Display RFC
RFC 4843 - An IPv6 Prefix for Overlay Routable Cryptographic Hash Identifiers (ORCHID)
(Formats: TXT)
(Obsoleted By: RFC 7343)
Network Working Group P. Nikander Request for Comments: 4843 Ericsson Research Nomadic Lab Category: Experimental J. Laganier DoCoMo Euro-Labs F. Dupont CELAR April 2007An IPv6 Prefix for Overlay Routable Cryptographic Hash Identifiers (ORCHID) Status of This Memo This memo defines an Experimental Protocol for the Internet community. It does not specify an Internet standard of any kind. Discussion and suggestions for improvement are requested. Distribution of this memo is unlimited. Copyright Notice Copyright (C) The IETF Trust (2007). Abstract This document introduces Overlay Routable Cryptographic Hash Identifiers (ORCHID) as a new, experimental class of IPv6-address- like identifiers. These identifiers are intended to be used as endpoint identifiers at applications and Application Programming Interfaces (API) and not as identifiers for network location at the IP layer, i.e., locators. They are designed to appear as application layer entities and at the existing IPv6 APIs, but they should not appear in actual IPv6 headers. To make them more like vanilla IPv6 addresses, they are expected to be routable at an overlay level. Consequently, while they are considered non-routable addresses from the IPv6 layer point-of-view, all existing IPv6 applications are expected to be able to use them in a manner compatible with current IPv6 addresses. This document requests IANA to allocate a temporary prefix out of the IPv6 addressing space for Overlay Routable Cryptographic Hash Identifiers. By default, the prefix will be returned to IANA in 2014, with continued use requiring IETF consensus. Nikander, et al. Experimental [Page 1] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1. Rationale and Intent . . . . . . . . . . . . . . . . . . . 3 1.2. ORCHID Properties . . . . . . . . . . . . . . . . . . . . 4 1.3. Expected use of ORCHIDs . . . . . . . . . . . . . . . . . 4 1.4. Action Plan . . . . . . . . . . . . . . . . . . . . . . . 4 1.5. Terminology . . . . . . . . . . . . . . . . . . . . . . . 4 2. Cryptographic Hash Identifier Construction . . . . . . . . . . 5 3. Routing Considerations . . . . . . . . . . . . . . . . . . . . 6 3.1. Overlay Routing . . . . . . . . . . . . . . . . . . . . . 6 4. Collision Considerations . . . . . . . . . . . . . . . . . . . 7 5. Design Choices . . . . . . . . . . . . . . . . . . . . . . . . 9 6. Security Considerations . . . . . . . . . . . . . . . . . . . 9 7. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 10 8. Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . 11 9. References . . . . . . . . . . . . . . . . . . . . . . . . . . 11 9.1. Normative References . . . . . . . . . . . . . . . . . . . 11 9.2. Informative References . . . . . . . . . . . . . . . . . . 11 1. Introduction This document introduces Overlay Routable Cryptographic Hash Identifiers (ORCHID), a new class of IP address-like identifiers. These identifiers are intended to be globally unique in a statistical sense (see Section 4), non-routable at the IP layer, and routable at some overlay layer. The identifiers are securely bound, via a secure hash function, to the concatenation of an input bitstring and a context tag. Typically, but not necessarily, the input bitstring will include a suitably encoded public cryptographic key. 1.1. Rationale and Intent These identifiers are expected to be used at the existing IPv6 Application Programming Interfaces (API) and application protocols between consenting hosts. They may be defined and used in different contexts, suitable for different overlay protocols. Examples of these include Host Identity Tags (HIT) in the Host Identity Protocol (HIP) [HIP-BASE] and Temporary Mobile Identifiers (TMI) for Mobile IPv6 Privacy Extension [PRIVACYTEXT]. As these identifiers are expected to be used along with IPv6 addresses at both applications and APIs, co-ordination is desired to make sure that an ORCHID is not inappropriately taken for a vanilla IPv6 address and vice versa. In practice, allocation of a separate prefix for ORCHIDs seems to suffice, making them compatible with IPv6 addresses at the upper layers while simultaneously making it trivial to prevent their usage at the IP layer. Nikander, et al. Experimental [Page 2] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 While being technically possible to use ORCHIDs between consenting hosts without any co-ordination with the IETF and the IANA, the authors would consider such practice potentially dangerous. A specific danger would be realised if the IETF community later decided to use the ORCHID prefix for some different purpose. In that case, hosts using the ORCHID prefix would be, for practical purposes, unable to use the prefix for the other new purpose. That would lead to partial balkanisation of the Internet, similar to what has happened as a result of historical hijackings of non-RFC 1918 [RFC1918] IPv4 addresses for private use. The whole need for the proposed allocation grows from the desire to be able to use ORCHIDs with existing applications and APIs. This desire leads to the potential conflict, mentioned above. Resolving the conflict requires the proposed allocation. One can argue that the desire to use these kinds of identifiers via existing APIs is architecturally wrong, and there is some truth in that argument. Indeed, it would be more desirable to introduce a new API and update all applications to use identifiers, rather than locators, via that new API. That is exactly what we expect to happen in the long run. However, given the current state of the Internet, we do not consider it viable to introduce any changes that, at once, require applications to be rewritten and host stacks to be updated. Rather than that, we believe in piece-wise architectural changes that require only one of the existing assets to be touched. ORCHIDs are designed to address this situation: to allow people to experiment with protocol stack extensions, such as secure overlay routing, HIP, or Mobile IP privacy extensions, without requiring them to update their applications. The goal is to facilitate large-scale experiments with minimum user effort. For example, there already exists, at the time of this writing, HIP implementations that run fully in user space, using the operating system to divert a certain part of the IPv6 address space to a user level daemon for HIP processing. In practical terms, these implementations are already using a certain IPv6 prefix for differentiating HIP identifiers from IPv6 addresses, allowing them both to be used by the existing applications via the existing APIs. This document argues for allocating an experimental prefix for such purposes, thereby paving the way for large-scale experiments with cryptographic identifiers without the dangers caused by address-space hijacking. Nikander, et al. Experimental [Page 3] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 1.2. ORCHID Properties ORCHIDs are designed to have the following properties: o Statistical uniqueness; also see Section 4 o Secure binding to the input parameters used in their generation (i.e., the context identifier and a bitstring). o Aggregation under a single IPv6 prefix. Note that this is only needed due to the co-ordination need as indicated above. Without such co-ordination need, the ORCHID namespace could potentially be completely flat. o Non-routability at the IP layer, by design. o Routability at some overlay layer, making them, from an application point of view, semantically similar to IPv6 addresses. As mentioned above, ORCHIDs are intended to be generated and used in different contexts, as suitable for different mechanisms and protocols. The context identifier is meant to be used to differentiate between the different contexts; see Section 4 for a discussion of the related API and kernel level implementation issues, and Section 5 for the design choices explaining why the context identifiers are used. 1.3. Expected use of ORCHIDs Examples of identifiers and protocols that are expected to adopt the ORCHID format include Host Identity Tags (HIT) in the Host Identity Protocol [HIP-BASE] and the Temporary Mobile Identifiers (TMI) in the Simple Privacy Extension for Mobile IPv6 [PRIVACYTEXT]. The format is designed to be extensible to allow other experimental proposals to share the same namespace. 1.4. Action Plan This document requests IANA to allocate an experimental prefix out of the IPv6 addressing space for Overlay Routable Cryptographic Hash Identifiers. 1.5. Terminology The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in [RFC2119]. Nikander, et al. Experimental [Page 4] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 2. Cryptographic Hash Identifier Construction An ORCHID is generated using the algorithm below. The algorithm takes a bitstring and a context identifier as input and produces an ORCHID as output. Input := any bitstring Hash Input := Context ID | Input Hash := Hash_function( Hash Input ) ORCHID := Prefix | Encode_100( Hash ) where: | : Denotes concatenation of bitstrings Input : A bitstring that is unique or statistically unique within a given context. The bitstring is intended to be associated with the to-be-created ORCHID in the given context. Context ID : A randomly generated value defining the expected usage context for the particular ORCHID and the hash function to be used for generation of ORCHIDs in this context. These values are allocated out of the namespace introduced for CGA Type Tags; see RFC 3972 and http://www.iana.org/assignments/cga-message-types. Hash_function : The one-way hash function (i.e., hash function with pre-image resistance and second pre-image resistance) to be used according to the document defining the context usage identified by the Context ID. For example, the current version of the HIP specification defines SHA1 [RFC3174] as the hash function to be used to generate ORCHIDs used in the HIP protocol [HIP-BASE]. Encode_100( ) : An extraction function in which output is obtained by extracting the middle 100-bit-long bitstring from the argument bitstring. Prefix : A constant 28-bit-long bitstring value (2001:10::/28). To form an ORCHID, two pieces of input data are needed. The first piece can be any bitstring, but is typically expected to contain a public cryptographic key and some other data. The second piece is a Nikander, et al. Experimental [Page 5] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 context identifier, which is a 128-bit-long datum, allocated as specified in Section 7. Each specific experiment (such as HIP HITs or MIP6 TMIs) is expected to allocate their own, specific context identifier. The input bitstring and context identifier are concatenated to form an input datum, which is then fed to the cryptographic hash function to be used according to the document defining the context usage identified by the Context ID. The result of the hash function is processed by an encoding function, resulting in a 100-bit-long value. This value is prepended with the 28-bit ORCHID prefix. The result is the ORCHID, a 128-bit-long bitstring that can be used at the IPv6 APIs in hosts participating to the particular experiment. The ORCHID prefix is allocated under the IPv6 global unicast address block. Hence, ORCHIDs are indistinguishable from IPv6 global unicast addresses. However, it should be noted that ORCHIDs do not conform with the IPv6 global unicast address format defined in Section 2.5.4 of [RFC4291] since they do not have a 64-bit Interface ID formatted as described in Section 2.5.1. of [RFC4291]. 3. Routing Considerations ORCHIDs are designed to serve as location independent endpoint- identifiers rather than IP-layer locators. Therefore, routers MAY be configured not to forward any packets containing an ORCHID as a source or a destination address. If the destination address is an ORCHID but the source address is a valid unicast source address, routers MAY be configured to generate an ICMP Destination Unreachable, Administratively Prohibited message. Due to the experimental nature of ORCHIDs, router software MUST NOT include any special handling code for ORCHIDs. In other words, the non-routability property of ORCHIDs, if implemented, MUST be implemented via configuration and NOT by hardwired software code. At this time, it is RECOMMENDED that the default router configuration not handle ORCHIDs in any special way. In other words, there is no need to touch existing or new routers due to this experiment. If such a reason should later appear, for example, due to a faulty implementation leaking ORCHIDs to the IP layer, the prefix can be and should be blocked by a simple configuration rule. 3.1. Overlay Routing As mentioned multiple times, ORCHIDs are designed to be non-routable at the IP layer. However, there are multiple ongoing research efforts for creating various overlay routing and resolution mechanisms for flat identifiers. For example, the Host Identity Nikander, et al. Experimental [Page 6] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 Indirection Infrastructure (Hi3) [Hi3] and Node Identity Internetworking Architecture (NodeID) [NodeID] proposals, outline ways for using a Distributed Hash Table to forward HIP packets based on the Host Identity Tag. What is common to the various research proposals is that they create a new kind of resolution or routing infrastructure on top of the existing Internet routing structure. In practical terms, they allow delivery of packets based on flat, non-routable identifiers, utilising information stored in a distributed database. Usually, the database used is based on Distributed Hash Tables. This effectively creates a new routing network on top of the existing IP-based routing network, capable of routing packets that are not addressed by IP addresses but some other kind of identifiers. Typical benefits from overlay routing include location independence, more scalable multicast, anycast, and multihoming support than in IP, and better DoS resistance than in the vanilla Internet. The main drawback is typically an order of magnitude of slower performance, caused by an easily largish number of extra look-up or forwarding steps needed. Consequently, in most practical cases, the overlay routing system is used only during initial protocol state set-up (cf. TCP handshake), after which the communicating endpoints exchange packets directly with IP, bypassing the overlay network. The net result of the typical overlay routing approaches is a communication service whose basic functionality is comparable to that provided by classical IP but provides considerably better resilience that vanilla IP in dynamic networking environments. Some experiments also introduce additional functionality, such as enhanced security or ability to effectively route through several IP addressing domains. The authors expect ORCHIDs to become fully routable, via one or more overlay systems, before the end of the experiment. 4. Collision Considerations As noted above, the aim is that ORCHIDs are globally unique in a statistical sense. That is, given the ORCHID referring to a given entity, the probability of the same ORCHID being used to refer to another entity elsewhere in the Internet must be sufficiently low so that it can be ignored for most practical purposes. We believe that the presented design meets this goal; see Section 5. Consider next the very rare case that some ORCHID happens to refer to two different entities at the same time, at two different locations in the Internet. Even in this case, the probability of this fact becoming visible (and therefore a matter of consideration) at any Nikander, et al. Experimental [Page 7] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 single location in the Internet is negligible. For the vast majority of cases, the two simultaneous uses of the ORCHID will never cross each other. However, while rare, such collisions are still possible. This section gives reasonable guidelines on how to mitigate the consequences in the case that such a collision happens. As mentioned above, ORCHIDs are expected to be used at the legacy IPv6 APIs between consenting hosts. The context ID is intended to differentiate between the various experiments, or contexts, sharing the ORCHID namespace. However, the context ID is not present in the ORCHID itself, but only in front of the input bitstring as an input to the hash function. While this may lead to certain implementation- related complications, we believe that the trade-off of allowing the hash result part of an ORCHID being longer more than pays off the cost. Because ORCHIDs are not routable at the IP layer, in order to send packets using ORCHIDs at the API level, the sending host must have additional overlay state within the stack to determine which parameters (e.g., what locators) to use in the outgoing packet. An underlying assumption here, and a matter of fact in the proposals that the authors are aware of, is that there is an overlay protocol for setting up and maintaining this additional state. It is assumed that the state-set-up protocol carries the input bitstring, and that the resulting ORCHID-related state in the stack can be associated back with the appropriate context and state-set-up protocol. Even though ORCHID collisions are expected to be extremely rare, two kinds of collisions may still happen. First, it is possible that two different input bitstrings within the same context may map to the same ORCHID. In this case, the state-set-up mechanism is expected to resolve the conflict, for example, by indicating to the peer that the ORCHID in question is already in use. A second type of collision may happen if two input bitstrings, used in different usage contexts, map to the same ORCHID. In this case, the main confusion is about which context to use. In order to prevent these types of collisions, it is RECOMMENDED that implementations that simultaneously support multiple different contexts maintain a node-wide unified database of known ORCHIDs, and indicate a conflict if any of the mechanisms attempt to register an ORCHID that is already in use. For example, if a given ORCHID is already being used as a HIT in HIP, it cannot simultaneously be used as a TMI in Mobile IP. Instead, if Mobile IP attempts to use the ORCHID, it will be notified (by the kernel) that the ORCHID in question is already in use. Nikander, et al. Experimental [Page 8] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 5. Design Choices The design of this namespace faces two competing forces: o As many bits as possible should be preserved for the hash result. o It should be possible to share the namespace between multiple mechanisms. The desire to have a long hash result requires that the prefix be as short as possible, and use few (if any) bits for additional encoding. The present design takes this desire to the maxim: all the bits beyond the prefix are used as hash output. This leaves no bits in the ORCHID itself available for identifying the context. Additionally, due to security considerations, the present design REQUIRES that the hash function used in constructing ORCHIDs be constant; see Section 6. The authors explicitly considered including a hash-extension mechanism, similar to the one in CGA [RFC3972], but decided to leave it out. There were two reasons: desire for simplicity, and the somewhat unclear IPR situation around the hash-extension mechanism. If there is a future revision of this document, we strongly advise the future authors to reconsider the decision. The desire to allow multiple mechanisms to share the namespace has been resolved by including the context identifier in the hash- function input. While this does not allow the mechanism to be directly inferred from a ORCHID, it allows one to verify that a given input bitstring and ORCHID belong to a given context, with high- probability; but also see Section 6. 6. Security Considerations ORCHIDs are designed to be securely bound to the Context ID and the bitstring used as the input parameters during their generation. To provide this property, the ORCHID generation algorithm relies on the second-preimage resistance (a.k.a. one-way) property of the hash function used in the generation [RFC4270]. To have this property and to avoid collisions, it is important that the allocated prefix is as short as possible, leaving as many bits as possible for the hash output. For a given Context ID, all mechanisms using ORCHIDs MUST use exactly the same mechanism for generating an ORCHID from the input bitstring. Allowing different mechanisms, without explicitly encoding the mechanism in the Context ID or the ORCHID itself, would allow so- called bidding-down attacks. That is, if multiple different hash Nikander, et al. Experimental [Page 9] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 functions were allowed to construct ORCHIDs valid for the same Context ID, and if one of the hash functions became insecure, that would allow attacks against even those ORCHIDs valid for the same Context ID that had been constructed using the other, still secure hash functions. Due to the desire to keep the hash output value as long as possible, the hash function is not encoded in the ORCHID itself, but rather in the Context ID. Therefore, the present design allows only one method per given Context ID for constructing ORCHIDs from input bitstrings. If other methods (perhaps using more secure hash functions) are later needed, they MUST use a different Context ID. Consequently, the suggested method to react to the hash result becoming too short, due to increased computational power, or to the used hash function becoming insecure due to advances in cryptology, is to allocate a new Context ID and cease to use the present one. As of today, SHA1 [RFC3174] is considered as satisfying the second- preimage resistance requirement. The current version of the HIP specification defines SHA1 [RFC3174] as the hash function to be used to generate ORCHIDs for the Context ID used by the HIP protocol [HIP-BASE]. In order to preserve a low enough probability of collisions (see Section 4), each method MUST utilize a mechanism that makes sure that the distinct input bitstrings are either unique or statistically unique within that context. There are several possible methods to ensure this; for example, one can include into the input bitstring a globally maintained counter value, a pseudo-random number of sufficient entropy (minimum 100 bits), or a randomly generated public cryptographic key. The Context ID makes sure that input bitstrings from different contexts never overlap. These together make sure that the probability of collisions is determined only by the probability of natural collisions in the hash space and is not increased by a possibility of colliding input bitstrings. 7. IANA Considerations IANA allocated a temporary non-routable 28-bit prefix from the IPv6 address space. By default, the prefix will be returned to IANA in 2014, continued use requiring IETF consensus. As per [RFC4773], the 28-bit prefix was drawn out of the IANA Special Purpose Address Block, namely 2001:0000::/23, in support of the experimental usage described in this document. IANA has updated the IPv6 Special Purpose Address Registry. Nikander, et al. Experimental [Page 10] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 During the discussions related to this document, it was suggested that other identifier spaces may be allocated from this block later. However, this document does not define such a policy or allocations. The Context Identifier (or Context ID) is a randomly generated value defining the usage context of an ORCHID and the hash function to be used for generation of ORCHIDs in this context. This document defines no specific value. We propose sharing the name space introduced for CGA Type Tags. Hence, defining new values would follow the rules of Section 8 of [RFC3972], i.e., on a First Come First Served basis. 8. Acknowledgments Special thanks to Geoff Huston for his sharp but constructive critique during the development of this memo. Tom Henderson helped to clarify a number of issues. This document has also been improved by reviews, comments, and discussions originating from the IPv6, Internet Area, and IETF communities. Julien Laganier is partly funded by Ambient Networks, a research project supported by the European Commission under its Sixth Framework Program. The views and conclusions contained herein are those of the authors and should not be interpreted as necessarily representing the official policies or endorsements, either expressed or implied, of the Ambient Networks project or the European Commission. 9. References 9.1. Normative References [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. [RFC3972] Aura, T., "Cryptographically Generated Addresses (CGA)", RFC 3972, March 2005. 9.2. Informative References [HIP-BASE] Moskowitz, R., "Host Identity Protocol", Work in Progress, February 2007. [Hi3] Nikander, P., Arkko, J., and B. Ohlman, "Host Identity Indirection Infrastructure (Hi3)", November 2004. Nikander, et al. Experimental [Page 11] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 [NodeID] Ahlgren, B., Arkko, J., Eggert, L., and J. Rajahalme, "A Node Identity Internetworking Architecture (NodeID)", April 2006. [PRIVACYTEXT] Dupont, F., "A Simple Privacy Extension for Mobile IPv6", Work in Progress, July 2006. [RFC1918] Rekhter, Y., Moskowitz, R., Karrenberg, D., Groot, G., and E. Lear, "Address Allocation for Private Internets", BCP 5, RFC 1918, February 1996. [RFC3174] Eastlake, D. and P. Jones, "US Secure Hash Algorithm 1 (SHA1)", RFC 3174, September 2001. [RFC4270] Hoffman, P. and B. Schneier, "Attacks on Cryptographic Hashes in Internet Protocols", RFC 4270, November 2005. [RFC4291] Hinden, R. and S. Deering, "IP Version 6 Addressing Architecture", RFC 4291, February 2006. [RFC4773] Huston, G., "Administration of the IANA Special Purpose IPv6 Address Block", RFC 4773, December 2006. Nikander, et al. Experimental [Page 12] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 Authors' Addresses Pekka Nikander Ericsson Research Nomadic Lab JORVAS FI-02420 Finland Phone: +358 9 299 1 EMail: pekka.nikander@nomadiclab.com Julien Laganier DoCoMo Communications Laboratories Europe GmbH Landsberger Strasse 312 Munich 80687 Germany Phone: +49 89 56824 231 EMail: julien.ietf@laposte.net Francis Dupont CELAR EMail: Francis.Dupont@fdupont.fr Nikander, et al. Experimental [Page 13] RFC 4843 Cryptographic Hash IDentifiers (ORCHID) April 2007 Full Copyright Statement Copyright (C) The IETF Trust (2007). This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights. This document and the information contained herein are provided on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Intellectual Property The IETF takes no position regarding the validity or scope of any Intellectual Property Rights or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; nor does it represent that it has made any independent effort to identify any such rights. Information on the procedures with respect to rights in RFC documents can be found in BCP 78 and BCP 79. Copies of IPR disclosures made to the IETF Secretariat and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use of such proprietary rights by implementers or users of this specification can be obtained from the IETF on-line IPR repository at http://www.ietf.org/ipr. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights that may cover technology that may be required to implement this standard. Please address the information to the IETF at ietf-ipr@ietf.org. Acknowledgement Funding for the RFC Editor function is currently provided by the Internet Society. Nikander, et al. Experimental [Page 14]