Title: IPv6 Transition Mechanisms, their Security and Management
1IPv6 Transition Mechanisms,their Security and
Management
- Georgios Koutepas
- National Technical University of Athens, Greece
- 6DISS Workshop
- March 5 2006
2Transition to IPv6
- Not an after-thought but designed to be part of
the new protocol since the beginning - Overview of transition requirements
- Gradual site transition a site may have only
some of its systems supporting IPv6 - Minimum transition requirements a site can
support IPv6 just by offering DNS services
without any upgrade in the rest of the
infrastructure - IP address compatibility the v4 addresses can be
converted to "corresponding" v6 addresses,
allowing the system to operate in both
environments - Ease of installation Operating Systems should
support IPv6 straightforwardly, without need for
software upgrades. - The answer SIT (Simple Internet Transition)
mechanisms included in IPv6
3IPv6 Transition Mechanisms
- SIT offers a scheme for
- The conversion of IPv4 addresses to IPv6
- Dual stack OS operation
- Tunnelling mechanisms via the encapsulation of v6
packets within v4 when passing over v4 clouds
(and vise-versa) - The Result
- Dual Stack mechanisms
- Translation Mechanisms
- Tunnelling Mechanisms
4Dual Stack mechanisms
5Translation Mechanisms
- NAT-PT (Network Address Translation - Protocol
Translation) - Potential problems
- Services based on protocol specific header info
cannot be supported end-to-end - "Classic" NAT security issues
- Others
- BIS (Bump in the Stack) - At the Transport Layer
- BIA (Bump in the API) - At the Application Layer
6Tunnelling Mechanisms
- How they work
- Encapsulation of IPv6 packets within IPv4 packets
and vice versa - Which means it can also be used for IPv4
connections over IPv6 native networks - Protocol in the IPv4 header 41
- The tunnel's end point performs the necessary
operations on the protocol 41 IPv4 packets - Reconnection of fragmented packets
- Packet forwarding in the IPv6 network
- Hop limit (equivalent to IPv4 TTL) reduction by
1 The tunnel is "transparent" to IPv6 - Nodes performing the (en/de)capsulation operation
have to be dual stack
7Types of tunnelling
- Based on the way we find the tunnel's other end
- (Pre)configured tunnel end-points
- Automatic. Tunnel end-point may be derived from
- 6to4 address
- IPv4 compatible IPv6 destination address
8Automatic Tunneling Mechanisms Tunnel Brokers
- The simplest way to IPv6 for single users (i.e.
using dialup, ADSL, etc.) - May create security problems OR opositely
protocol 41 may be banned by the sys-admins for
security reasons - Operation
- The user connects to a special web server (in the
IPv4 network) makes tunnel application - The server assigns an IPv6 address, creates a DNS
entry, informs the Tunnel Server, and sends a
configuration script to the user - The user runs the script, installs the
IPv6-over-IPv4 tunnel and onnects to the Tunnel
Server that routs the packets to the native IPv6
network
9Automatic Tunneling Mechanisms 6over4
- Deprecated...
- "Multicast tunnelling"
- Single IPv6 hosts use the IPv4 Multicast Network
to connect between them or the native IPv6
network via a 6over4 router (usually a 6to4
router) - The result is IPv6 hosts directly connected, even
using IPv6 Link Local addresses (derived
fromtheir IPv4 addresses)! - Also supports IPv6 multicast etc.
- 6over4 requires IPv4 Multicast support, which
does not exist widely.
10Automatic Tunneling Mechanisms ISATAP
- Intra Site automatic Tunnel Addressing Protocol
- Also uses the IPv4 infrastructure but without the
need for Multicast - Can operate under v4 NAT
- Operation
- The node (A.B.C.D)v4 gets the (FE805EFEABCD)v6
Link Local address - Using DNSv4 queries for the name ISATAP a
Potential Router List (PRL) is created (the
Router usually is a 6to4 system) - A Router Solicitation message is sent the answer
(Router-Advertisement message) gives the prefix
for creating the universal IPv6 address - ISATAP router-to-node communication using the
last 4 bytes of the destination address - Node-to-router IPv6 network via the ISATAP router
11Automatic Tunneling Mechanisms Teredo
- Useful for hosts behind NAT
- Encapsulates the IPv6 packets within UDP v4
packets to bypass the problem of NAT in many
cases restricting protocol 41 (IP encapsulated)
packets - The encapsulation takes place at the
communicating node itself rather than at a border
router (like it happens in 6to4) - The Teredo-relay then forwards the packets to the
native IPv6 network - Issues
- Complex implementation
- Can operate only with specific NAT types
- Limited number of Teredo-relays available in the
Internet - Used only there is no other available solution
12Automatic Tunneling Mechanisms 6to4 Overview
- Connects isolated IPv6 "clouds"
- Only the border routers need to implement the
6to4 functionality (and need to be dual stack
too) - Any site with single unicast IPv4 address can
transmit to the IPv6 network using the 2002/16
prefix - Many available relays to the IPv6 network, easy
to find by (IPv4) anycast addressing (from
192.88.99.0 - RFC 3068) - The most widely used mechanism, thanks to its
minimum requirements and ease of implementation
it is preferred to other automatic tunneling
methods and configured tunnels - However cannot be used behind NAT because it
requires an available universal IPv4 address
136to4 Architecture and Components
146to4 usage scenaria (1) 6to4 host to 6to4 host
- Native v6 communication and routing (RIPng)
156to4 usage scenaria (2) Between two 6to4 sites
- Useful for sites without native IPv6 ISP support
- Within the 6to4 sites the hosts use IPv6 natively
- Router advertisements and stateless address
autoconfiguration - DNSv6 host records - The other site can know
about the hosts it needs to communicate with - Non-local IPv6 addresses are sent to the default
(6to4) router - The IPv4 address within the 6to4 destination IPv6
address is used as the tunnel termination point
166to4 usage scenaria (2) Between two 6to4 sites
176to4 usage scenaria (3) Between a 6to4 site and
a native IPv6 network
- Connection to the native IPv6 network through a
6to4 Relay Router (an IPv6 router with a 6to4
"Pseudo-interface") - Usage of the Relay Router's IPv4 address or the
Anycast Address - 6to4 host to a native IPv6 host
- The 6to4 host uses DNS to find the destination
host - The 6to4 router forwards (via IPv4) the packet to
the "next-hop", the closest 6to4 relay router - The IPv6 router forward the packet to its final
destination - Native IPv6 host to a 6to4 host
- The 6to4 relay router advertises the 2002/16
prefix within the IPv6 network - A v6 host will use this information to send its
packet to the corresponding IPv6 router and
further to the 6to4 "pseudo-interface" via which
(by the IPv4 network) the packet reaches the 6to4
network and its final destination
186to4 usage scenaria (3) Between a 6to4 site and
a native IPv6 network
196to4 Securityor what can go wrong
- Vulnerabilities
- 6to4 routers must accept packets from ALL 6to4
relay routers - It's not possible to know if the relay router is
"Trusted" or even existent - 6to4 relay routers have to accept packets from
6to4 routers and native IPv6 hosts without any
checks - Threats
- DoS/DDoS against 6to4 components may result in
unavailability - 6to4 routers/relay routers may be used or
"reflected" DDoS attacks - "Service theft" unauthorized usage of relay
router services - Local IPv4 broadcast attacks
- Neighbor Discovery attacks
- "Sanity Checks" necessary!
206to4 Security an attack scenario
- Reflected DoS Attack
- It is supposed that bandwidth and processing
power limitations can prevent a large scale
attack
21Securing 6to4 components
- 6to4 routers
- Check for correspondence between the IPv4 part of
the packets and the 2002/16 IPv6 encapsulated
part - Implement "Sanity Checks"
- IPv4 Do not allow strange (e.g. loopback)
private, multicast, etc. addresses to be
encapsulated - IPv6 Reject "wrong" addresses, like link local,
multicast, etc. - Prevent routing of packets to other 6to4 sites
via 6to4 relay routers - Reject packets coming from another 6to4 site via
a relay router
22Securing 6to4 components (2)
- 6to4 relay routers
- Reject IPv4 packets from 6to4 routers that don't
have matching IPv4 src address (V4ADDR) and
equivalent 6to4 src address (2002V4ADR) in the
encapsulated IPv6 packet - Reject protocol 41 (IPv4) packets without
destination address 192.88.99.1 - Deny packets to the IPv6 network without a
universal IPv6 address - Reject packets from 6to4 routers to 6to4
addresses - Ingress Filtering and Access Control Lists for
the IPv6 part!
23A General Transition Roadmapfor an enterprise or
educational network
- Phase 1
- Network Design
- Define Wide and Local network segments
- Define special areas (due to requirements and
operations) - VLANs, DMZs etc. - Define management entities and their areas of
responsibility - Network management information flow
- Security requirements
- For users and applications
- For the network itself (protection of the
management information, protection of network
devices, security of management procedures) - Plan the steps to transition to the new protocol.
Examine the possibility of deploying transition
mechanisms (for communications between IPv6 areas
within an IPv4 network and vise-versa)
24A General Transition Roadmap (2)
- Phase 2
- Implementation of a mixed IPv4/IPv6 environment
- Gradual transition of non-critical systems to
IPv6 - Allows the evaluation of the operation and
stability of the network devices and non-critical
systems under IPv6 - Develops the transition procedures
- Disseminates the usages of transition mechanisms
(tunnels, gateways, etc.) for communications
between exclusive IPv6 areas - Phase 3
- Transition of all systems to IPv6
- Exclusive usage of IPv6 in the network
- Maintaining transition mechanisms for legacy
systems and contacts with IPv4 networks
25Any Questions ?