Tsunemasa Hayashi, NTT
   Internet Draft                          Haixiang He, Nortel Networks
   Expires: April 15, September 6, 2006                        Hiroaki Satou, NTT
                                                      Hiroshi Ohta, NTT
                                         Susheela Vaidya, Cisco Systems

                                                       October 12, 2005

                                                          March 5, 2006

    Issues Related to Receiver Access Control in the Current Multicast
                                 Protocols
                   <draft-ietf-mboned-rac-issues-01.txt>
                   <draft-ietf-mboned-rac-issues-02.txt>

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.

   Internet-Drafts are draft documents valid for a maximum of six
   months and may be updated, replaced, or obsolet  ed 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 15, September 6,  2006.

Copyright Notice

   Copyright (C) The Internet Society (2005) (2006)

Abstract

   This I-D memo evaluates the extent to which current multicasting
   protocols can be used to address common requirements for commercial, large-
   scale
   large-scale IP multicasting.  Four existing possible multicasting
   architectures (with or without some form of access or content
   control) are presented. Then each architecture is analyzed with
   respect to how it can or cannot satisfactorily address each issue.
   This I-D memo concludes that for many of these issues the possible
   architectures based on present standards as they now exist require
   non-standardized solutions to meet common use requirements. This I-D
   memo recommends for requirements to be defined that would set the
   groundwork for creating standardized ways to overcome framework(s) and solutions that sufficiently address
   these limitations.

   Copyright Notice...................................................1
   1. Introduction....................................................3
   2. Definitions and Abbreviations...................................4
   2.1 Definitions....................................................4
   2.2 Abbreviations..................................................4 Abbreviations..................................................5
   3. Common use models and network architecture implications.........5
   4. Issues in multicasting related to commercial and large-scale
   implementations....................................................6
   implementations....................................................7
   4.1 Access limits and resource issues..............................6 issues..............................7
   4.2 Capability to distinguish between receivers (end hosts)........6 hosts)........7
   4.3 Capability to distinguish between users (as opposed to merely
   hosts).............................................................7
   4.4 Minimizing Channel "leave latency"........................................7 Join Latency and Leave Latency..............8
   4.5 Surveillance of receiver by sender.............................7 sender.............................8
   4.5.1 Precise access log...........................................7 logging.......................................8
   4.5.2 How to share user information................................7 information................................8
   4.5.3 Trustworthy logs to monitor user activity....................8
   4.6 Notification to users of the result of the join request........8 request........9
   4.7 Sharing of Infrastructure for Support of Triple Play....................................................8 Play Services..9
   4.8 DRM Protection.................................................8 Protection.................................................9
   5. Description of existing architectures...........................8 architectures...........................9
   5.1 IGMP/MLD.......................................................8 IGMP/MLD.......................................................9
   5.2 IGMP/MLD plus L2/L3 Authentication with Access Control Policy.10 Policy.11
   5.3 Unicast Control with IGMP/MLD.................................11 IGMP/MLD.................................12
   5.4 IGMP/MLD with Multicast Encryption............................11 Encryption............................13
   6. Evaluation of architectures by issue...........................12 issue...........................14
   6.1 Access limit capabilities, compared by architecture...........12 architecture...........14
   6.2 Capability to distinguish between receivers, compared by
   architecture......................................................13
   architecture......................................................15
   6.3 Capability to distinguish between users, compared by architecture
   ..................................................................13
   architecture......................................................16
   6.4 Maintain guaranteed quality-level of data delivery (Voice,
   Video), compared by architecture..........................................14 architecture..................................17
   6.5 Fast leave for fast surfing capability, compared by architecture
   ..................................................................14
   ..................................................................17
   6.6 Surveillance of receiver by sender, compared by architecture..15 architecture..18
   6.7.Notification to users of the result of the join request compared
   by architecture...................................................15 architecture...................................................19
   6.8 Comparison summary............................................16 summary............................................19
   7. IANA considerations............................................16 considerations............................................20
   8. Security considerations........................................16 considerations........................................20
   9. Conclusion.....................................................16 Conclusion.....................................................20
   Normative References..............................................17 References..............................................20
   Comments..........................................................22
   Full Copyright Statement..........................................19 Statement..........................................23
   Intellectual Property.............................................19
   Acknowledgement...................................................19 Property.............................................23
   Expiration........................................................24
   Acknowledgement...................................................24

1. Introduction

   The intention of this I-D memo is to initiate a discussion on the state
   of current multicasting protocols deployed for commercial, large-scale large-
   scale multicasting and their capabilities to provide receiver access
   control.

   Existing IP multicasting protocols (as presented in Section 5) were
   designed to meet certain sets of requirements that do not
   necessarily include architectural considerations intended to support
   commercial services. This I-D memo presents a number of issues network
   providers may face when they attempt to apply current multicasting
   standards to commercial services.   The extent to which existing
   multicast protocols can or cannot satisfactorily deal with these
   issues is explored.  A few network models based on a range of
   different business models are presented as a basis for defining
   requirements.

   Multicasting can be useful to make the network more scalable when a
   large volume of information needs to be distributed to a large
   number of receivers.  However, multicasting according to current
   standards (e.g., IGMPv3[1] and MLDv2[2]) has drawbacks compared to
   unicasting in terms of its commercial applicability because of the
   insufficiency of access control and protect protection of network resources
   against malicious use or accidents.  In order to be applicable to
   large-scale commercial networks, multicast networks need to have the
   same capabilities which are currently supported by unicast networks.
   Such issues which are important to commercial, large-scale
   implementations of multicasting are listed.  Next, a few possible
   existing architectures used for multicasting with access control
   based on current standards are presented. Specifically 1) IGMP/ MLD,
   2) IGMP/MLD with L2 L2/L3 Authentication with ACL 3) Unicast Control
   with IGMP/MLD and 4) IGMP/MLD with Multicast Encryption will each be
   presented and described.  Each architecture is discussed with
   respect to the presented list of issues.

2. Definitions and Abbreviations

2.1 Definitions

   For the purposes of this I-D memo the following definitions apply:

   Accounting: actions for grasping each user's behavior, when she/he
   starts/stops to receive a channel, which channel she/he receives,
   etc.

   Authentication: action for identifying a user as a genuine one.

   Authorization: action for giving permission to access the content or
   network to a user.

   Receiver: an end-host or end-client which receives content.  A
   receiver may be distinguishable by a network ID such as MAC address
   or IP address.

   Triple Play: voice (VoIP), video, and broadband Internet access
   services.

   User: a human with a user account.  A user may possibly use multiple
   reception devices.  Multiple users may use the same reception device.

   Note: The definition of a receiver (device) and a user (human)
   should not be confused.

2.2 Abbreviations

   For the purposes of this draft the following abbreviations apply:

   ACL: Access Control List

   CDS: Content Delivery Services

   CSP:

   CP: Content Service Provider

   DRM: Data Rights Management

   KEI: Key Exchange Identifier

   NSP: Network Service Provider

   QoS: Quality of Service

3. Common use models and network architecture implications

   Issues such as user identification, access-control, tracking and
   billing are common requirements for commercial content delivery
   services (CDS) systems (and are important in many non-commercial CDS
   systems as well.)  These same requirements should be met for CDS
   systems that employ multicasting.

   In some cases a single entity may design and be responsible for a
   system that covers the various common high-level requirements of a
   commercial multicasting system such as 1) content serving, 2) the
   infrastructure to multicast it, 3) network and content access
   control mechanisms.  In many cases however the content provision and
   network provision roles are divided between separate entities.  The I-D
   draft-ietf-mboned-maccnt-00.txt
   memo draft-ietf-mboned-maccnt-04.txt [3, referred to hereafter in
   this memo as MACCNT-draft] provides more detail of the multiple
   versus single entity CDS network model.

   As such it should not be assumed that the entity responsible for the
   multicasting structure and the entity responsible for content
   serving are the same.  Indeed because the infrastructure for
   multicasting is expensive and many content holders are not likely to
   be competent at building and maintaining complicated infrastructures
   necessary for multicasting, many content holders would prefer to
   purchase transport and management services from a network service
   provider and thus share the infrastructure costs with other content
   holders.

   Similarly commercial network service providers do not generally
   specialize in providing content and are unlikely to build and
   maintain such a resource-intensive system without a certain level of
   demand from content holders.

   The business model of a single NSP network service provider (NSP)
   providing multicasting services to multiple CSP content providers CP has
   certain implications:

        -Need for user tracking and billing capabilities
        -Need for network access control and/or content access control
   satisfactory to the requirements of the CSP CP
        -Methods for sharing information between the NSP and CSP CP to make
   the above two possible

   When the NSP and CSP CP are the same single entity the general
   requirements are as follows.

        -Need for user tracking and user-billing capabilities
        -Need for access control and/or content protection at level the
   entity deems appropriate

   In the next section issues in multicasting related to commercial and
   large-scale implementations are presented.  Some presented issues
   are not pertinent to cases where the NSP and CSP CP are the same entity.

4. Issues in multicasting related to commercial and large-scale
   implementations

   This section lists issues related to receiver access control in
   current multicasting protocols which are especially important to
   commercial, large-scale multicasting.  More details concerning the
   requirements related to  To avoid unnecessary
   duplication with MACCNT-draft, detail for some of these issues are is
   provided through references in a separate I-D
   draft-ietf-mboned-maccnt-00.txt[3]. References to that document are
   provided as applicable below. the Normative Reference section.

4.1 Access limits and resource issues

   For commercial applications of multicasting, network and content
   providers generally wish to be able to control the number of groups
   a host can access at the same time. Also the network provider may
   wish to limit the number of users accessing a multicast stream
   because of bandwidth and processing issues between the receiver and
   the multicast server.  This section corresponds to MACCNT-draft[3],
   section 4.5.14.2 "Issue of network resource protection", and 4.2.1
   "Access control", but provides more detail.

   With best-effort services (e.g. mail transfer, web surfing) strict
   network resource allocation is not necessary, but for services with
   a guaranteed QoS level (e.g. IP television, teleconferencing, VoIP)
   it is necessary to allocate sufficient bandwidth and server
   resources to each service.  In order to guarantee certain QoS levels, it is
   important for network providers to be able to protect their network
   resources from being wasted (either maliciously or accidentally).  More detail on this the topic of network
   resource protection is provided in I-D draft-ietf-mboned-
   maccnt-00.txt, section "Issue of network
   resource protection." protection" of the MACCNT-draft[3].

4.2 Capability to distinguish between receivers (end hosts)

   Currently

   For detail on the sender cannot topic on the capability to distinguish between
   receivers, refer to MACCNT-draft[3], 4.1 the second paragraph which receivers (end hosts)
   are actually receiving its information
   begins with existing "With current protocols
   (IGMP/MLD.)  The sender must rely on the information from the
   multicasting routers. This can be complicated if (IGMP/MLD), the sender and
   routers are maintained by different entities. There is currently no
   standard way to share such information. cannot
   distinguish

4.3 Capability to distinguish between users (as opposed to merely
   hosts)

   Many content providers would like to have detailed information on
   which users (as opposed
   Detail related to merely hosts identified by physical
   addresses, etc.)  are consuming their content and information on
   their usage behavior. More detail on this the topic of user identification can be found in I-D
   draft-ietf-mboned-maccnt-00.txt,
   section "User identification."

4.4 Channel "leave latency"

   Commercial implementations of IP multicasting are likely to have
   strict requirements in terms identification" of user experience.  Leave latency is the time between when a user sends a signal that he/she wishes to
   "leave" a group and when MACCNT-draft[3], the network recognizes the "leave."
   A separate I-D draft-ietf-mboned-maccnt-00.txt provides more first
   paragraph.

4.4 Minimizing Channel Join Latency and Leave Latency

   More detail on this the topic of channel leave latency is provided in the
   section "Channel 'leave latency'" Join Latency and Leave Latency" of the MACCNT-
   draft[3].

4.5 Surveillance of receiver by sender

  4.5.1 Precise access log

   It is necessary logging

   For detail on the topic please refer to precisely log information such as who (host/user)
   is accessing what content at from what time (join action) until what
   time (leave action).  The result of MACCNT-draft[3], 4.6
   "Accounting and billing", especially the access-control decision (e.g.
   results of authorization) would also be valuable information. second paragraph which
   begins with "To assemble such..."

  4.5.2 How to share user information

   For commercial multicast applications where NSP and CSP CP are different
   entities, there are a number of issues regarding how to share user
   information between the NSP and CSP. CP.  For example, which entities
   should be able to access which information relating to user-based
   tracking? What is the user identifier that can be used between the
   entities to distinguish among users, and which entities should be
   able to recognize this identifier?  Another important issue is how
   the edge router should be able to access and then maintain user
   information. The current situation of present architectures is that
   only the NSP can get information about user activity, because user
   activities are only observable from join/leave information logged on
   edge devices which are under control of the NSP. This section
   corresponds to MACCNT-draft[3], section 4.5.1 "How to share user
   information", but provides more detail than in the MACCNT-draft.

  4.5.3 Trustworthy logs to monitor user activity
   An important issue for commercial multicasting applications is how
   the NSP can get trustworthy data on user activity which may be
   needed for billing and statistics purposes.  A standard way of
   logging user activity and protecting the integrity of the logs does
   not exist. Often network providers do not want to keep logs on
   untrusted user terminals which that can be tampered with.

4.6 Notification to users of the result of the join request

   It is necessary to provide information

   Details for this issue are presented in MACCNT-draft[3], section 4.6
   "Notification to users of the user about the status result of his/her the join request(granted/denied/other). request."

4.7 Sharing of Infrastructure for Support of Triple Play

   Ideally Services

   As stated in MACCNT-draft[3], section "Small impact on the existing
   products": "Ideally the NSP should be able to use the same
   infrastructure (such as access control) to support commercial
   multicast services for the so-called "triple play" services: voice (VoIP), video, and broadband
   Internet access services. 'triple play' services".

4.8 DRM Protection

   Digital Rights Management (DRM) is important but out of scope of
   this
   I-D. memo.

5. Description of existing architectures

   In this section, existing architectures used for multicasting based
   on current standards are defined.  In section 6 these architectures
   will be compared by the issues presented in section 4.

5.1 IGMP/MLD

   Internet Group Management Protocol(IGMP) or Multicast Listener
   Discovery (MLD) are protocols for layer 3 management of multicasting.
   In IP multicast a receiver sends a request to a first-hop multicast
   router to join a particular multicast group. The router is then
   responsible for forwarding the appropriate data from the sender to
   the receiver.

     +----------+    +----------+   +----------+         +----------+
     | Sender   |    | Router   |   | L2SW     |         | Receiver |
     |          |    |          |<---------------1,JOIN--|          |
     |          |    |          |   |          |         |          |
     |          |--------------------------------2,Data->|          |
     |          |    |          |   |          |         |          |
     |          |    |          |   |          |         |          |
     +----------+    +----------+   +----------+         +----------+

   For the sake of simplicity, the above diagram only shows the
   sequence of requests for a single receiver.  When multiple receivers
   are requesting the same channel stream the data would be copied at
   the multicasting router to serve the multiple streams.

5.2 IGMP/MLD plus L2/L3 Authentication with Access Control Policy

   With a basic implementation of IGMP/MLD implementation, IGMP/MLD, no authorization is
   performed on the receiver.  It is possible to combine an IGMP/MLD
   implementation with Layer 2 or Layer 3 Authentication to provide an
   access-control mechanism at the first point of attachment to the
   network, for example, using 802.X. 802.1X.

   For example, a receiver may request to an L2 authentication server
   for access to the network. The authentication controller then
   queries the policy server with the receiver's credentials (such as
   IP or MAC address), and if the receiver is determined to be an
   authorized user of the network ("success"), the router downloads the
   ACL from the policy sever. server.  For example, users which are not on the
   ACL are rejected.  Then the Layer 2 Switch is directed to open a
   port for the receiver to send a join request to the multicast router.
   The router is then responsible for forwarding the appropriate data
   from the sender to the receiver.

   Note: ACL is one existing method to realize an access control policy.
   Other methods exist.

     +----------+
     | Policy   |
     | Server   |\
     |          | \
     +----------+  \ 4,ACL Download
        |    ^      \
        |    |       \
        V    |        V
     +----------+    +----------+   +----------+          +----------+
     | L2       |    | Router   |   | L2SW     |          | Receiver |
     |          |    |          |   |          |          |          |
     | Auth.    |<----------------------------  1,Request-|          |
     |          |    |          |   |          |          |          |
     |          |--------2,Success------------>X(3,Auth)  |          |
     +----------+    |          |   |          |          |          |
                     |          |   |          |          |          |
     +----------+    |          |   |          |          |          |
     |          |    |          |<---------------5,Join---|          |
     | Sender   |    |          |   |          |          |          |
     |          |------------>x------------------6,Data-->|          |
     |          |    |        | |   |          |          |          |
     +----------+    +--------|-+   +----------+          +----------+
                              |
                              V
Key:
 Auth: Authentication

5.3 Unicast Control with IGMP/MLD

   The receiver first sends a unicast request to the sender which
   resides in the CP's domain.  This method is the same as that used in
   unicast video-on –demand video-on-demand (VoD) systems.  If authorization is
   successful the sender sends the multicast address information via
   unicast.  With this multicast address the receiver does a IGMP\MLD
   join as in described in 5.1. Generally this approach is relying on
   either some sort of content encryption or "security through
   obscurity" for content security. Also accounting becomes problematic
   because user credentials may not be identified.

     +----------+    +----------+   +----------+          +----------+
     | Sender   |    | Router   |   | L2SW     |          | Receiver |
     |          |    |          |   |          |          |          |
     |          |<------------------------------1,Request-|          |
     |          |    |          |   |          |          |          |
     |          |-------------------------------2,Success>|          |
     |          |    |          |   |          |          |          |
     |          |    |          |<--------------3,Join----|          |
     |          |    |          |   |          |          |          |
     |          |------------>x-----------------4,Data--->|          |
     |          |    |        | |   |          |          |          |
     |          |    |        | |   |          |          |          |
     +----------+    +--------|-+   +----------+          +----------+
                              |
                              V

5.4 IGMP/MLD with Multicast Encryption

   With a basic implementation of IGMP/MLD, no data protection is
   performed on data sent to the receiver.  No credential check is
   performed on the receiver and any receiver can receive and use the
   data.  The IGMP/MLD with Multicast Encryption model assumes that the
   sender is sending encrypted data and that for this data to be useful
   to the receiver it must first request and receive a key from a group
   controller and key server that is synchronized with the content
   encryption occurring on the sender's data.

     +----------+    +----------+   +----------+          +----------+
     | G.C. &   |    | Router   |   | L2SW     |          | Receiver |
     |          |    |          |   |          |          |          |
     | Key S.   |<------------------------------1,Request-|          |
     |          |    |          |   |          |          |          |
     |          |-------------------------------2,Key---->|          |
     +----------+    |          |   |          |          |          |
                     |          |   |          |          |          |
     +----------+    |          |   |          |          |          |
     |          |    |          |<---------------3,Join---|          |
     | Sender   |    |          |   |          |          |          |
     |          |------------>x------------------4,Data-->|          |
     |          |    |        | |   |          |          |          |
     +----------+    +--------|-+   +----------+          +----------+
                              |
                              V
   Key:
   G.C. & Key S.= Group Controller and Key Server

6. Evaluation of architectures by issue

   In this section the various issues raised in section four are
   analyzed by each of the architectures introduced in section five.

6.1 Access limit capabilities, compared by architecture

   Comparison of currently available architectures with respect to
   limiting the access of multicast groups

   - IGMP/MLD:  It is not possible to limit data reception.

   - L2/L3 authentication with access control policy:
   With an ACL it is possible to limit access of multicast groups.
   However it should be discussed as to how scalable this approach is
   because configuring an ACL could be a labor-intensive task.

   - IGMP/MLD with Unicast control
   It is possible for malicious users to reconfigure the receiver's
   terminal to ignore the Unicast control.  In this case, this
   maliciously reconfigured terminal could send a join message even if
   it is rejected by the network.  In such a case, the ineligible
   receiver would be able to receive the multicast.  As such, this
   method may not be strong enough to exclude ineligible access.

   -Multicast Encryption:
   It is possible for receivers to receive IP packets, even if they do
   not possess the keys to decrypt them. A receiver may also be able to
   store such received data until they discover a way to decrypt it.
   Another disadvantage of this method is that network resources are
   wasted if an ineligible receiver receives an encrypted content even
   if they do not have a valid key.

6.2 Capability to distinguish between receivers, compared by
   architecture

   Comparison of currently possible protocol-based solutions.

   -IGMP/MLD:
   The sender has no direct line of contact with the receiver and
   therefore cannot distinguish on a receiver-basis.    (If the edge-
   router's user interface is fixed to the receiver statically assigned then the join-log interface's
   log can be used, used to track joins, but this would mean portability is
   sacrificed.  Moreover, this method is not applicable to a case where
   the CSP CP and NSP are different companies because CSP the CP cannot access
   this join-log.) join-log. Sharing of the join-log could be done with a yet-to-
   be defined standard mechanism/format. )

   -L2/L3 authentication with access control policy:
   At the moment of L2/L3 authentication it is possible to recognize
   receivers, but if there are multiple content service providers (CSP) (CP) a single
   L2 Authorization Server cannot distinguish among the CSPs. CPs.  Therefore
   it would be necessary to gather the join logs.  (If the interface is
   fixed then the join-log can be used, but this would mean portability
   is sacrificed.  Moreover, this method is not applicable to a case
   where the CSP CP and NSP are different companies because the
   CSP CP cannot
   access this join-log.)

   -IGMP/MLD with Unicast control :

   It is possible to distinguish among receivers using Unicast control.
   This latency may not be a problem when users are switching between
   channels of the same CP in cases where the CP grants viewing
   privileges uniformly across all of its channels. However, other
   policies are possible that may be on a channel-basis, time-basis,
   etc. and in such cases channel changing has latency issues.

   -Multicast Encryption:
   If the Content Service Provider CP maintains the Key Server it is possible to distinguish on
   the receiver-level.  If the Network Service Provider maintains the
   key server it is necessary to devise a method for the NSP to notify
   the CSP. CP.

6.3 Capability to distinguish between users, compared by architecture

   Comparison of currently possible protocol-based solutions:

   -IGMP/MLD:
   Since there is no user-based information, it is not possible to
   distinguish on the user-level.

   -L2/L3 authentication with access control policy:
   At the moment of L2/L3 Authentication it is possible to distinguish
   on the user-level.

   However it is difficult to combine user and group logs: it would be
   necessary to match user IDs from L2-Auth logs and group IDs from the
   Join logs to match users and groups.

   -IGMP/MLD with unicast control :
   Distinguishing by user is possible using unicast control.

   -Multicast Encryption:
   If the Content Provider manages the Key Server it is possible to
   distinguish the user.
   If the Network Service Provider manages the Key Server it is
   necessary to notify the Content Provider.

6.4 Maintain guaranteed quality-level of data delivery (Voice, Video),
   compared by architecture

   Comparison of currently possible protocol-based solutions:

   -IGMP/MLD:
    It is not possible to reject a user attempting to access even if
   there are not sufficient resources.

   -L2/L3 authentication with access control policy:
   The AAA server does not know whether there are sufficient resources
   or not.  This method still can provide a guaranteed QoS if every
   channel has the same bandwidth and sufficient bandwidth are
   allocated to each user beforehand.  However, it is not possible to
   provide a guaranteed QoS by comparing the available bandwidth and
   the necessary bandwidth upon each user's request.

   -IGMP/MLD with Unicast control : control:
   When the CSP CP and NSP are separate entities it is not possible for the
   CSP
   CP to make a proper authorization decision because only the NSP
   grasps the network resource availability.

   -Multicast Encryption:
   It is not possible
   Having only encryption provides no access control and therefore
   provides no mechanism to reject a user attempting attempt to access even if
   there are not when
   sufficient resources because are not available (i.e. the user can receive
   data even without holding a valid key. key.)

6.5 Fast leave for fast surfing capability, compared by architecture

   Comparison of currently possible protocol-based solutions:

   -IGMP/MLD:
   It is possible to track on a per host level (based on host address)
   therefore fast leave for fast surfing capability can be achieved.

   -L2/L3 authentication with access control policy:
   It is possible to track on a per host level (based on host address)
   therefore fast leave for fast surfing capability can be achieved.

   -IGMP/MLD with Unicast control :
   Even if a quick leave is possible, changing to a new channel using
   Unicast Control is slow (latency problem).  This latency may not be
   a problem when users are switching between channels of the same CP
   in cases where the CP grants viewing privileges uniformly across all
   of its channels.  However, other policies are possible that may be
   on a channel-basis, time-basis, etc. and in such cases channel
   changing has latency issues.

   -Multicast Encryption:
   Even if a quick leave is possible, delivery of the Key Exchange
   Identifier(KEI) is slow.

6.6 Surveillance of receiver by sender, compared by architecture

   Comparison of currently possible protocol-based solutions:

   -IGMP/MLD:
   With this protocol it is possible to separately log join and leave
   actions, but it is difficult to match these join and leave actions
   because analyzing the logs requires heavy computation (related to
   the scalability with millions of users).

   -L2/L3 authentication with access control policy:
   In this solution, the leave action is not recorded unless some
   additional mechanism such as IGMP/MLD snooping is used.  In some
   cases, users disconnect their terminals without sending leave logout
   messages.  Also it is possible that the user is running multiple
   services and thus they will not log out even if they are finished
   watching video or other multicast content.  In this case, it is not
   possible to precisely determine for accounting purposes when each
   user's
   user disconnected.  Also, it may be a problem that unused bandwidth
   is being needlessly reserved.

   However MLD/IGMP reports/joins need to be refreshed periodically.
   The ACL entry in can be deactivated if the ACL should user no longer refreshes the
   report/join, but this means that user can be charged with unwatched
   programs (125 seconds default.) This lack of precise timing may be deactivated. a
   problem in certain cases such as for paid services.

   -IGMP/MLD with Unicast control :
   In this solution the leave action is not recorded.

   -Multicast Encryption:
   If logs are recorded for each renewal of keys, then it is possible
   to track activity on a per-user basis. However if logs are only
   recorded per content data download then such tracking is not
   possible.

   It should be noted that authentication of the source of each
   join/leave message is important.

6.7.Notification to users of the result of the join request compared by
   architecture

   Comparison of currently possible protocol-based solutions:

   -IGMP/MLD:
   After the join it is not possible to notify the user of the result
   of the join request.

   -L2/L3 authentication with access control policy:
   After the join it is not possible to notify the user of the result
   of the join request.

   -IGMP/MLD with Unicast control :
   After the join it is not possible to notify the user of the result
   of the join request.

   -Multicast Encryption:
   After the join it is not possible to notify the user of the result
   of the join request.

6.8 Comparison summary

   In this section a variety of existing architectures used for
   multicasting based on current standards were compared and evaluated.

   None of these architectures can sufficiently meet all of the common
   requirements for accounting, authentication and authorization in
   commercial, large-scale IP multicasting.  Therefore it is
   recommended that framework(s) for sufficiently addressing such
   requirements be explored.

7. IANA considerations

   This I-D memo does not raise any IANA consideration issues.

8. Security considerations

   This I-D memo does not raise any new security issues which are not
   already existing in original protocols.  Enhancement of multicast
   access control capabilities may enhance security performance.

9. Conclusion

   Issues such as user identification, access-control, tracking and
   billing are common requirements for many content delivery services
   (CDS) systems.  When CDS systems employ multicasting with
   architectures based on currently existing multicasting standards, it
   is often necessary to deploy non-standardized solutions to meet
   these common requirements. It is recommended that requirements be
   defined to serve as a basis set the groundwork for creating standardized ways to framework(s) and
   solutions that address the various issues discussed in this I-D memo
   which are limiting the application of multicasting especially to
   commercial, large-scale CDS services. Such requirements should take
   into consideration a range of possible architectures based on
   multiple business or usage models.

Normative References

   [1] B. Cain, et. al., "Internet Group Management Protocol, Version
       3", RFC3376, October 2002.

   [2] R. Vida, et. al., "Multicast Listener Discovery Version 2
       (MLDv2) for IPv6", RFC3810, June 2004.

   [3] Hayashi, et. al., "Accounting, Authentication and Authorization
       Issues in Well Managed IP Multicasting Services", draft-ietf-
       mboned-maccnt-req-01.txt, October 2005
       mboned-maccnt-req-04.txt, February 2006 [Work in Progress].

   Authors' Addresses

           Tsunemasa Hayashi
           NTT Network Innovation Laboratories
           1-1 Hikari-no-oka, Yokosuka-shi, Kanagawa, 239-0847 Japan
           Phone: +81 46 859 8790
           Email: hayashi.tsunemasa@lab.ntt.co.jp

           Haixiang He
           Nortel Networks
           600 Technology Park Drive
           Billerica, MA 01801, USA
           Phone: +1 978 288 7482
           Email: haixiang@nortelnetworks.com

           Hiroaki Satou
           NTT Network Service Systems Laboratories
           3-9-11 Midoricho, Musashino-shi, Tokyo, 180-8585 Japan
           Phone : +81 422 59 4683
           Email : satou.hiroaki@lab.ntt.co.jp

           Hiroshi Ohta
           NTT Network Service Systems Laboratories
           3-9-11 Midoricho, Musashino-shi, Tokyo, 180-8585 Japan
           Phone : +81 422 59 3617
           Email: ohta.hiroshi@lab.ntt.co.jp

           Susheela Vaidya
           Cisco Systems, Inc.
           170 W. Tasman Drive
           San Jose, CA  95134
           Phone: +1-408-525-1952
           Email: svaidya@cisco.com

Comments

   Comments are solicited and should be addressed to the mboned working
   group's mailing list at mboned@lists.uoregon.edu_and/or the authors.

Full Copyright Statement

   Copyright (C) The Internet Society (2005). (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 IETF's procedures with respect to rights in IETF Documents 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 ietf-
   ipr@ietf.org.

Expiration

   This Internet-Draft will expire on April 15, September 6, 2006.

Acknowledgement

   Funding for the RFC Editor function is currently provided by the
   Internet Society.