MIPSHOP E. Hepworth Internet-Draft Siemens Roke Manor Research Intended status: Informational T. Melia Expires: April 26, 2007 NEC S. Sreemanthula Nokia Research Center Y. Ohba Toshiba G. Vivek Intel J. Korhonen TeliaSonera R. Aguiar IT Sam(Zhongqi). Xia HUAWEI October 23, 2006 Mobility Independent Services: Problem Statement draft-melia-mipshop-mobility-services-ps-01 Status of this Memo By submitting this Internet-Draft, each author represents that any applicable patent or other IPR claims of which he or she is aware have been or will be disclosed, and any of which he or she becomes aware will be disclosed, in accordance with Section 6 of BCP 79. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt. The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html. This Internet-Draft will expire on April 26, 2007. Copyright Notice Hepworth, et al. Expires April 26, 2007 [Page 1] Internet-Draft Mobility Services October 2006 Copyright (C) The Internet Society (2006). Abstract There are on-going activities in the networking community to develop solutions that aid in IP handover mechanisms between heterogeneous wired and wireless access systems including, but not limited to, IEEE 802.21. Intelligent access selection, taking into account link layer attributes, requires the delivery of a variety of different information types to the terminal from different sources within the network and vice-versa. The protocol requirements for this signalling have both transport and security issues that must be considered. The signalling must not be constrained to specific link types, so there is at least a common component to the signalling problem which is within the scope of the IETF. This draft presents a problem statement for this core problem. Hepworth, et al. Expires April 26, 2007 [Page 2] Internet-Draft Mobility Services October 2006 Table of Contents 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4 2. Terminology . . . . . . . . . . . . . . . . . . . . . . . . . 4 3. Definition of Mobility Services . . . . . . . . . . . . . . . 5 4. Deployment Scenarios . . . . . . . . . . . . . . . . . . . . . 5 4.1. End-to-End Signalling and Transport over IP . . . . . . . 6 4.2. End-to-End Signalling and Partial Transport over IP . . . 6 4.3. End-to-End Signalling with a Proxy . . . . . . . . . . . . 7 4.4. End-to-End Network-to-Network Signalling . . . . . . . . . 7 5. Solution Components . . . . . . . . . . . . . . . . . . . . . 8 5.1. Payload Formats and Extensibility Considerations . . . . . 9 5.2. Requirements on the Mobility Service Transport Layer . . . 10 6. Security Considerations . . . . . . . . . . . . . . . . . . . 13 7. Conclusions and Open Issues . . . . . . . . . . . . . . . . . 14 8. References . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . . 15 Intellectual Property and Copyright Statements . . . . . . . . . . 18 Hepworth, et al. Expires April 26, 2007 [Page 3] Internet-Draft Mobility Services October 2006 1. Introduction This Internet Draft provides a problem statement for the exchange of information to support handover in heterogeneous link environments. This mobility support service allows more sophisticated handover operations by making available information about network characteristics, neighboring networks and associated characteristics, indications that a handover should take place, and suggestions for suitable target networks to which to handover. The mobility support services work complementarily with IP mobility mechanisms to enhance the overall performance and usability perception. There are two key attributes to the handover support service problem for inter-technology handovers: 1. The Information and Information Exchange mechanism: this includes the information elements that describe the information, and any signalling exchanges that are required to support the transfer of these Information Elements. 2. The Underlying Transport: this supports the above Information Exchange between devices in the network. The requirements on this transfer mechanism include transport issues, because of the volume of data to be sent, as well as discovery and security issues for this transport, as the signalling may cross administrative boundaries and is interdependent with AAA aspects. This draft has been motivated by on-going work within IEEE 802.21, but the following description intentionally describes the problem from a more general perspective. This document represents the views of the authors, and does not represent the official view of IEEE 802.21. The structure of this document is as follows. Section 3 defines mobility services. Section 4 provides a simple model for the protocol entities involved in the signalling and their possible relationships. Section 5 describes a decomposition of the signalling problem into service specific parts and a generic transport part. Section 5.2 describes more detailed requirements for the transport component. Section 6 provides security considerations, and Section 7 summarizes the conclusions and open issues. The appendixes describe some applications of the mobility services. 2. Terminology The following abbreviations are used in the document: Hepworth, et al. Expires April 26, 2007 [Page 4] Internet-Draft Mobility Services October 2006 o MN: mobile node o NN: network node, intended to represent some device in the network (the location of the node e.g. in the access network, home network is not specified, and for the moment it is assumed that they can reside anywhere). o EP: endpoint, intended to represent the terminating endpoints of the transport protocol used to support the signalling exchanges between nodes. o MME: A Mobility Management Entity implements network selection and handover decision algorithms and utilizes mobility signaling protocols and other protocols that aid in mobility functions. Generalizing, we call this functional entity Policy Decision Point which acts upon events and combines required actions with user profiles. The MME is able to collect information either from other network nodes or from the MN. 3. Definition of Mobility Services As mentioned in the introduction mobility (handover) support in heterogeneous wireless environments requires functional components located either in the mobile terminal or in the (access) network to exchange information and eventually to take decisions upon this information exchange. For instance traditional host-based handover solutions could be complemented with more sophisticated network- centric solutions reducing terminal complexity. Also, neighborhood discovery, potentially a complex operation in heterogeneous wireless scenarios, can result in a more simple step if implemented with an unified interface towards the access network. In this document the different supporting functions for handover management are generally referred as Mobility Services (MoS) having in common different requirements for the transport protocol. These requirements and associated functionalities are the focus of this document 4. Deployment Scenarios The deployment scenarios are outlined in the following sections. Note: while MN-to-MN signalling exchanges are theoretically possible, these are not currently being considered, and are out-of-scope. The following scenarios are discussed for understanding the overall problem of transporting MIH protocol and is not intended to show the Hepworth, et al. Expires April 26, 2007 [Page 5] Internet-Draft Mobility Services October 2006 scenarios are part of the requirements in the transport design. 4.1. End-to-End Signalling and Transport over IP In this case, the end-to-end signalling used to exchange the handover information elements (the Information Exchange) runs end-to-end between MN and NN. The underlying transport is also end-to-end +------+ +------+ | MN | | NN | | (EP) | | (EP) | +------+ +------+ Information Exchange <------------------------------------> /------------------------------------\ < Transport over IP > \------------------------------------/ Figure 1: End-to-end Signalling and Transport 4.2. End-to-End Signalling and Partial Transport over IP As before, the Information Exchange runs end-to-end between the MN and the second NN. However, in this scenario, some other transport means is used from the MN to the first NN, and the transport over IP is used only between NNs. This is analogous to the use of EAP end- to-end between Supplicant and Authentication Server, with an upper- layer multihop protocol such as RADIUS used as a backhaul transport protocol between an Access Point and the Authentication Server. +------+ +------+ +------+ | MN | | NN | | NN | | | | (EP) | | (EP) | +------+ +------+ +------+ Information Exchange <------------------------------------> (Transport over /------------------\ <--------------->< Transport over IP > e.g. L2) \------------------/ Figure 2: Partial Transport Hepworth, et al. Expires April 26, 2007 [Page 6] Internet-Draft Mobility Services October 2006 4.3. End-to-End Signalling with a Proxy In the final case, a number of proxies are inserted along the path between the two transport endpoints. The use of proxies is possible in both cases 1 and 2 above, but distinguished here as there are a number of options as to how the proxy may behave with regard to the transport and end-to-end signalling exchange. In this case, the proxy performs some processing on the Information Exchange before forwarding the information on. This can be viewed as concatenating signalling exchanges between a number of EPs. +------+ +---------+ +------+ | MN | | ProxyNN | | NN | | (EP) | | (EP) | | (EP) | +------+ +---------+ +------+ Information Exchange ------------------> -------------------> <------------------- <------------------ /---------------\ /----------------\ < Transport > < Transport > \---------------/ \----------------/ Figure 3: Information Exchange Approach The Proxy NN processes all layers of the protocol suite in the same way as an ordinary EP. There is a possibility for realizing other proxy scenarios. 4.4. End-to-End Network-to-Network Signalling In this case NN to NN signalling is envisioned. Such model should allow different network components to gather information from each other. This is useful for instance in conditions where network components need to take decisions and instruct mobile terminals of operation to be executed. Hepworth, et al. Expires April 26, 2007 [Page 7] Internet-Draft Mobility Services October 2006 +------+ +------+ | NN | | NN | | (EP) | | (EP) | +------+ +------+ Information Exchange -------------------> <------------------- /----------------\ < Transport > \----------------/ Figure 4: Information Exchange between different NN Network nodes exchange information about connected terminals status. 5. Solution Components Figure 5 shows a model where the Information Exchanges are implemented by a signalling protocol specific to a particular mobility service, and these are relayed over a generic transport layer (the Mobility Service Transport Layer). +----------------+ ^ |Mobility Support| | | Service 2 | | +----------------+ | (e.g. ES) | | Mobility Service |Mobility Support| +----------------+ | Signaling | Service 1 | +----------------+ | Layer | (e.g. IS) | |Mobility Support| | +----------------+ | Service 3 | | | (other) | | +----------------+ V ================================================ +---------------------------------------+ ^ Mobility Service | Mobility Service Transport Protocol | | Transport +---------------------------------------+ V Layer ================================================ +---------------------------------------+ | IP | +---------------------------------------+ Figure 5: Handover Services over IP The Mobility Service Transport Layer provides certain functionality (outlined in Section 5.2) to the higher layer mobility support Hepworth, et al. Expires April 26, 2007 [Page 8] Internet-Draft Mobility Services October 2006 services in order to support the exchange of information between communicating mobility service functions. The transport layer effectively provides a container capability to mobility support services, as well as any required transport and security operations required to provide communication without regard to the protocol semantics and data carried in the specific mobility services. The Mobility Support Services themselves may also define certain protocol exchanges to support the exchange of service specific Information Elements. It is likely that the responsibility for defining the contents and significance of the Information Elements is the responsibility of other standards bodies other than the IETF. Example mobility services include the Information Services, Event and Command services. 5.1. Payload Formats and Extensibility Considerations The format of the Mobility Service Transport Protocol is as follows: +----------------+----------------------------------------+ |Mobility Service| Opaque Payload | |Transport Header| (Mobility Support Service) | +----------------+----------------------------------------+ Figure 6: Protocol Structure The opaque payload encompasses the Mobility Support Service information that is to be transported. The definition of the Mobility Service Transport Header is something that is best addressed within the IETF. There are a number of issues with regard to the Mobility Support Service header and payload definition. These include: 1. Responsibility for defining the header: where should the contents of the Mobility Support Service header be defined, and should there be one or multiple header definitions (i.e. will a common header definition for all mobility support services be adequate?). Where there are commonalities, it may indicate that these aspects should actually be included in the Mobility Service Transport Header. 2. Payload Format: the format or the Mobility Support Service Data payload could be represented in a number of formats, e.g. TLV, ASN/1, XML or text. Ideally, a single payload representation should be defined, as support for multiple formats leads to unnecessary complexity. It is expected that a set of Data Objects will be defined for the Mobility Support Services to Hepworth, et al. Expires April 26, 2007 [Page 9] Internet-Draft Mobility Services October 2006 exchange. 3. Sharing of Data Objects: which refers to sharing the definitions of Data Objects between Mobility Support Services, e.g. if a Capabilities object is defined that is used by multiple Mobility Support Services, should the same definition be used by all of them. If this is the case, then a common identifier space is needed to identify the different Data Objects. There is a question about where the definition of Data Objects and the management of the identifier space should take place. The answers to some of the above issues may in part depend on how many standards groups are interested in defining their own Mobility Support Services. 5.2. Requirements on the Mobility Service Transport Layer The following section outlines some of the general transport requirements that should be supported by the Mobility Service Transport Protocol. Analysis has suggested that at least the following need to be taken into account: Discovery MNs need the ability to locate nodes that support particular mobility services in the network. There are no assumptions about the location of these mobility services within the network, therefore the discovery mechanism needs to operate across administrative boundaries. Issues such as speed of discovery, protection against spoofing, when discovery needs to take place, and the length of time over which the discovery information may remain valid all need to be considered. Approaches include: * Hard coding information into the MN, indicating either the IP address of the NN, or information about the NN that can be resolved onto an IP address. The configuration information could be managed dynamically, but assumes that the NN is independent of the access network to which the MN is currently attached. * Pushing information to the MN, where the information is delivered to the MN as part of other configuration operations, for example, in a Router Discovery exchange. The benefit of this approach is that no additional exchanges with the network would be required, but the limitations associated with modifying these protocols may limit applicability of the solution. Hepworth, et al. Expires April 26, 2007 [Page 10] Internet-Draft Mobility Services October 2006 * MN dynamically requesting information about a service, which may require both MN and NN support for a particular service discovery mechanism. This may require additional support by the access network (e.g. multicast or anycast) even when it may not be supporting the service directly itself. Numerous directory and configuration services already exist, and reuse of these mechanisms may be appropriate. There is an open question about whether multiple methods of discovery would be needed, and whether NNs would also need to discover other NNs. The definition of a service also needs to be determined, including the granularity of the description (for example, should the MN look for an "IS" service, or "IS-local information", and "IS-home network information" services. Information from a trusted source: The MN uses the Mobility Service information to make decisions about what steps to take next. It is essential that there is some way to ensure that the information received is from a trustworthy source. This includes cases where trusted proxies along the path have access to, and may modify, parts of the Mobility Service information. This requirement should reuse trust relationships that have already been established in the network, for example, on the relationships established by the AAA infrastructure after a mutual authentication, or on the certificate infrastructure required to support SEND. Low latency: Some of the Mobility Services generate time sensitive information. Therefore, there is a need to deliver the information over quite short timescales, and the required lifetime of a connection might be quite short lived. For reliable delivery, short-lived connections could be set up as and when needed, although there is a connection setup latency associated with this approach. Alternatively, a long-lived connection could be used, but this requires advanced warning of being needed and some way to maintain the state associated with the connection. It also assumes that the relationships between devices supporting the mobility service are fairly stable. Another alternative is connectionless operation, but this has interactions with other requirements such as reliable delivery. Reliability: Reliable delivery for some of the mobility services may be essential, but it is difficult to trade this off against the low latency requirement. It is also quite difficult to design a robust, high performance mechanism that can operate in heterogeneous environments, especially one where the link characteristics can vary quite dramatically. There are two main approaches that could be adopted: Hepworth, et al. Expires April 26, 2007 [Page 11] Internet-Draft Mobility Services October 2006 1. Assume the transport cannot be guaranteed to support reliable delivery. In this case, the Mobility Support Service itself will have to provide some sort of reliability mechanism to allow communicating endpoints to acknowledge receipt of information. 2. Assume the underlying transport will deal with most error situations, and provide a very basic acknowledgement mechanism that (if no acknowledgement is received) will indicate that something more serious has occurred than a packet drop (since these other types of error conditions are dealt with at the transport layer). Option 1 has a number of disadvantages associated with it, namely that ultimately the protocol design ends up re-inventing a lot of the functionality already available in lower layers at a higher layer where access to information about what is going on in the network is restricted. For example, how will the higher layer determine the cause of the error, if a message is lost due to network congestion, it is pointless sending the message again. It also adds to the complexity of the higher layer protocol, and makes successful deployment less certain (the protocol will have to be trialed in a number of network situations instead of re- using a protocol that has already been tested). Congestion Control: A Mobility Service may wish to transfer large amounts of data, placing a requirement for congestion control in the transport. There is an interaction between this requirement and that of the requirement for low latency since ways to deal with timely delivery of smaller asynchronous messages around the larger datagrams is required (mitigation of head of line blocking etc.). Secure delivery: The Mobility Service information must be delivered securely between trusted peers, where the transport may pass though untrusted intermediate nodes and networks. Design considerations include whether session based or host based security associations are required along the chain of NNs, and what the rate limitation requirements of requests/responses might be. Multiplexing: The transport service needs to be able to support different mobility services. This may require multiplexing and the ability to manage multiple discovery operations and peering relationships in parallel. Hepworth, et al. Expires April 26, 2007 [Page 12] Internet-Draft Mobility Services October 2006 Multihoming: For some information services exchanged with the MN, there is a possibility that the request and response messages can be carried over two different links e.g. a handover command request is on the current link while the response could be delivered on the new link. Depending on the IP mobility mechanism, there is some impact on the transport option for the mobility information services. This may potentially have some associated latency and security issues, for example, if the transport is over IP there is some transparency but Mobile IP may introduce additional delay and both TCP and UDP must use the permanent address of the MN. In addition to the above, it may be necessary for the transport to support multiple applications (or modes of operation) to support the particular requirements of the Information Exchange being carried out between nodes. This may require the ability to multiplex multiple information exchanges into a single transport exchange. Further information about transport requirements related to specific Mobility Services can be found in the different appendixes. 6. Security Considerations Network supported mobility services aim at improving decision making and management of dynamically connected hosts. The control and maintenance of mobile nodes becomes challenging where authentication and authorization credentials used to access a network are unavailable for the purpose of bootstrapping a security association for handover services. Information Services may not require authorization of the client, but both event and command services must authenticate message sources, particularly if they are mobile. Network side service entities will typically need to provide proof of authority to serve visiting devices. Where signalling or radio operations can result from received messages, significant disruption may result from processing bogus or modified messages. The effect of processing bogus messages depends largely upon the content of the message payload, which is handled by the handover services application. Regardless of the variation in effect, message delivery mechanisms need to provide protection against tampering, and spoofing. Sensitive and identifying information about a mobile device may be exchanged during handover service message exchange. Since handover decisions are to be made based upon message exchanges, it may be possible to trace an user's movement between cells, or predict future movements, by inspecting handover service messages. In order to Hepworth, et al. Expires April 26, 2007 [Page 13] Internet-Draft Mobility Services October 2006 prevent such tracking, message confidentiality should be available. This is particularly important since many mobile devices are associated with only one user, as divulgence of such information may violate the user's privacy. Additionally, identifying information may be exchanged during security association construction. As this information may be used to trace users across cell boundaries, identity protection should be available if possible, when establishing SAs. In addition, the user should not have to disclose its identity to the network (any more than it needed to during authentication) in order to access the Mobility Support Services. For example, if the local network is just aware that an anonymous user with a subscription to operatorXYX.com is accessing the network, the user should not have to divulge their true identity in order to access the Mobility Support Services available locally. Finally, the network nodes themselves will potentially be subject to denial of service attacks from MNs and these problems will be exacerbated if operation of the mobility service protocols imposes a heavy computational load on the NNs. The overall design has to consider at what stage (e.g. discovery, transport layer establishment, service specific protocol exchange) denial of service prevention or mitigation should be built in. 7. Conclusions and Open Issues This Internet draft outlined a broad problem statement for the signalling of information elements across a network to support media independent handover services. In order to enable this type of signalling service, a need for a generic transport solution with certain transport and security properties were outlined. Whilst the motivation for considering this problem has come from work within IEEE 802.21, a desirable goal is to ensure that solutions to this problem are applicable to a wider range of mobility services. It would be valuable to establish realistic performance goals for the solution to this common problem (i.e. transport and security aspects) using experience from previous IETF work in this area and knowledge about feasible deployment scenarios. This information could then be used as an input to other standards bodies in assisting them to design mobility services with feasible performance requirements. Much of the functionality required for this problem is available from existing IETF protocols or combination thereof. This document takes no position on whether an existing protocol can be adapted for the solution or whether new protocol development is required. In either Hepworth, et al. Expires April 26, 2007 [Page 14] Internet-Draft Mobility Services October 2006 case, we believe that the appropriate skills for development of protocols in this area lie in the IETF. 8. References [1] "Draft IEEE Standard for Local and Metropolitan Area Networks: Media Independent Handover Services", IEEE LAN/MAN Draft IEEE P802.21/D01.00, March 2006. [2] Adoba, B., "Architectural Implications of Link Indications draft-iab-link-indications-03.txt", June 2005. [3] Perkins, C., "IP Mobility Support for IPv4", RFC 3344, August 2002. [4] Johnson, D., Perkins, C., and J. Arkko, "Mobility Support in IPv6", RFC 3775, June 2004. [5] Moskowitz, R. and P. Nikander, "Host Identity Protocol (HIP) Architecture", RFC 4423, May 2006. [6] Eronen, P., "IKEv2 Mobility and Multihoming Protocol (MOBIKE)", RFC 4555, June 2006. [7] 3GPP, "3GPP system architecture evolution (SAE): Report on technical options and conclusions", 3GPP TR 23.882 0.10.1, February 2006. [8] Koodli, R., "Fast Handovers for Mobile IPv6", RFC 4068, July 2005. Authors' Addresses Eleanor Hepworth Siemens Roke Manor Research Roke Manor Romsey, SO51 5RE UK Email: eleanor.hepworth@roke.co.uk Hepworth, et al. Expires April 26, 2007 [Page 15] Internet-Draft Mobility Services October 2006 Telemaco Melia NEC Europe Network Laboratories Kufuerstenanlage 36 Heidelberg 69115 Germany Phone: +49 6221 90511 42 Email: telemaco.melia@netlab.nec.de Srivinas Sreemanthula Nokia Research Center 6000 Connection Dr. Irving, TX 75028 USA Email: srinivas.sreemanthula@nokia.com Yoshihiro Ohba Toshiba America Research, Inc. 1 Telcordia Drive Piscateway NJ 08854 USA Email: yohba@tari.toshiba.com Vivek Gupta Intel Corporation 2111 NE 25th Avenue Hillsboro, OR 97124 USA Phone: +1 503 712 1754 Email: vivek.g.gupta@intel.com Jouni Korhonen TeliaSonera Corporation. P.O.Box 970 FIN-00051 Sonera FINLAND Phone: +358 40 534 4455 Email: jouni.korhonen@teliasonera.com Hepworth, et al. Expires April 26, 2007 [Page 16] Internet-Draft Mobility Services October 2006 Rui L.A. Aguiar Instituto de Telecomunicacoes Universidade de Aveiro Aveiro 3810 Portugal Phone: +351 234 377900 Email: ruilaa@det.ua.pt Sam(Zhongqi) Xia Huawei Technologies Co.,Ltd HuaWei Bld., No.3 Xinxi Rd. Shang-Di Information Industry Base,Hai-Dian District Beijing 100085 P.R. China Phone: +86-10-82836136 Email: xiazhongqi@huawei.com Hepworth, et al. Expires April 26, 2007 [Page 17] Internet-Draft Mobility Services October 2006 Full Copyright Statement Copyright (C) The Internet Society (2006). 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 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. Acknowledgment Funding for the RFC Editor function is provided by the IETF Administrative Support Activity (IASA). Hepworth, et al. Expires April 26, 2007 [Page 18]