Pypacker: The fast and simple packet creating and parsing module

ethernet, ieee80211, intercept-packets, ipv4, ipv6, network, nfqueue, packets, parsing, pcap, protocol, python, security, tcp, testing, udp
pip install pypacker==5.1


Build Status version supported-versions supported-implementations License: GPL v2

General information

This is Pypacker: The fastest and simplest low-level packet manipulation library for Python. See below examples for what you can do with it.

If you like this project you can Donate with PayPal via PayPal.

What you can do with Pypacker

Create custom Packets (via keywords) or from raw bytes and change their data:

from pypacker.layer3.ip import IP
from pypacker.layer3.icmp import ICMP

# Packet via keywords
ip0 = IP(src_s="", dst_s="", p=1) +\
	ICMP(type=8) +\
	ICMP.Echo(id=123, seq=1, body_bytes=b"foobar")

# Packet from raw bytes. ip1_bts can also be retrieved via ip0.bin()
ip1_bts = b"E\x00\x00*\x00\x00\x00\x00@\x01;)\x7f\x00\x00\x01\xc0\xa8\x00\x01\x08\x00\xc0?\x00{\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00foobar"
ip1 = IP(ip1_bts) 
# Change source IPv4 address
ip0.src_s = ""
# Change ICMP payload
ip0[IP,ICMP,ICMP.Echo].body_bytes = b"foobar2"

# Output packet (similar result for ip1)
print("%s" % ip0)
        v_hl        : 0x45 = 69 = 0b1000101
        tos         : 0x0 = 0 = 0b0
        len         : 0x2B = 43 = 0b101011
        id          : 0x0 = 0 = 0b0
        off         : 0x0 = 0 = 0b0
        ttl         : 0x40 = 64 = 0b1000000
        p           : 0x1 = 1 = 0b1
        sum         : 0xB623 = 46627 = 0b1011011000100011
        src         : b'\x01\x02\x03\x04' =
        dst         : b'\xc0\xa8\x00\x01' =
        opts        : []
        type        : 0x8 = 8 = 0b1000
        code        : 0x0 = 0 = 0b0
        sum         : 0x8E3F = 36415 = 0b1000111000111111
        id          : 0x7B = 123 = 0b1111011
        seq         : 0x1 = 1 = 0b1
        ts          : 0x0 = 0 = 0b0
        bodybytes   : b'foobar2'

Read/write packets from/to file (pcap/tcpdump format):

from pypacker import ppcap
from pypacker.layer12 import ethernet
from pypacker.layer3 import ip
from pypacker.layer4 import tcp

preader = ppcap.Reader(filename="packets_ether.pcap")
pwriter = ppcap.Writer(filename="packets_ether_new.pcap", linktype=ppcap.DLT_EN10MB)

for ts, buf in preader:
	eth = ethernet.Ethernet(buf)

	if eth[ethernet.Ethernet, ip.IP, tcp.TCP] is not None:
		print("%d: %s:%s -> %s:%s" % (ts, eth[ip.IP].src_s, eth[tcp.TCP].sport,
			eth[ip.IP].dst_s, eth[tcp.TCP].dport))


Send/receive layer 2 packets:

from pypacker import psocket
from pypacker.layer12 import ethernet

psock = psocket.SocketHndl(timeout=10)

def filter_pkt(pkt):
	return == 80

# Receive raw bytes
for raw_bytes in psock:
	eth = ethernet.Ethernet(raw_bytes)
	print("Got packet: %r" % eth)
	# Send bytes
	# Receive raw bytes
	bts = psock.recv()
	# Send/receive based on source/destination data in packet
	pkts =
	# Use filter to get specific packets
	pkts = psock.recvp(filter_match_recv=filter_pkt)
	# stop on first packet


Intercept (and modificate) Packets eg for MITM:

# Add iptables rule:
# iptables -I INPUT 1 -p icmp -j NFQUEUE --queue-balance 0:2
import time

from pypacker import interceptor
from pypacker.layer3 import ip, icmp

# ICMP Echo request intercepting
def verdict_cb(ll_data, ll_proto_id, data, ctx):
	ip1 = ip.IP(data)
	icmp1 = ip1[icmp.ICMP]

	if icmp1 is None or icmp1.type != icmp.ICMP_TYPE_ECHO_REQ:
		return data, interceptor.NF_ACCEPT

	echo1 = icmp1[icmp.ICMP.Echo]

	if echo1 is None:
		return data, interceptor.NF_ACCEPT

	pp_bts = b"PYPACKER"
	print("changing ICMP echo request packet")
	echo1.body_bytes = echo1.body_bytes[:-len(pp_bts)] + pp_bts
	return ip1.bin(), interceptor.NF_ACCEPT

ictor = interceptor.Interceptor()
ictor.start(verdict_cb, queue_ids=[0, 1, 2])
print("now sind a ICMP echo request to localhost: ping")


  • Python 3.x (CPython, Pypy, Jython or whatever Interpreter)
  • Optional: netifaces >=0.10.6 (for utils)
  • Optional (for interceptor):
    • CPython
    • Linux based system with kernel support for NFQUEUE target. The config option is at:
      • Networking Options -> Network packet filtering -> Core Netfilter -> NFQUEUE target
    • iptables (alternatively nftables)
      • NFQUEUE related rulez can be added eg "iptables -I INPUT 1 -j NFQUEUE --queue-num 0"
    • libnetfilter_queue library (see


Some examples:

Usage examples

See examples/ and tests/


Tests are executed as follows:

  1. Add Pypacker directory to the PYTHONPATH.
  • cd pypacker
  • export PYTHONPATH=$(pwd):$PYTHONPATH
  1. execute tests
  • python tests/

Performance test results: pypacker

orC = Intel Core2 Duo CPU @ 1,866 GHz, 2GB RAM, CPython v3.6
orP = Intel Core2 Duo CPU @ 1,866 GHz, 2GB RAM, Pypy 5.10.1
rounds per test: 10000
>>> parsing (IP + ICMP)
orC = 86064 p/s
orP = 208346 p/s
>>> creating/direct assigning (IP only header)
orC = 41623 p/s
orP = 59370 p/s
>>> bin() without change (IP)
orC = 170356 p/s
orP = 292133 p/s
>>> output with change/checksum recalculation (IP)
orC = 10104 p/s
orP = 23851 p/s
>>> basic/first layer parsing (Ethernet + IP + TCP + HTTP)
orC = 62748 p/s
orP = 241047 p/s
>>> changing Triggerlist element value (Ethernet + IP + TCP + HTTP)
orC = 101552 p/s
orP = 201994 p/s
>>> changing Triggerlist/text based proto (Ethernet + IP + TCP + HTTP)
orC = 37249 p/s
orP = 272972 p/s
>>> direct assigning and concatination (Ethernet + IP + TCP + HTTP)
orC = 7428 p/s
orP = 14315 p/s
>>> full packet parsing (Ethernet + IP + TCP + HTTP)
orC = 6886 p/s
orP = 17040 p/s

Performance test results: pypacker vs. dpkt vs. scapy

Comparing pypacker, dpkt and scapy performance (parsing Ethernet + IP + TCP + HTTP)
orC = Intel Core2 Duo CPU @ 1,866 GHz, 2GB RAM, CPython v3.6
orC2 = Intel Core2 Duo CPU @ 1,866 GHz, 2GB RAM, CPython v2.7
rounds per test: 10000
>>> testing pypacker parsing speed
orC = 17938 p/s
>>> testing dpkt parsing speed
orC = 12431 p/s
>>> testing scapy parsing speed
orC2 = 726 p/s


Q: Where should I start learn to use Pypacker?

A: If you allready know Scapy starting by reading the examples should be OK. Otherwise there is a general introduction to pypacker included at the doc's which shows the usage and concepts of pypacker.

Q: How fast is pypacker?

A: See results above. For detailed results on your machine execute tests.

Q: Is there any documentation?

A: Pypacker is based on code of dpkt, which in turn didn't have any official and very little internal code documentation. This made understanding of the internal behaviour tricky. After all the code documentation was pretty much extended for Pypacker. Documentation can be found in these directories and files:

  • examples/ (many examples showing the usage of Pypacker)
  • wiki (general intro into pypacker)
  • (general Packet structure)

Protocols itself (see layerXYZ) generally don't have much documentation because those are documented by their respective RFCs/official standards.

Q: Which protocols are supported?

A: Currently minimum supported protocols are: Ethernet, Radiotap, IEEE80211, ARP, DNS, STP, PPP, OSPF, VRRP, DTP, IP, ICMP, PIM, IGMP, IPX, TCP, UDP, SCTP, HTTP, NTP, RTP, DHCP, RIP, SIP, Telnet, HSRP, Diameter, SSL, TPKT, Pmap, Radius, BGP

Q: How are protocols added?

A: Short answer: Extend Packet class and add the class variable __hdr__ to define header fields. Long answer: See examples/ for a very complete example.

Q: How can I contribute to this project?

A: Please use the Gitlab bug-tracker for bugs/feature request. Please read the bugtracker for already known bugs before filing a new one. Patches can be send via pull request.

Q: Under which license Pypacker is issued?

A: It's the GPLv2 License (see LICENSE file for more information).

Q: Are there any plans to support [protocol xyz]?

A: Support for particular protocols is added to Pypacker as a result of me needing this feature or people contributing that support - no formal plans for adding support for particular protocols in particular future releases exist.

Q: There is problem xyz with Pypacker using Windows 3.11/XP/7/8/mobile etc. Can you fix that?

A: The basic features should work with any OS. Optional ones may make trouble (eg interceptor).

Q: Calling copy.deepcopy(some_packet) raises an exception "TypeError: can't pickle Struct objects".

A: Try the following workaround to be able to pickle Struct objects:

import copy, copyreg
def pickle_struct(s):
	return struct.Struct, (s.format,)

copyreg.pickle(struct.Struct, pickle_struct)

Usage hints

Performance related

  • For maxmimum performance start accessing attributes at lowest level e.g. for filtering:
# This will lazy parse only needed layers behind the scenes
if ether.src == "...":
elif ip.src == "...":
elif == "...":
  • Avoid to convert packets using the "%s" or "%r" format as it triggers parsing behind the scene:
pkt = Ethernet() + IP() + TCP()
# This parses ALL layers
packet_print = "%s" % pkt
  • Avoid searching for a layer using single-value index-notation via pkt[L] as it parses all layers until L is found or highest layer is reached:
packet_found = pkt[Telnet]
# Alternative: Use multi-value index-notation. This will stop parsing at any non-matching layer:
packet_found = pkt[Ethernet,IP,TCP,Telnet]
  • Use pypy instead of cpython (~3x faster related to full packet parsing)

  • For even more performance disable auto fields (affects calling bin(...)):

pkt = ip.IP(src_s="", dst_s="") + tcp.TCP()
# Disable checksum calculation (and any other update) for IP and TCP (only THIS packet instance)
pkt.sum_au_active = False
pkt.tcp.sum_au_active = False
bts = pkt.bin(update_auto_fields=False)
sysctl -w net.core.rmem_max=12582912
sysctl -w net.core.rmem_default=12582912
sysctl -w net.core.wmem_max=12582912
sysctl -w net.core.wmem_default=12582912
sysctl -w net.core.optmem_max=2048000
sysctl -w net.core.netdev_max_backlog=5000
sysctl -w net.unix.max_dgram_qlen=1000
sysctl -w net.ipv4.tcp_rmem="10240 87380 12582912"
sysctl -w net.ipv4.tcp_wmem="10240 87380 12582912"
sysctl -w net.ipv4.tcp_mem="21228 87380 12582912"
sysctl -w net.ipv4.udp_mem="21228 87380 12582912"
sysctl -w net.ipv4.tcp_window_scaling=1
sysctl -w net.ipv4.tcp_timestamps=1
sysctl -w net.ipv4.tcp_sack=1

Misc related

  • Assemblation of TCP/UDP streams can be done by tshark using pipes with "-i -" and "-z follow,prot,mode,filter[,range]"