Transcript Modern Internet architecture & technology

Network Address
Translation (NAT)
1
Overview
Motivation
 End-to-end principle
 Role of IP addresses
 Basic NAT types and their behaviors
 NAT traversal: STUN

2
History

Early 1990s


IPv4 Address consumption concern
Two approaches
• IPv6 and NAT

NATs were initially intended to allow devices to
share an address pool dynamically



NAT goes against Internet end-to-end principle


3
First RFC about NAT in 1994
NAT vs. DHCP?
IETF hates NATs
No standardization -> backfire
Motivation

DSL and cable modem business model


ISP wants to save money


Even multi-homing
Security: Inbound traffic filtering

4
In PSTNs, there is extension
Changing next higher ISP becomes easier


Not simultaneous access, no servers
stateful firewall
End-to-end principle

5
RFC 1958: “An end-to-end protocol design
should not rely on the maintenance of state (i.e.,
information about the state of the end-to-end
communication) inside the network. Such state
should be maintained only in the endpoints, in
such a way that the state can only be destroyed
when the endpoint itself breaks (known as fatesharing). An immediate consequence of this is
that datagrams are better than classical virtual
circuits. The network's job is to transmit
datagrams as efficiently and flexibly as possible.
Everything else should be done at the fringes.”
Middle boxes







6
Middle box = “any intermediary
device performing functions other
than the normal, standard
functions of an IP router on the
datagram path between a source
host and destination host”
L2, L3, L4, L7, …
explicit vs. implicit
functional vs. optimizing
routing vs. processing
soft vs. hard state
fail-over vs. restart

Examples:
 NATs










SOCKS gateway
IP tunnel endpoint
Transport relay
Load balancers
Application firewalls
Transcoders (RFC 3234)
Proxies
Caches
Modified DNS servers
Anonymizers
Private Network


Private IP network is an IP network that is not
directly connected to the Internet
IP addresses in a private network can be
assigned arbitrarily.


Generally, private networks use addresses
from the following experimental address
ranges (non-routable addresses):



7
Not registered and not guaranteed to be globally
unique
10.0.0.0 – 10.255.255.255
172.16.0.0 – 172.31.255.255
192.168.0.0 – 192.168.255.255
Private Addresses
H1
10.0.1.2
H3
H2
10.0.1.2
10.0.1.3
10.0.1.1
10.0.1.3
10.0.1.1
Private network 1
Private network 1
Internet
R1
128.195.4.119
128.143.71.21
213.168.112.3
H5
8
H4
R2
Network Address Translation (NAT)



NAT is a router function where IP addresses (and
possibly port numbers) of IP datagrams are replaced
at the boundary of a private network
NAT is a method that enables hosts on private
networks to communicate with hosts on the Internet
NAT is run on routers that connect private networks
to the public Internet, to replace the IP address-port
pair of an IP packet with another IP address-port
pair.

Topology sensitive
• inside (private) vs. outside (public)
9
Basic operation of NAT

10
NAT device has address translation table
Main uses of NAT
11

Pooling of IP addresses

Supporting migration between
network service providers

IP masquerading

Load balancing of servers
Pooling of IP addresses


Scenario: Corporate network has many hosts but only
a small number of public IP addresses
NAT solution:



12
Corporate network is managed with a private address
space
NAT device, located at the boundary between the
corporate network and the public Internet, manages a
pool of public IP addresses
When a host from the corporate network sends an IP
datagram to a host in the public Internet, the NAT device
picks a public IP address from the address pool, and
binds this address to the private address of the host
Pooling of IP addresses
Private
network
Internet
Source
= 10.0.1.2
Destination = 213.168.112.3
Source
= 128.143.71.21
Destination = 213.168.112.3
NAT
device
private address: 10.0.1.2
public address:
H1
213.168.112.3
H5
Private
Address
Public
Address
10.0.1.2
Pool of addresses: 128.143.71.0-128.143.71.30
13
public address:
Supporting migration between network
service providers


Scenario: In CIDR, the IP addresses in a corporate network are
obtained from the service provider. Changing the service provider
requires changing all IP addresses in the network.
NAT solution:
Assign private addresses to the hosts of the corporate network
 NAT device has static address translation entries which bind the
private address of a host to the public address.
 Migration to a new network service provider merely requires an
update of the NAT device. The migration is not noticeable to the
hosts on the network.
Note:
 The difference to the use of NAT with IP address pooling is that the
mapping of public and private IP addresses is static.

14
Supporting migration between network
service providers
15
IP masquerading



Also called: Network address and port translation
(NAPT)
Scenario: Single public IP address is mapped to
multiple hosts in a private network.
NAT solution:


16
Assign private addresses to the hosts of the corporate
network
NAT device modifies the port numbers for outgoing
traffic
IP masquerading
17
Load balancing of servers


Scenario: Balance the load on a set of identical
servers, which are accessible from a single IP
address
NAT solution:




18
Here, the servers are assigned private addresses
NAT device acts as a proxy for requests to the server
from the public network
The NAT device changes the destination IP address of
arriving packets to one of the private addresses for a
server
A sensible strategy for balancing the load of the servers
is to assign the addresses of the servers in a roundrobin fashion.
Load balancing of servers
19
Concerns about NAT

Performance:
 Modifying the IP header by
changing the IP address
requires that NAT boxes
recalculate the IP header
checksum
 Modifying port number requires
that NAT boxes recalculate
TCP checksum

20
Fragmentation
 Care must be taken that a datagram that is fragmented before it
reaches the NAT device, is not assigned a different IP address or
different port numbers for each of the fragments.
Concerns about NAT

End-to-end connectivity:




21
NAT destroys universal end-to-end reachability of
hosts on the Internet.
A host in the public Internet often cannot initiate
communication to a host in a private network
• Hamper peer-to-peer applications
The problem is worse, when two hosts that are in
a private network need to communicate with each
other
Typically, the address-port mapping is maintained
soft-state (in minutes)
Concerns about NAT

IP address in application data:


22
Applications that carry IP addresses in the
payload of the application data generally do
not work across a private-public network
boundary.
Some NAT devices inspect the payload of
widely used application layer protocols and, if
an IP address is detected in the applicationlayer header or the application payload,
translate the address according to the address
translation table.
NAT and FTP

23
Normal FTP operation
NAT and FTP

24
NAT device with FTP support
NAT and FTP

25
FTP in passive mode and NAT.
NAT Traversal
26
NAPT Traversal
27
NAT types
Symmetric
 Port restricted cone
 Address restricted cone
 Full cone
 Hairpin

Different not only on a vendor-by-vendor basis but also on a model-by-model basis
28
Symmetric NAT
29
•NAT mapping btw src_addr/src_port and dest_addr/dest_port is fixed
•The most restrictive form
•It has been observed to be rare
Full-cone NAT


30
The least restrictive form
Binding of a local address/port can be used by any remote host
(Address) Restricted-cone NAT

31
NAT binding is accessible only by the destination host
 different port can be used
Port-restricted-cone NAT

NAT binding is accessible by any remote host


32
But only same port should be used
Typically, the internal host had previously sent a packet the remote host
Hairpin NAT
D

33
A local host can direct a packet to the public
address/port of an already mapped local host
Nondeterministic NATs
NATs
change their types of behavior when
a binding conflict occurs
Example
 Some
NATs attempt to preserve the port
number in the binding
 If the port number is available, the NAT
behaves like a full-cone NAT
 If that port is already occupied by other host,
the NAT may change the type, e.g. a
symmetric NAT
34
What is STUN? Who uses it?
STUN – Simple Traversal of User
Datagram Protocol through Network
Address Translators.
 The protocol is defined in RFC 3489.
 Protocols like SIP and applications
like Google Talk use STUN to gather
important information about the
network configuration.

35
What does it do?



36
STUN is a client-server protocol that “allows
entities behind a NAT to first discover the
presence of a NAT and the type of NAT, and
then to learn the addresses bindings
allocated by the NAT.”
In other words, it’s a means of discovering
the public IP and port numbers that a NAT
assigns to a node on a private LAN.
In addition, STUN does not require any
special network configuration and works
with a variety of existing networks, but not
all.
*http://tools.ietf.org/html/rfc3489
STUN and NAT terminology
A STUN Client is a node that
generates the STUN requests.
 A STUN Server is a node that
receives the STUN requests and
generates the STUN responses.
 NAT


37
usually part of a firewall or router
A picture i worth 1000 words
38
How STUN generally works

A STUN Client sends a STUN request to a
STUN Server. The Client then waits for the
Server to send a STUN response



39
STUN client is typically embedded in application
STUN server has two IP addresses
The trick is to analyze the response from the
server to determine the type of NAT router and
the associated bindings the router has given to
internal nodes.
The STUN Message

40
The following STUN attributes in the payload are
especially important:
 MAPPED-ADDRESS : Found in STUN responses. It
contains the IP address and port number of the STUN
request. I.e., the public IP and port of the STUN client.
 CHANGE-REQUEST : Found in STUN requests. It
contains flags for the IP address and port number of
the server. If set, the client is asking the server to
send the response from a different IP and port. (We
will see why later)
 CHANGED-ADDRESS – Found in STUN responses.
It contains the alternate IP address and port number
of the server due to CHANGE-REQUEST
NAT discovery (test 1)


41
To determine if a NAT router/firewall is
present, send a STUN request to the server.
Wait for a response and analyze it.
If the IP address and port number in the
MAPPED-ADDRESS attribute of the
payload in the STUN response equal the
local IP address and port number that it
bound to when sending the request, then
the client is NOT behind a NAT router.
Otherwise, it is behind a NAT router.
NAT discovery – Full Cone (test 2)

Full Cone NAT router – The client sets the
IP address and port number flags in the
CHANGE-REQUEST of the STUN request.
This causes the server to send the
response from the alternate IP and port
number.


42
If the client receives the STUN response,
then the client is behind a full cone router.
Otherwise, it is behind one of the other three
NAT routers.
NAT discovery – Symmetric (test 3)

Symmetric NAT – The client sends two
STUN requests. One request is sent to a
server at IP address X and port P, and
another to a server at IP address Y and port
P.


43
If the IP addresses and ports from the
MAPPED-ADDRESS attributes in the two
responses do not match, then it is behind a
Symmetric NAT router.
If they do match, then it is behind one of the
remaining two NAT routers.
NAT discovery – Restricted (test 4)

Restricted NAT – The port flag in the
CHANGE-REQUEST attribute of the
request is set. This instructs the server to
send a response from a different port.

44
If the response is received, it is behind a
restricted NAT router. If no response is
received, it is behind a port restricted NAT
router.
45
Limitations of STUN

Does not address incoming TCP connections.




46
STUNT and other proposals
Does not allow incoming UDP connections through a symmetric
NAT
STUN “is not a cure-all for the problems associated with NAT.”
“The problems in STUN have to do with the lack of standardized
behaviors and controls in NATs, which results in a proliferation of
devices whose behavior is highly unpredictable, extremely
variable, and uncontrollable. Ultimately, the solution is to make
the environment less hostile, and to introduce controls and
standardized behaviors into NAT. However, until such time as that
happens, STUN provides a good short term solution given the
terrible conditions under which it is forced to operate.”
NAT traversal
47
48
49
50
51
52
53
54
55
56
57
58
59
Other issues

Symmetric NATs
The first packet for hole punching will
be dropped
 And the port number (for the other
peer) is changed from the one for the
server
 Port prediction technique

• May not work

60
Nested NATs