RFC 7707
Network Reconnaissance in IPv6 Networks
Internet Engineering Task Force (IETF) F. Gont Request for Comments: 7707 Huawei Technologies Obsoletes: 5157 T. Chown Category: Informational Jisc ISSN: 2070-1721 March 2016 Network Reconnaissance in IPv6 NetworksAbstract
IPv6 offers a much larger address space than that of its IPv4 counterpart. An IPv6 subnet of size /64 can (in theory) accommodate approximately 1.844 * 10^19 hosts, thus resulting in a much lower host density (#hosts/#addresses) than is typical in IPv4 networks, where a site typically has 65,000 or fewer unique addresses. As a result, it is widely assumed that it would take a tremendous effort to perform address-scanning attacks against IPv6 networks; therefore, IPv6 address-scanning attacks have been considered unfeasible. This document formally obsoletes RFC 5157, which first discussed this assumption, by providing further analysis on how traditional address- scanning techniques apply to IPv6 networks and exploring some additional techniques that can be employed for IPv6 network reconnaissance. Status of This Memo This document is not an Internet Standards Track specification; it is published for informational purposes. This document is a product of the Internet Engineering Task Force (IETF). It represents the consensus of the IETF community. It has received public review and has been approved for publication by the Internet Engineering Steering Group (IESG). Not all documents approved by the IESG are a candidate for any level of Internet Standard; see Section 2 of RFC 5741. Information about the current status of this document, any errata, and how to provide feedback on it may be obtained at http://www.rfc-editor.org/info/rfc7707.
Copyright Notice Copyright (c) 2016 IETF Trust and the persons identified as the document authors. All rights reserved. This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (http://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 3 2. Conventions . . . . . . . . . . . . . . . . . . . . . . . . . 4 3. Requirements for the Applicability of Network Reconnaissance Techniques . . . . . . . . . . . . . . . . . . . . . . . . . 4 4. IPv6 Address Scanning . . . . . . . . . . . . . . . . . . . . 6 4.1. Address Configuration in IPv6 . . . . . . . . . . . . . . 6 4.1.1. Stateless Address Autoconfiguration (SLAAC) . . . . . 6 4.1.2. Dynamic Host Configuration Protocol for IPv6 (DHCPv6) 11 4.1.3. Manually Configured Addresses . . . . . . . . . . . . 12 4.1.4. IPv6 Addresses Corresponding to Transition/Coexistence Technologies . . . . . . . . . 14 4.1.5. IPv6 Address Assignment in Real-World Network Scenarios . . . . . . . . . . . . . . . . . . . . . . 14 4.2. IPv6 Address Scanning of Remote Networks . . . . . . . . 17 4.2.1. Reducing the Subnet ID Search Space . . . . . . . . . 18 4.3. IPv6 Address Scanning of Local Networks . . . . . . . . . 19 4.4. Existing IPv6 Address-Scanning Tools . . . . . . . . . . 20 4.4.1. Remote IPv6 Network Address Scanners . . . . . . . . 20 4.4.2. Local IPv6 Network Address Scanners . . . . . . . . . 21 4.5. Mitigations . . . . . . . . . . . . . . . . . . . . . . . 21 4.6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . 22 5. Alternative Methods to Glean IPv6 Addresses . . . . . . . . . 23 5.1. Leveraging the Domain Name System (DNS) for Network Reconnaissance . . . . . . . . . . . . . . . . . . . . . 23 5.1.1. DNS Advertised Hosts . . . . . . . . . . . . . . . . 23 5.1.2. DNS Zone Transfers . . . . . . . . . . . . . . . . . 23 5.1.3. DNS Brute Forcing . . . . . . . . . . . . . . . . . . 23 5.1.4. DNS Reverse Mappings . . . . . . . . . . . . . . . . 24 5.2. Leveraging Local Name Resolution and Service Discovery Services . . . . . . . . . . . . . . . . . . . . . . . . 24 5.3. Public Archives . . . . . . . . . . . . . . . . . . . . . 25
5.4. Application Participation . . . . . . . . . . . . . . . . 25
5.5. Inspection of the IPv6 Neighbor Cache and Routing Table . 25
5.6. Inspection of System Configuration and Log Files . . . . 26
5.7. Gleaning Information from Routing Protocols . . . . . . . 26
5.8. Gleaning Information from IP Flow Information Export
(IPFIX) . . . . . . . . . . . . . . . . . . . . . . . . . 26
5.9. Obtaining Network Information with traceroute6 . . . . . 26
5.10. Gleaning Information from Network Devices Using SNMP . . 27
5.11. Obtaining Network Information via Traffic Snooping . . . 27
6. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . 27
7. Security Considerations . . . . . . . . . . . . . . . . . . . 27
8. References . . . . . . . . . . . . . . . . . . . . . . . . . 28
8.1. Normative References . . . . . . . . . . . . . . . . . . 28
8.2. Informative References . . . . . . . . . . . . . . . . . 29
Appendix A. Implementation of a Full-Fledged IPv6 Address-
Scanning Tool . . . . . . . . . . . . . . . . . . . 34
A.1. Host-Probing Considerations . . . . . . . . . . . . . . . 34
A.2. Implementation of an IPv6 Local Address-Scanning Tool . . 35
A.3. Implementation of an IPv6 Remote Address-Scanning Tool . 36
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . 37
Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 38
1. Introduction
The main driver for IPv6 [RFC2460] deployment is its larger address
space [CPNI-IPv6]. This larger address space not only allows for an
increased number of connected devices but also introduces a number of
subtle changes in several aspects of the resulting networks. One of
these changes is the reduced host density (the number of hosts
divided by the number of addresses) of typical IPv6 subnetworks, when
compared to their IPv4 counterparts. [RFC5157] describes how this
significantly lower IPv6 host density is likely to make classic
network address-scanning attacks less feasible, since even by
applying various heuristics, the address space to be scanned remains
very large. RFC 5157 goes on to describe some alternative methods
for attackers to glean active IPv6 addresses and provides some
guidance for administrators and implementors, e.g., not using
sequential addresses with DHCPv6.
With the benefit of more than five years of additional IPv6
deployment experience, this document formally obsoletes RFC 5157. It
emphasizes that while address-scanning attacks are less feasible,
they may, with appropriate heuristics, remain possible. At the time
that RFC 5157 was written, observed address-scanning attacks were
typically across ports on the addresses of discovered servers; since
then, evidence that some classic address scanning is occurring is
being witnessed. This text thus updates the analysis on the
feasibility of address-scanning attacks in IPv6 networks, and it
explores a number of additional techniques that can be employed for IPv6 network reconnaissance. Practical examples and guidance are also included in the appendices. On one hand, raising awareness about IPv6 network reconnaissance techniques may allow (in some cases) network and security administrators to prevent or detect such attempts. On the other hand, network reconnaissance is essential for the so-called "penetration tests" typically performed to assess the security of production networks. As a result, we believe the benefits of a thorough discussion of IPv6 network reconnaissance are twofold. Section 4 analyzes the feasibility of address-scanning attacks (e.g., ping sweeps) in IPv6 networks and explores a number of possible improvements to such techniques. Appendix A describes how the aforementioned analysis can be leveraged to produce address-scanning tools (e.g., for penetration testing purposes). Finally, the rest of this document discusses a number of miscellaneous techniques that could be leveraged for IPv6 network reconnaissance.2. Conventions
Throughout this document, we consider that bits are numbered from left to right, starting at 0, and that bytes are numbered from left to right, starting at 0.3. Requirements for the Applicability of Network Reconnaissance Techniques
Throughout this document, a number of network reconnaissance techniques are discussed. Each of these techniques has different requirements on the side of the practitioner, with respect to whether they require local access to the target network and whether they require login access (or similar access credentials) to the system on which the technique is applied. The following table tries to summarize the aforementioned requirements and serves as a cross index to the corresponding sections.
+---------------------------------------------+----------+----------+ | Technique | Local | Login | | | access | access | +---------------------------------------------+----------+----------+ | Remote Address Scanning (Section 4.2) | No | No | +---------------------------------------------+----------+----------+ | Local Address Scanning (Section 4.3) | Yes | No | +---------------------------------------------+----------+----------+ | DNS Advertised Hosts (Section 5.1.1) | No | No | +---------------------------------------------+----------+----------+ | DNS Zone Transfers (Section 5.1.2) | No | No | +---------------------------------------------+----------+----------+ | DNS Brute Forcing (Section 5.1.3) | No | No | +---------------------------------------------+----------+----------+ | DNS Reverse Mappings (Section 5.1.4) | No | No | +---------------------------------------------+----------+----------+ | Leveraging Local Name Resolution and | Yes | No | | Service Discovery Services (Section 5.2) | | | +---------------------------------------------+----------+----------+ | Public Archives (Section 5.3) | No | No | +---------------------------------------------+----------+----------+ | Application Participation (Section 5.4) | No | No | +---------------------------------------------+----------+----------+ | Inspection of the IPv6 Neighbor Cache and | No | Yes | | Routing Table (Section 5.5) | | | +---------------------------------------------+----------+----------+ | Inspecting System Configuration and Log | No | Yes | | Files (Section 5.6) | | | +---------------------------------------------+----------+----------+ | Gleaning Information from Routing Protocols | Yes | No | | (Section 5.7) | | | +---------------------------------------------+----------+----------+ | Gleaning Information from IP Flow | No | Yes | | Information Export (IPFIX) (Section 5.8) | | | +---------------------------------------------+----------+----------+ | Obtaining Network Information with | No | No | | traceroute6 (Section 5.9) | | | +---------------------------------------------+----------+----------+ | Gleaning Information from Network Devices | No | Yes | | Using SNMP (Section 5.10) | | | +---------------------------------------------+----------+----------+ | Obtaining Network Information via Traffic | Yes | No | | Snooping (Section 5.11) | | | +---------------------------------------------+----------+----------+ Table 1: Requirements for the Applicability of Network Reconnaissance Techniques
4. IPv6 Address Scanning
This section discusses how traditional address-scanning techniques (e.g., "ping sweeps") apply to IPv6 networks. Section 4.1 provides an essential analysis of how address configuration is performed in IPv6, identifying patterns in IPv6 addresses that can be leveraged to reduce the IPv6 address search space when performing IPv6 address- scanning attacks. Section 4.2 discusses IPv6 address scanning of remote networks. Section 4.3 discusses IPv6 address scanning of local networks. Section 4.4 discusses existing IPv6 address-scanning tools. Section 4.5 provides advice on how to mitigate IPv6 address- scanning attacks. Finally, Appendix A discusses how the insights obtained in the following subsections can be incorporated into a fully fledged IPv6 address-scanning tool.4.1. Address Configuration in IPv6
IPv6 incorporates two automatic address-configuration mechanisms: Stateless Address Autoconfiguration (SLAAC) [RFC4862] and Dynamic Host Configuration Protocol for IPv6 (DHCPv6) [RFC3315]. Support for SLAAC for automatic address configuration is mandatory, while support for DHCPv6 is optional -- however, most current versions of general- purpose operating systems support both. In addition to automatic address configuration, hosts, typically servers, may employ manual configuration, in which all the necessary information is manually entered by the host or network administrator into configuration files at the host. The following subsections describe each of the possible configuration mechanisms/approaches in more detail.4.1.1. Stateless Address Autoconfiguration (SLAAC)
The basic idea behind SLAAC is that every host joining a network will send a multicasted solicitation requesting network configuration information, and local routers will respond to the request providing the necessary information. SLAAC employs two different ICMPv6 message types: ICMPv6 Router Solicitation and ICMPv6 Router Advertisement messages. Router Solicitation messages are employed by hosts to query local routers for configuration information, while Router Advertisement messages are employed by local routers to convey the requested information.
Router Advertisement messages convey a plethora of network configuration information, including the IPv6 prefix that should be used for configuring IPv6 addresses on the local network. For each local prefix learned from a Router Advertisement message, an IPv6 address is configured by appending a locally generated Interface Identifier (IID) to the corresponding IPv6 prefix. The following subsections describe currently deployed policies for generating the IIDs used with SLAAC.4.1.1.1. Interface Identifiers Embedding IEEE Identifiers
The traditional SLAAC IIDs are based on the link-layer address of the corresponding network interface card. For example, in the case of Ethernet addresses, the IIDs are constructed as follows: 1. The "Universal" bit (bit 6, from left to right) of the address is set to 1. 2. The word 0xfffe is inserted between the Organizationally Unique Identifier (OUI) and the rest of the Ethernet address. For example, the Media Access Control (MAC) address 00:1b:38:83:88:3c would lead to the IID 021b:38ff:fe83:883c. A number of considerations should be made about these identifiers. Firstly, one 16-bit word (bytes 3-4) of the resulting address always has a fixed value (0xfffe), thus reducing the search space for the IID. Secondly, the high-order three bytes of the IID correspond to the OUI of the network interface card vendor. Since not all possible OUIs have been assigned, this further reduces the IID search space. Furthermore, of the assigned OUIs, many could be regarded as corresponding to legacy devices and thus are unlikely to be used for Internet-connected IPv6-enabled systems, yet further reducing the IID search space. Finally, in some scenarios, it could be possible to infer the OUI in use by the target network devices, yet narrowing down the possible IIDs even more. NOTE: For example, an organization known for being provisioned by vendor X is likely to have most of the nodes in its organizational network with OUIs corresponding to vendor X. These considerations mean that in some scenarios, the original IID search space of 64 bits may be effectively reduced to 2^24 or n * 2^24 (where "n" is the number of different OUIs assigned to the target vendor).
Furthermore, if just one host address is detected or known within a
subnet, it is not unlikely that, if systems were ordered in a batch,
they may have sequential MAC addresses. Additionally, given a MAC
address observed in one subnet, sequential or nearby MAC addresses
may be seen in other subnets in the same site.
NOTE:
[RFC7136] notes that all bits of an IID should be treated as
"opaque" bits. Furthermore, [DEFAULT-IIDS] is currently in the
process of changing the default IID generation scheme to align
with [RFC7217] (as described below in Section 4.1.1.5), such that
IIDs are semantically opaque and do not follow any patterns.
Therefore, the traditional IIDs based on link-layer addresses are
expected to become less common over time.
4.1.1.2. Interface Identifiers of Virtualization Technologies
IIDs resulting from virtualization technologies can be considered a
specific subcase of IIDs embedding IEEE identifiers (please see
Section 4.1.1.1): they employ IEEE identifiers, but part of the IID
has specific patterns. The following subsections describe IIDs of
some popular virtualization technologies.
4.1.1.2.1. VirtualBox
All automatically generated MAC addresses in VirtualBox virtual
machines employ the OUI 08:00:27 [VBox2011]. This means that all
addresses resulting from traditional SLAAC will have an IID of the
form a00:27ff:feXX:XXXX, thus effectively reducing the IID search
space from 64 bits to 24 bits.
4.1.1.2.2. VMware ESX Server
The VMware ESX server (versions 1.0 to 2.5) provides yet a more
interesting example. Automatically generated MAC addresses have the
following pattern [vmesx2011]:
1. The OUI is set to 00:05:69.
2. The next 16 bits of the MAC address are set to the same value as
the last 16 bits of the console operating system's primary IPv4
address.
3. The final 8 bits of the MAC address are set to a hash value based
on the name of the virtual machine's configuration file.
This means that, assuming the console operating system's primary IPv4 address is known, the IID search space is reduced from 64 bits to 8 bits. On the other hand, manually configured MAC addresses in the VMware ESX server employ the OUI 00:50:56, with the low-order three bytes of the MAC address being in the range 00:00:00-3F:FF:FF (to avoid conflicts with other VMware products). Therefore, even in the case of manually configured MAC addresses, the IID search space is reduced from 64 bits to 22 bits.4.1.1.2.3. VMware vSphere
VMware vSphere [vSphere] supports these default MAC address generation algorithms: o Generated addresses * Assigned by the vCenter server * Assigned by the ESXi host o Manually configured addresses By default, MAC addresses assigned by the vCenter server use the OUI 00:50:56 and have the format 00:50:56:XX:YY:ZZ, where XX is calculated as (0x80 + vCenter Server ID (in the range 0x00-0x3F)), and XX and YY are random two-digit hexadecimal numbers. Thus, the possible IID range is 00:50:56:80:00:00-00:50:56:BF:FF:FF; therefore, the search space for the resulting SLAAC addresses will be 22 bits. MAC addresses generated by the ESXi host use the OUI 00:0C:29 and have the format 00:0C:29:XX:YY:ZZ, where XX, YY, and ZZ are the last three octets in hexadecimal format of the virtual machine Universally Unique Identifier (UUID) (based on a hash calculated with the UUID of the ESXi physical machine and the path to a configuration file). Thus, the MAC addresses will be in the range 00:0C:29:00:00:00-00:0C:29:FF:FF:FF; therefore, the search space for the resulting SLAAC addresses will be 24 bits. Finally, manually configured MAC addresses employ the OUI 00:50:56, with the low-order three bytes being in the range 00:00:00-3F:FF:FF (to avoid conflicts with other VMware products). Therefore, the resulting MAC addresses will be in the range 00:50:56:00:00:00-00:50:56:3F:FF:FF, and the search space for the corresponding SLAAC addresses will be 22 bits.
4.1.1.3. Temporary Addresses
Privacy concerns [Gont-DEEPSEC2011] [RFC7721] regarding IIDs embedding IEEE identifiers led to the introduction of "Privacy Extensions for Stateless Address Autoconfiguration in IPv6" [RFC4941], also known as "temporary addresses" or "privacy addresses". Essentially, "temporary addresses" produce random addresses by concatenating a random identifier to the autoconfiguration IPv6 prefix advertised in a Router Advertisement message. NOTE: In addition to their unpredictability, these addresses are typically short-lived, such that even if an attacker were to learn of one of these addresses, they would be of use for a limited period of time. A typical implementation may keep a temporary address preferred for 24 hours, and configured but deprecated for seven days. It is important to note that "temporary addresses" are generated in addition to the stable addresses [RFC7721] (such as the traditional SLAAC addresses based on IEEE identifiers): stable SLAAC addresses are meant to be employed for "server-like" inbound communications, while "temporary addresses" are meant to be employed for "client- like" outbound communications. This means that implementation/use of "temporary addresses" does not prevent an attacker from leveraging the predictability of stable SLAAC addresses, since "temporary addresses" are generated in addition to (rather than as a replacement of) the stable SLAAC addresses (such as those derived from IEEE identifiers). The benefit that temporary addresses offer in this context is that they reduce the exposure of the host addresses to any third parties that may observe traffic sent from a host where temporary addresses are enabled and used by default. But, in the absence of firewall protection for the host, its stable SLAAC address remains liable to be scanned from off-site.4.1.1.4. Constant, Semantically Opaque IIDs
In order to mitigate the security implications arising from the predictable IPv6 addresses derived from IEEE identifiers, Microsoft Windows produced an alternative scheme for generating "stable addresses" (in replacement of the ones embedding IEEE identifiers). The aforementioned scheme is believed to be an implementation of RFC 4941 [RFC4941], but without regenerating the addresses over time. The resulting IIDs are constant across system bootstraps, and also constant across networks.
Assuming no flaws in the aforementioned algorithm, this scheme would
remove any patterns from the SLAAC addresses.
NOTE:
However, since the resulting IIDs are constant across networks,
these addresses may still be leveraged for host-tracking purposes
[RFC7217] [RFC7721].
The benefit of this scheme is thus that the host may be less readily
detected by applying heuristics to an address-scanning attack, but,
in the absence of concurrent use of temporary addresses, the host is
liable to be tracked across visited networks.
4.1.1.5. Stable, Semantically Opaque IIDs
In response to the predictability issues discussed in Section 4.1.1.1
and the privacy issues discussed in [RFC7721], the IETF has
standardized (in [RFC7217]) a method for generating IPv6 IIDs to be
used with IPv6 SLAAC, such that addresses configured using this
method are stable within each subnet, but the IIDs change when hosts
move from one subnet to another. The aforementioned method is meant
to be an alternative to generating IIDs based on IEEE identifiers,
such that the benefits of stable addresses can be achieved without
sacrificing the privacy of users.
Implementation of this method (in replacement of IIDs based on IEEE
identifiers) eliminates any patterns from the IID, thus benefiting
user privacy and reducing the ease with which addresses can be
scanned.
4.1.2. Dynamic Host Configuration Protocol for IPv6 (DHCPv6)
DHCPv6 can be employed as a stateful address configuration mechanism,
in which a server (the DHCPv6 server) leases IPv6 addresses to IPv6
hosts. As with the IPv4 counterpart, addresses are assigned
according to a configuration-defined address range and policy, with
some DHCPv6 servers assigning addresses sequentially, from a specific
range. In such cases, addresses tend to be predictable.
NOTE:
For example, if the prefix 2001:db8::/64 is used for assigning
addresses on the local network, the DHCPv6 server might
(sequentially) assign addresses from the range 2001:db8::1 -
2001:db8::100.
In most common scenarios, this means that the IID search space will
be reduced from the original 64 bits to 8 or 16 bits. [RFC5157]
recommended that DHCPv6 instead issue addresses randomly from a large
pool; that advice is repeated here. [IIDS-DHCPv6] specifies an algorithm that can be employed by DHCPv6 servers to produce stable addresses that do not follow any specific pattern, thus resulting in an IID search space of 64 bits.4.1.3. Manually Configured Addresses
In some scenarios, node addresses may be manually configured. This is typically the case for IPv6 addresses assigned to routers (since routers do not employ automatic address configuration) but also for servers (since having a stable address that does not depend on the underlying link-layer address is generally desirable). While network administrators are mostly free to select the IID from any value in the range 1 - 2^64, for the sake of simplicity (i.e., ease of remembering), they tend to select addresses with one of the following patterns: o low-byte addresses: in which most of the bytes of the IID are set to 0 (except for the least significant byte) o IPv4-based addresses: in which the IID embeds the IPv4 address of the network interface (as in 2001:db8::192.0.2.1) o service port addresses: in which the IID embeds the TCP/UDP service port of the main service running on that node (as in 2001:db8::80 or 2001:db8::25) o wordy addresses: which encode words (as in 2001:db8::bad:cafe) Each of these patterns is discussed in detail in the following subsections.4.1.3.1. Low-Byte Addresses
The most common form of low-byte addresses is that in which all the bytes of the IID (except the least significant bytes) are set to zero (as in 2001:db8::1, 2001:db8::2, etc.). However, it is also common to find similar addresses in which the two lowest-order 16-bit words (from the right to left) are set to small numbers (as in 2001::db8::1:10, 2001:db8::2:10, etc.). Yet it is not uncommon to find IPv6 addresses in which the second lowest-order 16-bit word (from right to left) is set to a small value in the range 0x0000:0x00ff, while the lowest-order 16-bit word (from right to left) varies in the range 0x0000:0xffff. It should be noted that all of these address patterns are generally referred to as "low-byte
addresses", even when, strictly speaking, it is not only the lowest- order byte of the IPv6 address that varies from one address to another. In the worst-case scenario, the search space for this pattern is 2^24 (although most systems can be found by searching 2^16 or even 2^8 addresses).4.1.3.2. IPv4-Based Addresses
The most common form of these addresses is that in which an IPv4 address is encoded in the lowest-order 32 bits of the IPv6 address (usually as a result of the address notation of the form 2001:db8::192.0.2.1). However, it is also common for administrators to encode each of the bytes of the IPv4 address in each of the 16-bit words of the IID (as in, e.g., 2001:db8::192:0:2:1). Therefore, the search space for addresses following this pattern is that of the corresponding IPv4 prefix (or twice the size of that search space if both forms of "IPv4-based addresses" are to be searched).4.1.3.3. Service-Port Addresses
Addresses following this pattern include the service port (e.g., 80 for HTTP) in the lowest-order byte of the IID and have the rest of the bytes of the IID set to zero. There are a number of variants for this address pattern: o The lowest-order 16-bit word (from right to left) may contain the service port, and the second lowest-order 16-bit word (from right to left) may be set to a number in the range 0x0000-0x00ff (as in, e.g., 2001:db8::1:80). o The lowest-order 16-bit word (from right to left) may be set to a value in the range 0x0000-0x00ff, while the second lowest-order 16-bit word (from right to left) may contain the service port (as in, e.g., 2001:db8::80:1). o The service port itself might be encoded in decimal or in hexadecimal notation (e.g., an address embedding the HTTP port might be 2001:db8::80 or 2001:db8::50) -- with addresses encoding the service port as a decimal number being more common. Considering a maximum of 20 popular service ports, the search space for addresses following this pattern is, in the worst-case scenario, 10 * 2^11.
4.1.3.4. Wordy Addresses
Since the IPv6 address notation allows for a number of hexadecimal digits, it is not difficult to encode words into IPv6 addresses (as in, e.g., 2001:db8::bad:cafe). Addresses following this pattern are likely to be explored by means of "dictionary attacks"; therefore, computing the corresponding search space is not straightforward.4.1.4. IPv6 Addresses Corresponding to Transition/Coexistence Technologies
Some transition/coexistence technologies might be leveraged to reduce the target search space of remote address-scanning attacks, since they specify how the corresponding IPv6 address must be generated. For example, in the case of Teredo [RFC4380], the 64-bit IID is generated from the IPv4 address observed at a Teredo server along with a UDP port number. For obvious reasons, the search space for these addresses will depend on the specific transition/coexistence technology being employed.4.1.5. IPv6 Address Assignment in Real-World Network Scenarios
Figures 1, 2, and 3 provide a summary of the results obtained by [Gont-LACSEC2013] when measuring the address patterns employed by web servers, name servers, and mail servers, respectively. Figure 4 provides a rough summary of the results obtained by [Malone2008] for IPv6 routers. Figure 5 provides a summary of the results obtained by [Ford2013] for clients.
+---------------+------------+ | Address type | Percentage | +---------------+------------+ | IEEE-based | 1.44% | +---------------+------------+ | Embedded-IPv4 | 25.41% | +---------------+------------+ | Embedded-Port | 3.06% | +---------------+------------+ | ISATAP | 0.00% | +---------------+------------+ | Low-byte | 56.88% | +---------------+------------+ | Byte-pattern | 6.97% | +---------------+------------+ | Randomized | 6.24% | +---------------+------------+ Figure 1: Measured Web Server Addresses +---------------+------------+ | Address type | Percentage | +---------------+------------+ | IEEE-based | 0.67% | +---------------+------------+ | Embedded-IPv4 | 22.11% | +---------------+------------+ | Embedded-Port | 6.48% | +---------------+------------+ | ISATAP | 0.00% | +---------------+------------+ | Low-byte | 56.58% | +---------------+------------+ | Byte-pattern | 11.07% | +---------------+------------+ | Randomized | 3.09% | +---------------+------------+ Figure 2: Measured Name Server Addresses
+---------------+------------+ | Address type | Percentage | +---------------+------------+ | IEEE-based | 0.48% | +---------------+------------+ | Embedded-IPv4 | 4.02% | +---------------+------------+ | Embedded-Port | 1.07% | +---------------+------------+ | ISATAP | 0.00% | +---------------+------------+ | Low-byte | 92.65% | +---------------+------------+ | Byte-pattern | 1.20% | +---------------+------------+ | Randomized | 0.59% | +---------------+------------+ Figure 3: Measured Mail Server Addresses +--------------+------------+ | Address type | Percentage | +--------------+------------+ | Low-byte | 70.00% | +--------------+------------+ | IPv4-based | 5.00% | +--------------+------------+ | SLAAC | 1.00% | +--------------+------------+ | Wordy | <1.00% | +--------------+------------+ | Randomized | <1.00% | +--------------+------------+ | Teredo | <1.00% | +--------------+------------+ | Other | <1.00% | +--------------+------------+ Figure 4: Measured Router Addresses
+---------------+------------+ | Address type | Percentage | +---------------+------------+ | IEEE-based | 7.72% | +---------------+------------+ | Embedded-IPv4 | 14.31% | +---------------+------------+ | Embedded-Port | 0.21% | +---------------+------------+ | ISATAP | 1.06% | +---------------+------------+ | Randomized | 69.73% | +---------------+------------+ | Low-byte | 6.23% | +---------------+------------+ | Byte-pattern | 0.74% | +---------------+------------+ Figure 5: Measured Client Addresses NOTE: "ISATAP" stands for "Intra-Site Automatic Tunnel Addressing Protocol" [RFC5214]. It should be clear from these measurements that a very high percentage of host and router addresses follow very specific patterns. Figure 5 shows that while around 70% of clients observed in this measurement appear to be using temporary addresses, a significant number of clients still expose IEEE-based addresses and addresses using embedded IPv4 (thus also revealing IPv4 addresses). Besides, as noted in Section 4.1.1.3, temporary addresses are employed along with stable IPv6 addresses; thus, hosts employing a temporary address may still be the subject of address-scanning attacks that target their stable address(es). [ADDR-ANALYSIS] contains a spatial and temporal analysis of IPv6 addresses corresponding to clients and routers.4.2. IPv6 Address Scanning of Remote Networks
Although attackers have been able to get away with "brute-force" address-scanning attacks in IPv4 networks (thanks to the lesser search space), successfully performing a brute-force address-scanning attack of an entire /64 network would be infeasible. As a result, it is expected that attackers will leverage the IPv6 address patterns discussed in Section 4.1 to reduce the IPv6 address search space.
IPv6 address scanning of remote networks should consider an additional factor not present for the IPv4 case: since the typical IPv6 subnet is a /64, scanning an entire /64 could, in theory, lead to the creation of 2^64 entries in the Neighbor Cache of the last-hop router. Unfortunately, a number of IPv6 implementations have been found to be unable to properly handle a large number of entries in the Neighbor Cache; hence, these address-scanning attacks may have the side effect of resulting in a Denial-of-Service (DoS) attack [CPNI-IPv6] [RFC6583]. [RFC7421] discusses the "default" /64 boundary for host subnets and the assumptions surrounding it. While there are reports of sites implementing IPv6 subnets of size /112 or smaller to reduce concerns about the above attack, such smaller subnets are likely to make address-scanning attacks more feasible, in addition to encountering the issues with non-/64 host subnets discussed in [RFC7421].4.2.1. Reducing the Subnet ID Search Space
When address scanning a remote network, consideration is required to select which subnet IDs to choose. A typical site might have a /48 allocation, which would mean up to 65,000 or so IPv6 /64 subnets to be scanned. However, in the same way the search space for the IID can be reduced, we may also be able to reduce the subnet ID search space in a number of ways, by guessing likely address plan schemes or using any complementary clues that might exist from other sources or observations. For example, there are a number of documents available online (e.g., [RFC5375]) that provide recommendations for the allocation of address space, which address various operational considerations, including Regional Internet Registry (RIR) assignment policy, ability to delegate reverse DNS zones to different servers, ability to aggregate routes efficiently, address space preservation, ability to delegate address assignment within the organization, ability to add/allocate new sites/prefixes to existing entities without updating Access Control Lists (ACLs), and ability to de-aggregate and advertise subspaces via various Autonomous System (AS) interfaces. Address plans might include use of subnets that: o Run from low ID upwards, e.g., 2001:db8:0::/64, 2001:db8:1::/64, etc. o Use building numbers, in hexadecimal or decimal form. o Use Virtual Local Area Network (VLAN) numbers.
o Use an IPv4 subnet number in a dual-stack target, e.g., a site
with a /16 for IPv4 might use /24 subnets, and the IPv6 address
plan may reuse the third byte of the IPv4 address as the IPv6
subnet ID.
o Use the service "color", as defined for service-based prefix
coloring, or semantic prefixes. For example, a site using a
specific coloring for a specific service such as Voice over IP
(VoIP) may reduce the subnet ID search space for those devices.
The net effect is that the address space of an organization may be
highly structured, and allocations of individual elements within this
structure may be predictable once other elements are known.
In general, any subnet ID address plan may convey information, or be
based on known information, which may in turn be of advantage to an
attacker.
4.3. IPv6 Address Scanning of Local Networks
IPv6 address scanning in Local Area Networks (LANs) could be
considered, to some extent, a completely different problem than that
of scanning a remote IPv6 network. The main difference is that use
of link-local multicast addresses can relieve the attacker of
searching for unicast addresses in a large IPv6 address space.
NOTE:
While a number of other network reconnaissance vectors (such as
network snooping, leveraging Neighbor Discovery traffic, etc.) are
available when scanning a local network, this section focuses only
on address-scanning attacks (a la "ping sweep").
An attacker can simply send probe packets to the all-nodes link-local
multicast address (ff02::1), such that responses are elicited from
all local nodes.
Since Windows systems (Vista, 7, etc.) do not respond to ICMPv6 Echo
Request messages sent to multicast addresses, IPv6 address-scanning
tools typically employ a number of additional probe packets to elicit
responses from all the local nodes. For example, unrecognized IPv6
options of type 10xxxxxx elicit Internet Control Message Protocol
version 6 (ICMPv6) Parameter Problem, code 2, error messages.
Many address-scanning tools discover only IPv6 link-local addresses
(rather than, e.g., the global addresses of the target systems):
since the probe packets are typically sent with the attacker's IPv6
link-local address, the "victim" nodes send the response packets
using the IPv6 link-local address of the corresponding network
interface (as specified by the IPv6 address-selection rules [RFC6724]). However, sending multiple probe packets, with each packet employing source addresses from different prefixes, typically helps to overcome this limitation.4.4. Existing IPv6 Address-Scanning Tools
4.4.1. Remote IPv6 Network Address Scanners
IPv4 address-scanning tools have traditionally carried out their task by probing an entire address range (usually the entire address range comprised by the target subnetwork). One might argue that the reason for which they have been able to get away with such somewhat "rudimentary" techniques is that the scale or challenge of the task is so small in the IPv4 world that a "brute-force" attack is "good enough". However, the scale of the "address-scanning" task is so large in IPv6 that attackers must be very creative to be "good enough". Simply sweeping an entire /64 IPv6 subnet would just not be feasible. Many address-scanning tools do not even support sweeping an IPv6 address range. On the other hand, the alive6 tool from [THC-IPV6] supports sweeping address ranges, thus being able to leverage some patterns found in IPv6 addresses, such as the incremental addresses resulting from some DHCPv6 setups. Finally, the scan6 tool from [IPv6-Toolkit] supports sweeping address ranges and can also leverage all the address patterns described in Section 4.1 of this document. Clearly, a limitation of many of the currently available tools for IPv6 address scanning is that they lack an appropriately tuned "heuristics engine" that can help reduce the search space, such that the problem of IPv6 address scanning becomes tractable. It should be noted that IPv6 network monitoring and management tools also need to build and maintain information about the hosts in their network. Such systems can no longer scan internal systems in a reasonable time to build a database of connected systems. Rather, such systems will need more efficient approaches, e.g., by polling network devices for data held about observed IP addresses, MAC addresses, physical ports used, etc. Such an approach can also enhance address accountability, by mapping IPv4 and IPv6 addresses to observed MAC addresses. This of course implies that any access control mechanisms for querying such network devices, e.g., community strings for SNMP, should be set appropriately to avoid an attacker being able to gather address information remotely.
4.4.2. Local IPv6 Network Address Scanners
There are a variety of publicly available local IPv6 network address- scanners: o Current versions of nmap [nmap2015] implement this functionality. o The Hacker's Choice (THC) IPv6 Attack Toolkit [THC-IPV6] includes a tool (alive6) that implements this functionality. o SI6 Network's IPv6 Toolkit [IPv6-Toolkit] includes a tool (scan6) that implements this functionality.4.5. Mitigations
IPv6 address-scanning attacks can be mitigated in a number of ways. A non-exhaustive list of the possible mitigations includes: o Employing [RFC7217] (stable, semantically opaque IIDs) in replacement of addresses based on IEEE identifiers, such that any address patterns are eliminated. o Employing Intrusion Prevention Systems (IPSs) at the perimeter. o Enforcing IPv6 packet filtering where applicable (see, e.g., [RFC4890]). o Employing manually configured MAC addresses if virtual machines are employed and "resistance" to address-scanning attacks is deemed desirable, such that even if the virtual machines employ IEEE-derived IIDs, they are generated from non-predictable MAC addresses. o Avoiding use of sequential addresses when using DHCPv6. Ideally, the DHCPv6 server would allocate random addresses from a large pool (see, e.g., [IIDS-DHCPv6]). o Using the "default" /64 size IPv6 subnet prefixes. o In general, avoiding being predictable in the way addresses are assigned. It should be noted that some of the aforementioned mitigations are operational, while others depend on the availability of specific protocol features (such as [RFC7217]) on the corresponding nodes.
Additionally, while some resistance to address-scanning attacks is
generally desirable (particularly when lightweight mitigations are
available), there are scenarios in which mitigation of some address-
scanning vectors is unlikely to be a high priority (if at all
possible). And one should always remember that security by obscurity
is not a reasonable defense in itself; it may only be one (relatively
small) layer in a broader security environment.
Two of the techniques discussed in this document for local address-
scanning attacks are those that employ multicasted ICMPv6 Echo
Requests and multicasted IPv6 packets containing unsupported options
of type 10xxxxxx. These two vectors could be easily mitigated by
configuring nodes to not respond to multicasted ICMPv6 Echo Requests
(default on Windows systems) and by updating the IPv6 specifications
(and/or possibly configuring local nodes) such that multicasted
packets never elicit ICMPv6 error messages (even if they contain
unsupported options of type 10xxxxxx).
NOTE:
[SMURF-AMPLIFIER] proposed such an update to the IPv6
specifications.
In any case, when it comes to local networks, there are a variety of
network reconnaissance vectors. Therefore, even if address-scanning
vectors were mitigated, an attacker could still rely on, e.g.,
protocols employed for the so-called "service discovery protocols"
(see Section 5.2) or eventually rely on network snooping as a last
resort for network reconnaissance. There is ongoing work in the IETF
on extending mDNS, or at least DNS-based service discovery, to work
across a whole site, rather than in just a single subnet, which will
have associated security implications.
4.6. Conclusions
In the previous subsections, we have shown why a /64 host subnet may
be more vulnerable to address-based scanning than might intuitively
be thought and how an attacker might reduce the target search space
when performing an address-scanning attack.
We have described a number of mitigations against address-scanning
attacks, including the replacement of traditional SLAAC with stable
semantically opaque IIDs (which requires support from system
vendors). We have also offered some practical guidance in regard to
the principle of avoiding predictability in host addressing schemes.
Finally, examples of address-scanning approaches and tools are
discussed in the appendices.
While most early IPv6-enabled networks remain dual stack, they are more likely to be scanned and attacked over IPv4 transport, and one may argue that the IPv6-specific considerations discussed here are not of an immediate concern. However, an early IPv6 deployment within a dual-stack network may be seen by an attacker as a potentially "easier" target if the implementation of security policies is not as strict for IPv6 (for whatever reason). As IPv6-only networks become more common, the above considerations will be of much greater importance.
(page 23 continued on part 2)
