8_t *)in);
+			md->length     += 128 * 8;
+			in             += 128;
+			inlen          -= 128;
+		} else {
+			unsigned long n = std::min(inlen,(128 - md->curlen));
+			Utils::copy(md->buf + md->curlen,in,n);
+			md->curlen += n;
+			in             += n;
+			inlen          -= n;
+			if (md->curlen == 128) {
+				sha512_compress(md,md->buf);
+				md->length += 8*128;
+				md->curlen = 0;
+			}
+		}
+	}
+}
+
+static ZT_INLINE void sha512_done(sha512_state *const md,uint8_t *out)
+{
+	int i;
+
+	md->length += md->curlen * 8ULL;
+	md->buf[md->curlen++] = (uint8_t)0x80;
+
+	if (md->curlen > 112) {
+		while (md->curlen < 128) {
+			md->buf[md->curlen++] = (uint8_t)0;
+		}
+		sha512_compress(md, md->buf);
+		md->curlen = 0;
+	}
+
+	while (md->curlen < 120) {
+		md->buf[md->curlen++] = (uint8_t)0;
+	}
+
+	STORE64H(md->length, md->buf+120);
+	sha512_compress(md, md->buf);
+
+	for (i = 0; i < 8; i++) {
+		STORE64H(md->state[i], out+(8*i));
+	}
+}
+
+} // anonymous namespace
+
+void SHA512(void *digest,const void *data,unsigned int len)
+{
+	sha512_state state;
+	sha512_init(&state);
+	sha512_process(&state,(uint8_t *)data,(unsigned long)len);
+	sha512_done(&state,(uint8_t *)digest);
+}
+
+void SHA384(void *digest,const void *data,unsigned int len)
+{
+	uint8_t tmp[64];
+	sha512_state state;
+	sha384_init(&state);
+	sha512_process(&state,(uint8_t *)data,(unsigned long)len);
+	sha512_done(&state,tmp);
+	Utils::copy<48>(digest,tmp);
+}
+
+void SHA384(void *digest,const void *data0,unsigned int len0,const void *data1,unsigned int len1)
+{
+	uint8_t tmp[64];
+	sha512_state state;
+	sha384_init(&state);
+	sha512_process(&state,(uint8_t *)data0,(unsigned long)len0);
+	sha512_process(&state,(uint8_t *)data1,(unsigned long)len1);
+	sha512_done(&state,tmp);
+	Utils::copy<48>(digest,tmp);
+}
 
 #endif // !ZT_HAVE_NATIVE_SHA512
 
-// Internally re-export to included C code, which includes some fast crypto code ported in on some platforms.
-// This eliminates the need to link against a third party SHA512() from this code
-extern "C" void ZT_sha512internal(void *digest,const void *data,unsigned int len)
+void HMACSHA384(const uint8_t key[ZT_SYMMETRIC_KEY_SIZE],const void *msg,const unsigned int msglen,uint8_t mac[48])
 {
-	ZeroTier::SHA512::hash(digest,data,len);
+	uint64_t kInPadded[16]; // input padded key
+	uint64_t outer[22]; // output padded key | H(input padded key | msg)
+
+	const uint64_t k0 = Utils::loadMachineEndian< uint64_t >(key);
+	const uint64_t k1 = Utils::loadMachineEndian< uint64_t >(key + 8);
+	const uint64_t k2 = Utils::loadMachineEndian< uint64_t >(key + 16);
+	const uint64_t k3 = Utils::loadMachineEndian< uint64_t >(key + 24);
+	const uint64_t k4 = Utils::loadMachineEndian< uint64_t >(key + 32);
+	const uint64_t k5 = Utils::loadMachineEndian< uint64_t >(key + 40);
+
+	const uint64_t ipad = 0x3636363636363636ULL;
+	kInPadded[0] = k0 ^ ipad;
+	kInPadded[1] = k1 ^ ipad;
+	kInPadded[2] = k2 ^ ipad;
+	kInPadded[3] = k3 ^ ipad;
+	kInPadded[4] = k4 ^ ipad;
+	kInPadded[5] = k5 ^ ipad;
+	kInPadded[6] = ipad;
+	kInPadded[7] = ipad;
+	kInPadded[8] = ipad;
+	kInPadded[9] = ipad;
+	kInPadded[10] = ipad;
+	kInPadded[11] = ipad;
+	kInPadded[12] = ipad;
+	kInPadded[13] = ipad;
+	kInPadded[14] = ipad;
+	kInPadded[15] = ipad;
+
+	const uint64_t opad = 0x5c5c5c5c5c5c5c5cULL;
+	outer[0] = k0 ^ opad;
+	outer[1] = k1 ^ opad;
+	outer[2] = k2 ^ opad;
+	outer[3] = k3 ^ opad;
+	outer[4] = k4 ^ opad;
+	outer[5] = k5 ^ opad;
+	outer[6] = opad;
+	outer[7] = opad;
+	outer[8] = opad;
+	outer[9] = opad;
+	outer[10] = opad;
+	outer[11] = opad;
+	outer[12] = opad;
+	outer[13] = opad;
+	outer[14] = opad;
+	outer[15] = opad;
+
+	// H(output padded key | H(input padded key | msg))
+	SHA384(reinterpret_cast<uint8_t *>(outer) + 128,kInPadded,128,msg,msglen);
+	SHA384(mac,outer,176);
 }
+
+void KBKDFHMACSHA384(const uint8_t key[ZT_SYMMETRIC_KEY_SIZE],const char label,const char context,const uint32_t iter,uint8_t out[ZT_SYMMETRIC_KEY_SIZE])
+{
+	uint8_t kbkdfMsg[13];
+
+	Utils::storeBigEndian<uint32_t>(kbkdfMsg,(uint32_t)iter);
+
+	kbkdfMsg[4] = (uint8_t)'Z';
+	kbkdfMsg[5] = (uint8_t)'T'; // preface our labels with something ZT-specific
+	kbkdfMsg[6] = (uint8_t)label;
+	kbkdfMsg[7] = 0;
+
+	kbkdfMsg[8] = (uint8_t)context;
+
+	// Output key length: 384 bits (as 32-bit big-endian value)
+	kbkdfMsg[9] = 0;
+	kbkdfMsg[10] = 0;
+	kbkdfMsg[11] = 0x01;
+	kbkdfMsg[12] = 0x80;
+
+	static_assert(ZT_SYMMETRIC_KEY_SIZE == ZT_SHA384_DIGEST_SIZE,"sizeof(out) != ZT_SHA384_DIGEST_SIZE");
+	HMACSHA384(key,&kbkdfMsg,sizeof(kbkdfMsg),out);
+}
+
+} // namespace ZeroTier
diff --git a/node/SHA512.hpp b/node/SHA512.hpp
index eefaead16..12ac9583a 100644
--- a/node/SHA512.hpp
+++ b/node/SHA512.hpp
@@ -1,5 +1,5 @@
 /*
- * Copyright (c)2019 ZeroTier, Inc.
+ * Copyright (c)2013-2020 ZeroTier, Inc.
  *
  * Use of this software is governed by the Business Source License included
  * in the LICENSE.TXT file in the project's root directory.
@@ -14,18 +14,74 @@
 #ifndef ZT_SHA512_HPP
 #define ZT_SHA512_HPP
 
-#define ZT_SHA512_DIGEST_LEN 64
+#include "Constants.hpp"
+
+#ifdef __APPLE__
+#include <CommonCrypto/CommonDigest.h>
+#endif
+
+#define ZT_SHA512_DIGEST_SIZE 64
+#define ZT_SHA384_DIGEST_SIZE 48
+#define ZT_SHA512_BLOCK_SIZE 128
+#define ZT_SHA384_BLOCK_SIZE 128
+
+#define ZT_HMACSHA384_LEN 48
 
 namespace ZeroTier {
 
-/**
- * SHA-512 digest algorithm
- */
-class SHA512
+// SHA384 and SHA512 are actually in the standard libraries on MacOS and iOS
+#ifdef __APPLE__
+#define ZT_HAVE_NATIVE_SHA512 1
+static ZT_INLINE void SHA512(void *digest,const void *data,unsigned int len)
 {
-public:
-	static void hash(void *digest,const void *data,unsigned int len);
-};
+	CC_SHA512_CTX ctx;
+	CC_SHA512_Init(&ctx);
+	CC_SHA512_Update(&ctx,data,len);
+	CC_SHA512_Final(reinterpret_cast<unsigned char *>(digest),&ctx);
+}
+static ZT_INLINE void SHA384(void *digest,const void *data,unsigned int len)
+{
+	CC_SHA512_CTX ctx;
+	CC_SHA384_Init(&ctx);
+	CC_SHA384_Update(&ctx,data,len);
+	CC_SHA384_Final(reinterpret_cast<unsigned char *>(digest),&ctx);
+}
+static ZT_INLINE void SHA384(void *digest,const void *data0,unsigned int len0,const void *data1,unsigned int len1)
+{
+	CC_SHA512_CTX ctx;
+	CC_SHA384_Init(&ctx);
+	CC_SHA384_Update(&ctx,data0,len0);
+	CC_SHA384_Update(&ctx,data1,len1);
+	CC_SHA384_Final(reinterpret_cast<unsigned char *>(digest),&ctx);
+}
+#endif
+
+#ifndef ZT_HAVE_NATIVE_SHA512
+void SHA512(void *digest,const void *data,unsigned int len);
+void SHA384(void *digest,const void *data,unsigned int len);
+void SHA384(void *digest,const void *data0,unsigned int len0,const void *data1,unsigned int len1);
+#endif
+
+/**
+ * Compute HMAC SHA-384 using a 256-bit key
+ *
+ * @param key Secret key
+ * @param msg Message to HMAC
+ * @param msglen Length of message
+ * @param mac Buffer to fill with result
+ */
+void HMACSHA384(const uint8_t key[ZT_SYMMETRIC_KEY_SIZE],const void *msg,unsigned int msglen,uint8_t mac[48]);
+
+/**
+ * Compute KBKDF (key-based key derivation function) using HMAC-SHA-384 as a PRF
+ *
+ * @param key Source master key
+ * @param label A label indicating the key's purpose in the ZeroTier system
+ * @param context An arbitrary "context" or zero if not applicable
+ * @param iter Key iteration for generation of multiple keys for the same label/context
+ * @param out Output to receive derived key
+ */
+void KBKDFHMACSHA384(const uint8_t key[ZT_SYMMETRIC_KEY_SIZE],char label,char context,uint32_t iter,uint8_t out[ZT_SYMMETRIC_KEY_SIZE]);
 
 } // namespace ZeroTier
 
diff --git a/node/Switch.cpp b/node/Switch.cpp
index c24e2d231..1e7e2ca23 100644
--- a/node/Switch.cpp
+++ b/node/Switch.cpp
@@ -96,7 +96,7 @@ void Switch::onRemotePacket(void *tPtr,const int64_t localSocket,const InetAddre
 				if ((now - _lastBeaconResponse) >= 2500) { // limit rate of responses
 					_lastBeaconResponse = now;
 					Packet outp(peer->address(),RR->identity.address(),Packet::VERB_NOP);
-					outp.armor(peer->key(),true);
+					outp.armor(peer->key(),true,peer->aesKeysIfSupported());
 					path->send(RR,tPtr,outp.data(),outp.size(),now);
 				}
 			}
@@ -879,7 +879,6 @@ void Switch::requestWhois(void *tPtr,const int64_t now,const Address &addr)
 		int32_t flowId = ZT_QOS_NO_FLOW;
 		Packet outp(upstream->address(),RR->identity.address(),Packet::VERB_WHOIS);
 		addr.appendTo(outp);
-		RR->node->expectReplyTo(outp.packetId());
 		send(tPtr,outp,true,flowId);
 	}
 }
@@ -1042,14 +1041,16 @@ void Switch::_sendViaSpecificPath(void *tPtr,SharedPtr<Peer> peer,SharedPtr<Path
 	unsigned int chunkSize = std::min(packet.size(),mtu);
 	packet.setFragmented(chunkSize < packet.size());
 
-	peer->recordOutgoingPacket(viaPath, packet.packetId(), packet.payloadLength(), packet.verb(), flowId, now);
-
 	if (trustedPathId) {
 		packet.setTrusted(trustedPathId);
 	} else {
-		packet.armor(peer->key(),encrypt);
+		Packet::Verb v = packet.verb();
+		packet.armor(peer->key(),encrypt,peer->aesKeysIfSupported());
+		RR->node->expectReplyTo(packet.packetId());
 	}
 
+	peer->recordOutgoingPacket(viaPath, packet.packetId(), packet.payloadLength(), packet.verb(), flowId, now);
+
 	if (viaPath->send(RR,tPtr,packet.data(),chunkSize,now)) {
 		if (chunkSize < packet.size()) {
 			// Too big for one packet, fragment the rest
diff --git a/service/SoftwareUpdater.cpp b/service/SoftwareUpdater.cpp
index 6ace5fd84..8c76ab3eb 100644
--- a/service/SoftwareUpdater.cpp
+++ b/service/SoftwareUpdater.cpp
@@ -112,10 +112,10 @@ void SoftwareUpdater::setUpdateDistribution(bool distribute)
 						// If update meta is called e.g. foo.exe.json, then foo.exe is the update itself
 						const std::string binPath(udd + ZT_PATH_SEPARATOR_S + u->substr(0,u->length() - 5));
 						const std::string metaHash(OSUtils::jsonBinFromHex(d.meta[ZT_SOFTWARE_UPDATE_JSON_UPDATE_HASH]));
-						if ((metaHash.length() == ZT_SHA512_DIGEST_LEN)&&(OSUtils::readFile(binPath.c_str(),d.bin))) {
-							std::array<uint8_t,ZT_SHA512_DIGEST_LEN> sha512;
-							SHA512::hash(sha512.data(),d.bin.data(),(unsigned int)d.bin.length());
-							if (!memcmp(sha512.data(),metaHash.data(),ZT_SHA512_DIGEST_LEN)) { // double check that hash in JSON is correct
+						if ((metaHash.length() == 64)&&(OSUtils::readFile(binPath.c_str(),d.bin))) {
+							std::array<uint8_t,64> sha512;
+							SHA512(sha512.data(),d.bin.data(),(unsigned int)d.bin.length());
+							if (!memcmp(sha512.data(),metaHash.data(),64)) { // double check that hash in JSON is correct
 								d.meta[ZT_SOFTWARE_UPDATE_JSON_UPDATE_SIZE] = d.bin.length(); // override with correct value -- setting this in meta json is optional
 								std::array<uint8_t,16> shakey;
 								memcpy(shakey.data(),sha512.data(),16);
@@ -333,10 +333,10 @@ bool SoftwareUpdater::check(const int64_t now)
 			const std::string binPath(_homePath + ZT_PATH_SEPARATOR_S ZT_SOFTWARE_UPDATE_BIN_FILENAME);
 			try {
 				// (1) Check the hash itself to make sure the image is basically okay
-				uint8_t sha512[ZT_SHA512_DIGEST_LEN];
-				SHA512::hash(sha512,_download.data(),(unsigned int)_download.length());
-				char hexbuf[(ZT_SHA512_DIGEST_LEN * 2) + 2];
-				if (OSUtils::jsonString(_latestMeta[ZT_SOFTWARE_UPDATE_JSON_UPDATE_HASH],"") == Utils::hex(sha512,ZT_SHA512_DIGEST_LEN,hexbuf)) {
+				uint8_t sha512[64];
+				SHA512(sha512,_download.data(),(unsigned int)_download.length());
+				char hexbuf[(64 * 2) + 2];
+				if (OSUtils::jsonString(_latestMeta[ZT_SOFTWARE_UPDATE_JSON_UPDATE_HASH],"") == Utils::hex(sha512,64,hexbuf)) {
 					// (2) Check signature by signing authority
 					const std::string sig(OSUtils::jsonBinFromHex(_latestMeta[ZT_SOFTWARE_UPDATE_JSON_UPDATE_SIGNATURE]));
 					if (Identity(ZT_SOFTWARE_UPDATE_SIGNING_AUTHORITY).verify(_download.data(),(unsigned int)_download.length(),sig.data(),(unsigned int)sig.length())) {