draft-ietf-mboned-interdomain-peering-bcp-14.xml

<?xml version='1.0' encoding='utf-8'?>
<!DOCTYPE rfc SYSTEM "rfc2629.dtd">
<rfc submissionType="IETF" docName="draft-ietf-mboned-interdomain-peering-bcp-14" category="bcp" ipr="trust200902"><?rfc compact="yes"?>
	<?rfc text-list-symbols="ooo*+-"?>
	<?rfc sortrefs="no"?>
	<?rfc symrefs="yes"?>
	<?rfc strict="yes"?>
	<?rfc toc="yes"?>
	<front>
	<title abbrev="Multicast Across Inter-Domain Peering Points">Use of Multicast Across Inter-Domain Peering Points</title>
	<author role="editor" fullname="Percy S. Tarapore" initials="P.S." surname="Tarapore">
	<organization>AT&amp;T</organization>
	<address><phone>1-732-420-4172</phone><email>tarapore@att.com</email>
	</address>
	</author>

	<author fullname="Robert Sayko" initials="R." surname="Sayko">
	<organization>AT&amp;T</organization>
	<address><phone>1-732-420-3292</phone><email>rs1983@att.com</email>
	</address>
	</author>

	<author fullname="Greg Shepherd" initials="G." surname="Shepherd">
	<organization>Cisco</organization>
	<address><email>shep@cisco.com</email>
	</address>
	</author>

	<author role="editor" fullname="Toerless Eckert" initials="T." surname="Eckert">
	<organization abbrev="Huawei">Futurewei Technologies Inc.</organization>
	<address><email>tte+ietf@cs.fau.de</email>
	</address>
	</author>

	<author fullname="Ram Krishnan" initials="R." surname="Krishnan">
	<organization>SupportVectors</organization>
	<address><email>ramkri123@gmail.com</email>
	</address>
	</author>

	<date day="30" month="October" year="2017"/>
	<workgroup>MBONED Working Group</workgroup>
	<abstract><t>
   This document examines the use of Source Specific Multicast (SSM)
   across inter-domain peering points for a specified set of deployment
   scenarios. The objective is to describe the setup process for
   multicast-based delivery across administrative domains for these
   scenarios and document supporting functionality to enable this
   process.</t>

	</abstract>
	</front>

	<middle>
	<section title="Introduction" anchor="section-1"><t>

   Content and data from several types of applications (e.g., live
   video streaming, software downloads) are well suited for delivery
   via multicast means. The use of multicast for delivering such
   content or other data offers significant savings of utilization of resources
   in any given administrative domain. End user demand for such
   content or other data is growing. Often, this requires transporting the
   content or other data across administrative domains via inter-domain peering
   points.</t>

<t>The objective of this Best Current Practices document is twofold:

<list style="symbols">

<t>Describe the technical process and establish guidelines for
   setting up multicast-based delivery of application content or other data
   across inter-domain peering points via a set of use cases.</t>

<t>Catalog all required information exchange between the
   administrative domains to support multicast-based delivery.
   This enables operators to initiate necessary processes to
   support inter-domain peering with multicast.</t>

</list></t>

<t>The scope and assumptions for this document are as follows:
<list style="symbols">

<t>Administrative Domain 1 (AD-1) sources content to one or more End Users (EUs) in one or more Administrative Domain 2 (AD-2). AD-1 and AD-2 want to use IP multicast to allow supporting large and growing EU populations with minimum amount of duplicated traffic to send across network links.

<list style="symbols">

    <t>This document does not detail the case where EUs are originating content. To support that additional service, it is recommended to use some method (outside the scope of this document) by which the content from EUs is transmitted to the application in AD-1 that this document refers to as the multicast source and let it send out the traffic as IP multicast. From that point on, the descriptions in this document apply, except that they are not complete because they do not cover the transport or operational aspects of the leg from EU to AD-1.</t>

   <t>This document does not detail the case where AD-1 and AD-2 are
   not directly connected to each other but only via one or more AD-3
   (transit providers). The cases described in this document where tunnels
   are used between AD-1 and AD-2 can be applied to such scenarios, but
   SLA ("Service Level Agreement") control for example would be different. Other additional issues
   will likely exist as well in such scenarios. This is for further study.</t>

</list></t>

        <t>For the purpose of this document, the term "peering point"
        refers to a network connection ("link") between two administrative network
        domains over which traffic is exchanged between them. This is also
        referred to as a Network-to-Network Interface (NNI). Unless otherwise
        noted, the peering point is assumed to be a private peering point,
        where the network connection is a physically or virtually isolated
        network connection solely between AD-1 and AD-2. The other case
        is that of a broadcast peering point which is a common option in
        public Internet Exchange Points (IXP). See <xref target="section-4.2.2"/>
        for more details about that option.</t>

	<t>Administrative Domain 1 (AD-1) is enabled with native
         multicast. A peering point exists between AD-1 and AD-2.</t>

	<t>It is understood that several protocols are available for this
         purpose including PIM-SM and Protocol Independent Multicast -
         Source Specific Multicast (PIM-SSM) <xref target="RFC7761"/>, Internet Group
         Management Protocol (IGMP) <xref target="RFC3376"/>, and Multicast Listener
         Discovery (MLD) <xref target="RFC3810"/>.</t>

	<t>As described in <xref target="section-2"/>, the source IP address of the
         multicast stream in the originating AD (AD-1) is known. Under
         this condition, PIM-SSM use is beneficial as it allows the
         receiver's upstream router to directly send a JOIN message to
         the source without the need of invoking an intermediate
         Rendezvous Point (RP). Use of SSM also presents an improved
         threat mitigation profile against attack, as described in
         <xref target="RFC4609"/>. Hence, in the case of inter-domain peering, it is
         recommended to use only SSM protocols; the setup of inter-
         domain peering for ASM (Any-Source Multicast) is not in scope
         for this document.</t>

	<t>The rest of the document assumes that PIM-SSM and BGP are used across
         the peering point plus AMT and/or GRE according to scenario. The use
         of other protocols is beyond the scope of this document.</t>

	<t>An Automatic Multicast Tunnel (AMT) <xref target="RFC7450"/> is setup at the
         peering point if either the peering point or AD-2 is not
         multicast enabled. It is assumed that an AMT Relay will be
         available to a client for multicast delivery. The selection of
         an optimal AMT relay by a client is out of scope for this
         document. Note that AMT use is necessary only when native
         multicast is unavailable in the peering point (Use Case 3.3) or
         in the downstream administrative domain (Use Cases 3.4, and
         3.5).</t>

	<t>The collection of billing data is assumed to be done at the
         application level and is not considered to be a networking
         issue. The settlements process for end user billing and/or
         inter-provider billing is out of scope for this document.</t>

	<t>Inter-domain network connectivity troubleshooting is only
         considered within the context of a cooperative process between
         the two domains.</t>

</list>
</t>


<t>
   This document also attempts to identify ways by which the peering
   process can be improved. Development of new methods for improvement
   is beyond the scope of this document.</t>

	</section>

	<section title="Overview of Inter-domain Multicast Application Transport" anchor="section-2">

	<t>A multicast-based application delivery scenario is as follows:

	<list style="symbols">
        <t>Two independent administrative domains are interconnected via a peering point.</t>

	<t>The peering point is either multicast enabled (end-to-end
        native multicast across the two domains) or it is connected by
        one of two possible tunnel types:

        <list style="symbols">
            <t>A Generic Routing Encapsulation (GRE) Tunnel <xref target="RFC2784"/>
            allowing multicast tunneling across the peering point, or</t>
  
	  <t>An Automatic Multicast Tunnel (AMT) <xref target="RFC7450"/>.</t>
  
	</list>
	</t>

	<t>A service provider controls one or more application sources in
        AD-1 which will send multicast IP packets via one or more
        (S,G)s (multicast traffic flows,
        see <xref target="section-4.2.1"/> if you are unfamiliar with IP multicast). 

        It is assumed that the service being provided is
        suitable for delivery via multicast (e.g. live video streaming
        of popular events, software downloads to many devices, etc.),
        and that the packet streams will carried by a suitable
        multicast transport protocol.</t>

	<t>An End User (EU) controls a device connected to AD-2, which
        runs an application client compatible with the service
        provider's application source.</t>

	<t>The application client joins appropriate (S,G)s in order to
        receive the data necessary to provide the service to the EU.
        The mechanisms by which the application client learns the
        appropriate (S,G)s are an implementation detail of the
        application, and are out of scope for this document.</t>

	</list>
	</t>

	<t>
   The assumption here is that AD-1 has ultimate responsibility for
   delivering the multicast based service on behalf of the content
   source(s). All relevant interactions between the two domains
   described in this document are based on this assumption.</t>

	<t>
   Note that domain 2 may be an independent network domain (e.g.: Tier
   1 network operator domain). Alternately, domain 2 could also be an
   Enterprise network domain operated by a single customer of AD-1. The peering
   point architecture and requirements may have some unique aspects
   associated with the Enterprise case.</t>

	<t>
   The Use Cases describing various architectural configurations for
   the multicast distribution along with associated requirements is
   described in section 3. Unique aspects related to the Enterprise
   network possibility will be described in this section. <xref target="section-4"/>
   contains a comprehensive list of pertinent information that needs to
   be exchanged between the two domains in order to support functions
   to enable the application transport.</t>

<t>
  Note that domain 2 may be an independent network domain (e.g., Tier
   1 network operator domain). Alternately, domain 2 could also be an
   Enterprise network domain operated by a single customer.
</t>

<t>
  The Use Cases describing various architectural configurations for
   the multicast distribution along with associated requirements is
   described in <xref target="section-3"/>. The peering
   point architecture and requirements may have some unique aspects
   associated with the Enterprise case. These unique aspects will also be
   described in <xref target="section-3"/>. <xref target="section-4"/>
   contains a comprehensive list of pertinent information that needs to
   be exchanged between the two domains in order to support functions
   to enable the application transport.
</t>

   </section>
	<section title="Inter-domain Peering Point Requirements for Multicast" anchor="section-3">
<t>
   The transport of applications using multicast requires that the
   inter-domain peering point is enabled to support such a process.
   There are five Use Cases for consideration in this document.</t>

<section title="Native Multicast" anchor="section-3.1">

	<t>
   This Use Case involves end-to-end Native Multicast between the two
   administrative domains and the peering point is also native
   multicast enabled - see <xref target="native-pic"/>.</t>

<figure anchor="native-pic" title="Content Distribution via End to End Native Multicast">
<artwork><![CDATA[
   -------------------               -------------------
  /       AD-1        \             /        AD-2       \
 / (Multicast Enabled) \           / (Multicast Enabled) \
/                       \         /                       \
| +----+                |         |                       |
| |    |       +------+ |         |  +------+             |   +----+
| | AS |------>|  BR  |-|---------|->|  BR  |-------------|-->| EU |
| |    |       +------+ |   I1    |  +------+             |I2 +----+
\ +----+                /         \                       /
 \                     /           \                     /
  \                   /             \                   /
   -------------------               -------------------

AD = Administrative Domain (Independent Autonomous System)
AS = Application (e.g., Content) Multicast Source
BR = Border Router
I1 = AD-1 and AD-2 Multicast Interconnection (e.g., MBGP)
I2 = AD-2 and EU Multicast Connection
]]></artwork>
</figure>


        <t>Advantages of this configuration are:
	<list style="symbols"><t>Most efficient use of bandwidth in both domains.</t>

	<t>Fewer devices in the path traversed by the multicast stream when
        compared to an AMT enabled peering point.</t>

	</list>
	</t>

	<t>
   From the perspective of AD-1, the one disadvantage associated with
   native multicast into AD-2 instead of individual unicast to every EU
   in AD-2 is that it does not have the ability to count the number of
   End Users as well as the transmitted bytes delivered to them. This
   information is relevant from the perspective of customer billing and
   operational logs. It is assumed that such data will be collected by
   the application layer. The application layer mechanisms for
   generating this information need to be robust enough such that all
   pertinent requirements for the source provider and the AD operator
   are satisfactorily met. The specifics of these methods are beyond
   the scope of this document.</t>

        <t>Architectural guidelines for this configuration are as follows:
	<list style="letters"><t>Dual homing for peering points between domains is recommended
        as a way to ensure reliability with full BGP table visibility.</t>

	<t>If the peering point between AD-1 and AD-2 is a controlled
        network environment, then bandwidth can be allocated
        accordingly by the two domains to permit the transit of non-
        rate adaptive multicast traffic. If this is not the case, then
        the multicast traffic must support rate-adaption (see <xref target="BCP145"/>).</t>

	<t>The sending and receiving of multicast traffic between two
        domains is typically determined by local policies associated
        with each domain. For example, if AD-1 is a service provider
        and AD-2 is an enterprise, then AD-1 may support local policies
        for traffic delivery to, but not traffic reception from, AD-2.
        Another example is the use of a policy by which AD-1 delivers
        specified content to AD-2 only if such delivery has been
        accepted by contract.</t>

	<t>Relevant information on multicast streams delivered to End
        Users in AD-2 is assumed to be collected by available
        capabilities in the application layer. The precise nature and
        formats of the collected information will be determined by
        directives from the source owner and the domain operators.</t>


	</list>
	</t>

</section>

<section title="Peering Point Enabled with GRE Tunnel" anchor="section-3.2">

<t>
   The peering point is not native multicast enabled in this Use Case.
   There is a Generic Routing Encapsulation Tunnel provisioned over the
   peering point. See <xref target="gre-pic"/>.</t>

<figure anchor="gre-pic" title="Content Distribution via GRE Tunnel">
<artwork><![CDATA[
    -------------------              -------------------
   /       AD-1        \            /        AD-2       \
  / (Multicast Enabled) \          / (Multicast Enabled) \
 /                       \        /                       \
 | +----+          +---+ |  (I1)  | +---+                 |
 | |    |   +--+   |uBR|-|--------|-|uBR|   +--+          |   +----+
 | | AS |-->|BR|   +---+-|        | +---+   |BR| -------->|-->| EU |
 | |    |   +--+ <.......|........|........>+--+          |I2 +----+
 \ +----+                /   I1   \                       /
  \                     /   GRE    \                     /
   \                   /   Tunnel   \                   /
    -------------------              -------------------

AD = Administrative Domain (Independent Autonomous System)
AS = Application (e.g., Content) Multicast Source
uBR = unicast Border Router - not necessarily multicast enabled
      may be the same router as BR
BR = Border Router - for multicast
I1 = AD-1 and AD-2 Multicast Interconnection (e.g., MBGP)
I2 = AD-2 and EU Multicast Connection
]]></artwork>
</figure>

<t>In this case, the interconnection I1 between AD-1 and
   AD-2 in <xref target="gre-pic"/> is multicast enabled via a Generic Routing
   Encapsulation Tunnel (GRE) <xref target="RFC2784"/> between the two BR
   and encapsulating the multicast protocols across it.</t>

<t>Normally, this approach is choosen if the uBR physcially connected to the
   peering link can or should not be enabled for IP multicast. This approach
   may also be beneficial if BR and uBR are the same device, but the
   peering link is a broadcast domain (IXP), see <xref target="broadcast-peering"/>.</t>

<t> The routing configuration is
   basically unchanged: Instead of BGP (SAFI2) across the native IP
   multicast link between AD-1 and AD-2, BGP (SAFI2) is now run across
   the GRE tunnel.</t>


<t>Advantages of this configuration:
<list style="symbols">

    <t>Highly efficient use of bandwidth in both domains, although not
    as efficient as the fully native multicast Use Case.</t>
    
    <t>Fewer devices in the path traversed by the multicast stream
    when compared to an AMT enabled peering point.</t>
    
    <t>Ability to support partial and/or incremental IP multicast deployments in AD-
    1 and/or AD-2: Only the path(s) between AS/BR (AD-1) and BR/EU (AD-2) need to
    be multicast enabled. The uBRs may not support IP multicast
    or enabling it could be seen as operationally risky on that important edge node whereas dedicated
    BR nodes for IP multicast may be more acceptable at least initially. BR can also be located such that 
    only parts of the domain may need to support native IP multicast (e.g.: only the
    core in AD-1 but not edge networks towards uBR).</t>
    
    <t>GRE is an existing technology and is relatively simple to implement.</t>

</list>
</t>

	<t>Disadvantages of this configuration:
	<list style="symbols"><t>Per Use Case 3.1, current router technology cannot count the
        number of end users or the number bytes transmitted.</t>

	<t>GRE tunnel requires manual configuration.</t>

	<t>The GRE must be established prior to stream starting.</t>

	<t>The GRE tunnel is often left pinned up.</t>

	</list>
	</t>

	<t>
   Architectural guidelines for this configuration include the
   following:</t>

	<t>
   Guidelines (a) through (d) are the same as those described in Use
   Case 3.1. Two additional guidelines are as follows:</t>

   <t><list hangIndent="3" style="hanging">
        <t hangText="e.">
      GRE tunnels are typically configured manually between peering
      points to support multicast delivery between domains.</t>


        <t hangText="f.">
      It is recommended that the GRE tunnel (tunnel server)
      configuration in the source network is such that it only
      advertises the routes to the application sources and not to the
      entire network. This practice will prevent unauthorized delivery
      of applications through the tunnel (e.g., if application - e.g.,
      content - is not part of an agreed inter-domain partnership).</t>

	</list>
	</t>

</section>
<section title="Peering Point Enabled with an AMT - Both Domains Multicast Enabled" anchor="section-3.3">

	<t>
   Both administrative domains in this Use Case are assumed to be
   native multicast enabled here; however, the peering point is not.</t>

	<t>
   The peering point is enabled with an Automatic Multicast Tunnel. The
   basic configuration is depicted in Figure 2.</t>

<figure title="- AMT Interconnection between AD-1 and AD-2" anchor="ref-amt-interconnection-between-ad-1-and-ad-2">
<artwork><![CDATA[
    -------------------              -------------------
   /       AD-1        \            /        AD-2       \
  / (Multicast Enabled) \          / (Multicast Enabled) \
 /                       \        /                       \
 | +----+          +---+ |   I1   | +---+                 |
 | |    |   +--+   |uBR|-|--------|-|uBR|   +--+          |   +----+
 | | AS |-->|AR|   +---+-|        | +---+   |AG| -------->|-->| EU |
 | |    |   +--+ <.......|........|........>+--+          |I2 +----+
 \ +----+                /  AMT   \                       /
  \                     /  Tunnel  \                     /
   \                   /            \                   /
    -------------------              -------------------

AD = Administrative Domain (Independent Autonomous System)
AS = Application (e.g., Content) Multicast Source
AR = AMT Relay
AG = AMT Gateway
uBR = unicast Border Router - not multicast enabled
      otherwise AR=uBR (AD-1), uBR=AG (AD-2)
I1 = AMT Interconnection between AD-1 and AD-2
I2 = AD-2 and EU Multicast Connection
]]></artwork>
</figure>

<t> Advantages of this configuration:
<list style="symbols">
    <t>Highly efficient use of bandwidth in AD-1.</t>

    <t>AMT is an existing technology and is relatively simple to
       implement. Attractive properties of AMT include the following:

    <list style="symbols">
        <t>Dynamic interconnection between Gateway-Relay pair across the peering point.</t>

	<t>Ability to serve clients and servers with differing policies.</t>

    </list></t>
</list>
</t>

<t>Disadvantages of this configuration:
<list style="symbols">
    <t>Per Use Case 3.1 (AD-2 is native multicast), current router
    technology cannot count the number of end users or the number
    of bytes transmitted to all end users.</t>

    <t>Additional devices (AMT Gateway and Relay pairs) may be
    introduced into the path if these services are not incorporated
    in the existing routing nodes.</t>

    <t>Currently undefined mechanisms for the AG to automatically
       select the optimal AR.</t>

</list>
</t>

	<t>Architectural guidelines for this configuration are as follows:</t>

	<t>
   Guidelines (a) through (d) are the same as those described in Use
   Case 3.1. In addition,</t>

   <t><list hangIndent="3" style="hanging">
        <t hangText="e.">
     It is recommended that AMT Relay and Gateway pairs be
     configured at the peering points to support multicast delivery
     between domains. AMT tunnels will then configure dynamically
     across the peering points once the Gateway in AD-2 receives the
     (S, G) information from the EU.</t>

	</list>
	</t>

</section>
<section title="Peering Point Enabled with an AMT - AD-2 Not Multicast Enabled" anchor="section-3.4">

	<t>
   In this AMT Use Case, the second administrative domain AD-2 is not
   multicast enabled. Hence, the interconnection between AD-2 and the
   End User is also not multicast enabled. This Use Case is depicted in
   Figure 3.</t>

<figure anchor="amt-pic" title="AMT Tunnel Connecting AD-1 AMT Relay and EU Gateway">
<artwork><![CDATA[
   -------------------               -------------------
   /       AD-1        \            /        AD-2       \
  / (Multicast Enabled) \          / (Non Multicast      \
 /                       \        /              Enabled) \ N(large)
 | +----+          +---+ |        | +---+                 |  #EU
 | |    |   +--+   |uBR|-|--------|-|uBR|                 |   +----+
 | | AS |-->|AR|   +---+-|        | +---+    ................>|EU/G|
 | |    |   +--+ <.......|........|...........            |I2 +----+
 \ +----+                / N x AMT\                       /
  \                     /  Tunnel  \                     /
   \                   /            \                   /
    -------------------              -------------------

AS = Application Multicast Source
uBR = unicast Border Router - not multicast enabled,
      otherwise AR = uBR (in AD-1).
AR = AMT Relay
EU/G = Gateway client embedded in EU device
I2 = AMT Tunnel Connecting EU/G to AR in AD-1 through Non-Multicast
   Enabled AD-2.
]]></artwork>
</figure>

<t>This Use Case is equivalent to having unicast distribution of the
application through AD-2. The total number of AMT tunnels would be
   equal to the total number of End Users requesting the application.
   The peering point thus needs to accommodate the total number of AMT
   tunnels between the two domains. Each AMT tunnel can provide the
   data usage associated with each End User.</t>

<t>Advantages of this configuration:
<list style="symbols">
    <t>Efficient use of bandwidth in AD-1 (The closer AR is to uBR, the more efficient).</t>

    <t>Ability for AD-1 to introduce IP multicast based content delivery without any support by network devices in AD-2: Only application side in the EU device needs to perform AMT gateway library functionality to receive traffic from AMT relay.</t>

    <t>Allows for AD-2 to "upgrade" to Use Case 3.5 (see below) at a later time without any change in AD-1 at that time.</t>

    <t>AMT is an existing technology and is relatively simple to implement. Attractive properties of AMT include the following:
    <list style="symbols">
        <t>Dynamic interconnection between Gateway-Relay pair across the peering point.</t>
	<t>Ability to serve clients and servers with differing policies.</t>
    </list> </t>

    <t>Each AMT tunnel serves as a count for each End User and is also able to track data usage (bytes) delivered to the EU.</t>

</list>
</t>

<t> Disadvantages of this configuration:
<list style="symbols">
    <t>Additional devices (AMT Gateway and Relay pairs) are introduced into the transport path.</t>
    <t>Assuming multiple peering points between the domains, the EU Gateway needs to be able to find the "correct" AMT Relay in AD-1.</t>
</list>
</t>

<t>Architectural guidelines for this configuration are as follows:</t>

	<t>
   Guidelines (a) through (c) are the same as those described in Use
   Case 3.1.</t>
 
   <t><list hangIndent="3" style="hanging">
        <t hangText="d.">
     It is necessary that proper procedures are implemented such
     that the AMT Gateway at the End User device is 
     able to find the correct AMT Relay for each (S,G) content stream.
     Standard mechanisms for that selection are still
     subject to ongoing work. This includes use of anycast gateway
     addresses, anycast DNS names, explicit configuration that is mapping (S,G)
     to a relay address or letting the application in the EU/G provide
     the relay address to the embedded AMT gateway function.</t>

        <t hangText="e.">
     The AMT tunnel capabilities are expected to be sufficient for
     the purpose of collecting relevant information on the multicast
     streams delivered to End Users in AD-2.</t>

	</list>
	</t>

</section>
<section title="AD-2 Not Multicast Enabled - Multiple AMT Tunnels Through AD-2" anchor="section-3.5">

        <t>This is a variation of Use Case 3.4 as follows:</t>

<figure anchor="amt-pic2" title="AMT Tunnel Connecting AMT Relay and Relays">
<artwork><![CDATA[
   -------------------               -------------------
   /       AD-1        \            /        AD-2       \
  / (Multicast Enabled) \          / (Non Multicast      \
 /                 +---+ \  (I1)  / +---+        Enabled) \
 | +----+          |uBR|-|--------|-|uBR|                 |
 | |    |   +--+   +---+ |        | +---+           +---+ |   +----+
 | | AS |-->|AR|<........|....    | +---+           |AG/|....>|EU/G|
 | |    |   +--+         |  ......|.|AG/|..........>|AR2| |I3 +----+
 \ +----+                /   I1   \ |AR1|   I2      +---+ /
  \                     /  single  \+---+                /
   \                   / AMT Tunnel \                   /
    -------------------              -------------------

uBR = unicast Border Router - not multicast enabled
      otherwise AR=uBR (AD-1) or ubr=AGAR1 (AD-2)
AS = Application Source
AR = AMT Relay in AD-1
AGAR1 = AMT Gateway/Relay node in AD-2 across Peering Point
I1 = AMT Tunnel Connecting AR in AD-1 to GW in AGAR1 in AD-2
AGAR2 = AMT Gateway/Relay node at AD-2 Network Edge
I2 = AMT Tunnel Connecting Relay in AGAR1 to GW in AGAR2
EU/G = Gateway client embedded in EU device
I3 = AMT Tunnel Connecting EU/G to AR in AGAR2
]]></artwork>
</figure>

	<t>
   Use Case 3.4 results in several long AMT tunnels crossing the entire
   network of AD-2 linking the EU device and the AMT Relay in AD-1
   through the peering point. Depending on the number of End Users,
   there is a likelihood of an unacceptably high amount of traffic
   due to the large number of AMT tunnels - and unicast streams - through
   the peering point. This situation can be alleviated as follows:</t>

<t><list style="symbols">

    <t>Provisioning of strategically located AMT nodes in AD-2
       AD-2. An AMT node comprises co-location of an AMT Gateway and
       an AMT Relay. No change is required by AD-1 compared to 3.4.
       This can be done whenever AD-2 seems fit (too much traffic across
       peering point.</t>

    <t>One such node is at the AD-2 side of the peering
       point (node AGAR1 in above Figure).</t>

    <t>Single AMT tunnel established across peering point linking AMT
       Relay in AD-1 to the AMT Gateway in the AMT node AGAR1 in AD-2.</t>

    <t>AMT tunnels linking AMT node AGAR1 at peering point in AD-2 to
       other AMT nodes located at the edges of AD-2: e.g., AMT tunnel
       I2 linking AMT Relay in AGAR1 to AMT Gateway in AMT node AGAR2
       in Figure 4.</t>

    <t>AMT tunnels linking EU device (via Gateway client embedded in
       device) and AMT Relay in appropriate AMT node at edge of AD-2:
       e.g., I3 linking EU Gateway in device to AMT Relay in AMT node
       AGAR2.</t>

    <t>In the most simple option (not shown), AD-2 only deploys a single
       AGAR1 and lets EU/G build AMT tunnels directly to it. This 
       setup already solves the problem of replicated traffic across
       the peering point. As soon as there is need to support more
       AMT tunnels to EU/G, then additional AGAR2 nodes can be deployed
       by AD-2.</t>

</list>
</t>

	<t>
   The advantage for such a chained set of AMT tunnels is that the
   total number of unicast streams across AD-2 is significantly
   reduced, thus freeing up bandwidth. Additionally, there will be a
   single unicast stream across the peering point instead of possibly,
   an unacceptably large number of such streams per Use Case 3.4.
   However, this implies that several AMT tunnels will need to be
   dynamically configured by the various AMT Gateways based solely on
   the (S,G) information received from the application client at the EU
   device. A suitable mechanism for such dynamic configurations is
   therefore critical.</t>

	<t>Architectural guidelines for this configuration are as follows:</t>

	<t>
   Guidelines (a) through (c) are the same as those described in Use
   Case 3.1.</t>

   <t><list hangIndent="3" style="hanging">
        <t hangText="d.">
     It is necessary that proper procedures are implemented such
     that the various AMT Gateways (at the End User devices and the AMT
     nodes in AD-2) are able to find the correct AMT Relay in other AMT
     nodes as appropriate. Standard mechanisms for that selection are still
     subject to ongoing work. This includes use of anycast gateway
     addresses, anycast DNS names, or explicit configuration that is mapping (S,G)
     to a relay address. On the EU/G, this mapping information may come
     from the application.</t>

        <t hangText="e.">
     The AMT tunnel capabilities are expected to be sufficient for
     the purpose of collecting relevant information on the multicast
     streams delivered to End Users in AD-2.</t>

	</list>
	</t>

  </section>
</section>

<section title="Functional Guidelines" anchor="section-4">

<t> Supporting functions and related interfaces over the peering point
   that enable the multicast transport of the application are listed in
   this section. Critical information parameters that need to be
   exchanged in support of these functions are enumerated, along with
   guidelines as appropriate. Specific interface functions for
   consideration are as follows.</t>

<section title="Network Interconnection Transport Guidelines" anchor="section-4.1">

	<t>
   The term "Network Interconnection Transport" refers to the
   interconnection points between the two Administrative Domains. The
   following is a representative set of attributes that will need to be
   agreed to between the two administrative domains to support
   multicast delivery.</t>

	<t><list style="symbols">
          <t>Number of Peering Points.</t>

	  <t>Peering Point Addresses and Locations.</t>

	  <t>Connection Type - Dedicated for Multicast delivery or shared
          with other services.</t>

	  <t>Connection Mode - Direct connectivity between the two AD's or
          via another ISP.</t>

	  <t>Peering Point Protocol Support - Multicast protocols that will
          be used for multicast delivery will need to be supported at
          these points. Examples of protocols include eBGP <xref target="RFC4760"/> and
          MBGP <xref target="RFC4760"/>.</t>
  
	  <t>Bandwidth Allocation - If shared with other services, then
          there needs to be a determination of the share of bandwidth
          reserved for multicast delivery. See section 4.1.1 below for more details.</t>
  
	  <t>QoS Requirements - Delay and/or latency specifications that need to be
          specified in an SLA.</t>
  
	  <t>AD Roles and Responsibilities - the role played by each AD for
          provisioning and maintaining the set of peering points to
          support multicast delivery.</t>
  
	</list>
	</t>

  <section title="Bandwidth Management" anchor="section-4.1.1">

      <t>Like IP unicast traffic, IP multicast traffic carried across non-controlled
      networks must comply to Congestion Control Principles as described in
      <xref target="BCP41"/> and explained in detail for UDP IP multicast in <xref target="BCP145"/>.</t>
      
      <t>Non-controlled networks (such as the Internet) are those where there is no policy for
      managing bandwidth other than best effort with fair share of bandwidth under congestion.
      As a simplified rule of thumb, complying to congestion control principles means to reduce
      bandwidth under congestion in a way that is fair to competing competing (typically TCP) flow 
      ("rate adaptive").</t>
      
      <t>In many instances, multicast content delivery evolves from intra-domain
      deployments where it is handled as a controlled network service and of
      not complyng to congestion control principles. It was
      given a reserved amount of bandwidth and admitted to the network so
      that congestion never occurs. Therefore the congestion control issue
      should be given specific attention when evolving to an interdomain peering
      deployment.</t>

      <t>In the case where end-to-end IP multicast traffic passes across the
      network of two ADs (and their subsidiaries/customers), both ADs must agree
      on a consistent traffic management policy. If for example AD-1 sources
      non congestion aware IP multicast traffic and AD-2 carries it
      as best effort traffic across links shared with other Internet traffic and
      subject to congestion, this will not work: Under congestion, some amount
      of that traffic will be dropped, rendering the remaining packets
      often as undecodeable garbage clogging up the network in AD-2 and because this
      is not congestion aware, the loss does not reduce this rate. 
      Competing traffic will not get their fair share under
      congestion, and EUs will be frusted by extremely bad quality of both their
      IP multicast and other (e.g.: TCP) traffic. Note that this is not an IP multicast
      technology issue, but solely a transport/application layer issue:
      The problem would equally happen if AD-1 would send non-rate adaptive unicast traffic,,
      for example legacy IPTV video-on-demand traffic which typically is also non
      congestion aware. Because rate adaption in IP unicast video is commonplace today
      because of ABR (Adaptive Bitrate Video), it is very unlikely for this to happen though
      in reality with IP unicast.</t>

      <t>While the rules for traffic management apply whether or not IP multicast
      is tunneled or not, the one feature that can make AMT tunnels
      more difficult is the unpredictability of bandwidth requirements across
      underlying links because of the way they can be used:

      With native IP multicast or GRE tunnels, the amount
      of bandwidth depends on the amount of content, not the number of EUs -
      and is therefore easier to plan for. AMT tunnels terminating in EU/G on the other
      hand scale with the number of EUs. In the vicinity of the AMT relay they can
      introduce very large amount of replicated traffic and it is not always feasible
      to provision enough bandwidth for all possible EU to get the highest quality for
      all their content during peak utilization in such setups - unless the AMT
      relays are very close to the EU edge. Therefore it is also recommended
      to use IP multicast rate adaptation even inside controlled networks when using
      AMT tunnels directly to EU/G.</t>

      <t>Note that rate-adaptive IP multicast traffic in general does not mean that the
      sender is reducing the bitrate, but rather that the EUs that experience congestion
      are joining to a lower bitrate (S,G) stream of the content, similar to adaptive
      bitrate streaming over TCP. Migration from non rate-adaptive
      to rate adaptive bitrate in IP multicast does therefore also change the
      dynamic (S,G) join behavior in the network resulting in potentially higher
      performance requirement for IP multicast protocols (IGMP/PIM), especially 
      on the last hops where dynamic changes occur (including AMT gateway/relays):
      In non rate-adaptive IP multicast, only "channel change" causes state change,
      in rate-adaptive also the congestion situation causes state change.</t>

      <t>Even though not fully specified in this document, peerings that rely on
      GRE/AMT tunnels may be across one or more transit ADs instead of an exclusive
      (non-shared, L1/L2) path. Unless those transit
      ADs are explicitly contracted to provide other than "best effort" transit
      for the tunneled traffic, the IP multicast traffic tunneled must be rate adaptive
      to not violate BCP41 across those transit ADs.</t>

  </section>

</section>

<section title="Routing Aspects and Related Guidelines" anchor="section-4.2">

	<t>
   The main objective for multicast delivery routing is to ensure that
   the End User receives the multicast stream from the "most optimal"
   source <xref target="INF_ATIS_10"/> which typically:</t>

	<t><list style="symbols"><t>Maximizes the multicast portion of the transport and minimizes
        any unicast portion of the delivery, and</t>

	<t>Minimizes the overall combined network(s) route distance.</t>

	</list>
	</t>

	<t>
   This routing objective applies to both Native and AMT; the actual
   methodology of the solution will be different for each. Regardless,
   the routing solution is expected:</t>

	<t><list style="symbols"><t>To be scalable,</t>

	<t>To avoid or minimize new protocol development or modifications,
         and</t>

	<t>To be robust enough to achieve high reliability and
         automatically adjust to changes and problems in the multicast
         infrastructure.</t>

	</list>
	</t>

	<t>
   For both Native and AMT environments, having a source as close as
   possible to the EU network is most desirable; therefore, in some
   cases, an AD may prefer to have multiple sources near different
   peering points. However, that is entirely an implementation issue.</t>

<section title="Native Multicast Routing Aspects" anchor="section-4.2.1">


	<t>
   Native multicast simply requires that the Administrative Domains
   coordinate and advertise the correct source address(es) at their
   network interconnection peering points(i.e., border routers). An
   example of multicast delivery via a Native Multicast process across
   two Administrative Domains is as follows assuming that the
   interconnecting peering points are also multicast enabled:</t>

	<t><list style="symbols"><t>Appropriate information is obtained by the EU client who is a
        subscriber to AD-2 (see Use Case 3.1). This information is in
        the form of metadata and it contains instructions directing the
        EU client to launch an appropriate application if necessary, as
        well as additional information for the application about the
        source location and the group (or stream) id in the form of the
        "S,G" data. The "S" portion provides the name or IP address of
        the source of the multicast stream. The metadata may also
        contain alternate delivery information such as specifying the
        unicast address of the stream.</t>

	<t>The client uses the join message with S,G to join the multicast
        stream <xref target="RFC4604"/>. To facilitate this process, the two AD's need to do the following:
	<list style="symbols"><t>Advertise the source id(s) over the Peering Points.</t>

	<t>Exchange relevant Peering Point information such as Capacity
        and Utilization.</t>

	<t>Implement compatible multicast protocols to ensure proper
        multicast delivery across the peering points.</t>

	</list>
	</t>

	</list>
	</t>

</section>
<section title="GRE Tunnel over Interconnecting Peering Point" anchor="section-4.2.2">

	<t>
   If the interconnecting peering point is not multicast enabled and
   both AD's are multicast enabled, then a simple solution is to
   provision a GRE tunnel between the two AD's - see Use Case 3.2.2.
   The termination points of the tunnel will usually be a network
   engineering decision, but generally will be between the border
   routers or even between the AD 2 border router and the AD 1 source
   (or source access router). The GRE tunnel would allow end-to-end
   native multicast or AMT multicast to traverse the interface.
   Coordination and advertisement of the source IP is still required.</t>

	<t>
   The two AD's need to follow the same process as described in 4.2.1
   to facilitate multicast delivery across the Peering Points.</t>

</section>
<section title="Routing Aspects with AMT Tunnels" anchor="section-4.2.3">

	<t>
   Unlike Native Multicast (with or without GRE), an AMT Multicast
   environment is more complex. It presents a dual layered problem
   because there are two criteria that should be simultaneously met:</t>

<t><list style="symbols">
<t>Find the closest AMT relay to the end-user that also has
    multicast connectivity to the content source, and</t>

<t>Minimize the AMT unicast tunnel distance.</t>
</list></t>

<t>There are essentially two components to the AMT specification</t>

<t><list style="hanging">
<t hangText="AMT Relays:">These serve the purpose of tunneling UDP multicast
        traffic to the receivers (i.e., End-Points). The AMT Relay will
        receive the traffic natively from the multicast media source and
        will replicate the stream on behalf of the downstream AMT
        Gateways, encapsulating the multicast packets into unicast
        packets and sending them over the tunnel toward the AMT Gateway.
        In addition, the AMT Relay may perform various usage and
        activity statistics collection. This results in moving the
        replication point closer to the end user, and cuts down on
        traffic across the network. Thus, the linear costs of adding
        unicast subscribers can be avoided. However, unicast replication
        is still required for each requesting End-Point within the
        unicast-only network.</t>

<t hangText="AMT Gateway (GW):"> The Gateway will reside on an End-Point - this
        could be any type of IP host such as a Personal Computer (PC), mobile phone,
        Set Top Box (STB) or appliances.  The AMT
        Gateway receives join and leave requests from the Application
        via an Application Programming Interface (API). In this manner,
        the Gateway allows the End-Point to conduct itself as a true
        Multicast End-Point. The AMT Gateway will encapsulate AMT
        messages into UDP packets and send them through a tunnel (across
        the unicast-only infrastructure) to the AMT Relay.</t>

</list>
</t>

	<t>
   The simplest AMT Use Case (section 3.3) involves peering points that
   are not multicast enabled between two multicast enabled AD's. An AMT
   tunnel is deployed between an AMT Relay on the AD 1 side of the
   peering point and an AMT Gateway on the AD 2 side of the peering
   point. One advantage to this arrangement is that the tunnel is
   established on an as needed basis and need not be a provisioned
   element. The two AD's can coordinate and advertise special AMT Relay
   Anycast addresses with each other. Alternately, they may decide to
   simply provision Relay addresses, though this would not be an
   optimal solution in terms of scalability.</t>

	<t>
   Use Cases 3.4 and 3.5 describe more complicated AMT situations as
   AD-2 is not multicast enabled. For these cases, the End User device
   needs to be able to setup an AMT tunnel in the most optimal manner.
   There are many methods by which relay selection can be done
   including the use of DNS based queries and static lookup tables
   <xref target="RFC7450"/>. The choice of the method is implementation dependent and
   is up to the network operators. Comparison of various methods is out
   of scope for this document; it is for further study.</t>

	<t>
   An illustrative example of a relay selection based on DNS queries
   and Anycast IP addresses process for Use Cases 3.4 and 3.5 is
   described here. Using an Anycast IP address for AMT Relays allows
   for all AMT Gateways to find the "closest" AMT Relay - the nearest
   edge of the multicast topology of the source. Note that this is
   strictly illustrative; the choice of the method is up to the network
   operators. The basic process is as follows:</t>

	<t><list style="symbols"><t>Appropriate metadata is obtained by the EU client application. The
     metadata contains instructions directing the EU client to an
     ordered list of particular destinations to seek the requested
     stream and, for multicast, specifies the source location and the
     group (or stream) ID in the form of the "S,G" data. The "S"
     portion provides the URI (name or IP address) of the source of the
     multicast stream and the "G" identifies the particular stream
     originated by that source. The metadata may also contain alternate
     delivery information such as the address of the unicast form of
     the content to be used, for example, if the multicast stream
     becomes unavailable.</t>

	<t>Using the information from the metadata, and possibly information
     provisioned directly in the EU client, a DNS query is initiated in
     order to connect the EU client/AMT Gateway to an AMT Relay.</t>

	<t>Query results are obtained, and may return an Anycast address or a
     specific unicast address of a relay. Multiple relays will
     typically exist. The Anycast address is a routable "pseudo-address" shared among the relays that can gain multicast access to
     the source.</t>

	<t>If a specific IP address unique to a relay was not obtained, the
     AMT Gateway then sends a message (e.g., the discovery message) to
     the Anycast address such that the network is making the routing
     choice of particular relay - e.g., closest relay to the EU. 
     Details are outside the scope for this document. See <xref target="RFC4786"/>.</t>

	<t>The contacted AMT Relay then returns its specific unicast IP
     address (after which the Anycast address is no longer required).
     Variations may exist as well.</t>

	<t>The AMT Gateway uses that unicast IP address to initiate a three-
     way handshake with the AMT Relay.</t>

	<t>AMT Gateway provides "S,G" to the AMT Relay (embedded in AMT
     protocol messages).</t>

	<t>AMT Relay receives the "S,G" information and uses the S,G to join
     the appropriate multicast stream, if it has not already subscribed
     to that stream.</t>

	<t>AMT Relay encapsulates the multicast stream into the tunnel
     between the Relay and the Gateway, providing the requested content
     to the EU.</t>

	</list>
	</t>

</section>

<section title="Public Peering Routing Aspects" anchor="section-4.2.4">

<figure anchor="broadcast-peering" title="Broadcast Peering Point">
<artwork><![CDATA[

           AD-1a            AD-1b
           BR                BR
            |                 |
          --+-+---------------+-+-- broadcast peering point LAN
              |                 |
              BR               BR
             AD-2a            AD-2b
    
]]></artwork>
</figure>

<t>A broadcast peering point is an L2 subnet connecting 3 or more ADs.
It is common in IXPs and usually consists of ethernet switch(es)
operated by the IXP connecting to BRs operated by the ADs.</t>

<t>In an example setup domain AD-2a peers with AD-1a and wants to
receive IP multicast from it. Likewise AD-2b peers with
AD-1b and wants to receive IP multicast from it.</t>

<t>Assume one or more IP multicast (S,G) traffic streams can be served
by both AD-1a and AD-1b, for example because both AD-1a and AD-1b
do contract this content from the same content source.</t>

<t>In this case, AD-2a and AD-2b can not control anymore which upstream
domain, AD-1a or AD-1b will forward this (S,G) into the LAN. AD-2a BR
requests the (S,G) from AD-1a BR and AD-2b BR requests the same (S,G)
from AD-1b BR. To avoid duplicate packets, an (S,G) can be forwarded by
 only one router onto the
LAN, and PIM-SM/PIM-SSM detects requests for duplicate transmission and resolve
it via the so-called "assert" protocol operation which results in only
one BR forwarding the traffic. Assume this is AD-1a BR. 
AD-2b will then receive the multicast traffic unexpectedly from a provider
with whom it does not have a mutual agreement for the traffic. Quality issues
in EUs behind AD-2b caused by AD-1a will cause a lot of responsiblity
and troubleshooting issues.</t>

<t>In face of this technical issues, we describe the following options how IP multicast
can be carried across broadcast peering point LANs:</t>

<t><list style="numbers">

<t>IP multicast is tunneled across the LAN. Any of the GRE/AMT tunneling
solutions mentioned in this document are applicable. This is the
one case where specifically a GRE tunnel between the upstream BR
(e.g.: AD-1a) and downstream BR (e.g.: AD-2a) is recommended
as opposed to tunneling across uBRs which are not the actual BRs.</t>

<t>The LAN has only one upstream AD that is sourcing IP multicast
and native IP multicast is used. This is an efficient way to distribute
the same IP multicast content to multiple downstream ADs. Misbehaving
downstream BRs can still disrupt the delivery of IP multicast from the
upstream BR to other downstream BRs, therefore strict rules must be
follow to prohibit that case. The downstream BRs must ensure that
they will always consider only the upstream BR as a source for
multicast traffic: e.g.: no BGP SAFI-2 peerings between the downstream ADs
across the peering point LAN, so that only the upstream BR is the only possible next-hop reachable
across this LAN. And routing policies configured to avoid fall back to the use of SAFI-1 (unicast) routes for IP multicast if unicast BGP peering is not limited in the same way.</t>

<t>The LAN has multiple upstreams, but they are federated and
agree on a consistent policy for IP multicast traffic across the LAN.
One policy is that each possible source is only announced by one upstream
BR. Another policy is that sources are redundantly announced (problematic case mentioned in above example), 
but the upstream domains also provide mutual operational insight to help troubleshooting
(outside the scope of this document).</t>

</list></t>

</section>
</section>
<section title="Back Office Functions - Provisioning and Logging Guidelines" anchor="section-4.3">

	<t>
   Back Office refers to the following:</t>

	<t><list style="symbols">
	<t>Servers and Content Management systems that support the delivery
      of applications via multicast and interactions between AD's.</t>

	<t>Functionality associated with logging, reporting, ordering,
      provisioning, maintenance, service assurance, settlement, etc.</t>

	</list>
	</t>

<section title="Provisioning Guidelines" anchor="section-4.3.1">

	<t>
   Resources for basic connectivity between AD's Providers need to be
   provisioned as follows:</t>

	<t><list style="symbols">
	<t>Sufficient capacity must be provisioned to support multicast-based
      delivery across AD's.</t>

	<t>Sufficient capacity must be provisioned for connectivity between
      all supporting back-offices of the AD's as appropriate. This
      includes activating proper security treatment for these back-
      office connections (gateways, firewalls, etc) as appropriate.</t>

	<t>Routing protocols as needed, e.g. configuring routers to support
      these.</t>

	</list>
	</t>

	<t>
   Provisioning aspects related to Multicast-Based inter-domain
   delivery are as follows.</t>

	<t>
   The ability to receive requested application via multicast is
   triggered via receipt of the necessary metadata. Hence, this
   metadata must be provided to the EU regarding multicast URL - and
   unicast fallback if applicable. AD-2 must enable the delivery of
   this metadata to the EU and provision appropriate resources for this
   purpose.</t>

	<t>
   Native multicast functionality is assumed to be available across
   many ISP backbones, peering and access networks. If, however, native
   multicast is not an option (Use Cases 3.4 and 3.5), then:</t>

	<t><list style="symbols">
	<t>EU must have multicast client to use AMT multicast obtained either
      from Application Source (per agreement with AD-1) or from AD-1 or
      AD-2 (if delegated by the Application Source).</t>

	<t>If provided by AD-1/AD-2, then the EU could be redirected to a
      client download site (note: this could be an Application Source
      site). If provided by the Application Source, then this Source
      would have to coordinate with AD-1 to ensure the proper client is
      provided (assuming multiple possible clients).</t>

	<t>Where AMT Gateways support different application sets, all AD-2
      AMT Relays need to be provisioned with all source &amp; group
      addresses for streams it is allowed to join.</t>

	<t>DNS across each AD must be provisioned to enable a client GW to
      locate the optimal AMT Relay (i.e. longest multicast path and
      shortest unicast tunnel) with connectivity to the content's
      multicast source.</t>

	</list>
	</t>

	<t>
   Provisioning Aspects Related to Operations and Customer Care are
   stated as follows.</t>

	<t>
   Each AD provider is assumed to provision operations and customer
   care access to their own systems.</t>

	<t>
   AD-1's operations and customer care functions must have visibility
   to what is happening in AD-2's network or to the service provided by
   AD-2, sufficient to verify their mutual goals and operations, e.g.
   to know how the EU's are being served. This can be done in two ways:</t>

	<t><list style="symbols">
	<t>Automated interfaces are built between AD-1 and AD-2 such that
      operations and customer care continue using their own systems.
      This requires coordination between the two AD's with appropriate
      provisioning of necessary resources.</t>

	<t>AD-1's operations and customer care personnel are provided access
      directly to AD-2's system. In this scenario, additional
      provisioning in these systems will be needed to provide necessary
      access. Additional provisioning must be agreed to by the two AD's
      to support this option.</t>

	</list>
	</t>

</section>

<section title="Interdomain Authentication Guidelines" anchor="section-4.3.2">


	<t>
   All interactions between pairs of AD's can be discovered and/or be
   associated with the account(s) utilized for delivered applications.
   Supporting guidelines are as follows:</t>

	<t><list style="symbols">
	<t>A unique identifier is recommended to designate each master
      account.</t>

	<t>AD-2 is expected to set up "accounts" (logical facility generally
      protected by credentials such as login passwords) for use by AD-1. Multiple
      accounts and multiple types or partitions of accounts can apply, e.g.
      customer accounts, security accounts, etc.</t>

	</list>
	</t>

<t>The reason to specifically mention the need for AD-1 to initiate
interactions with AD-2 (and use some account for that), as opposed
to the opposite direction is based on the recommended workflow
initiated by customers (see <xref target="section-4.4"/>): The customer
contacts content source (part of AD-1), when AD-1 sees the need to propagate 
the issue, it will interact with AD-2 using the aforementioned guidelines.</t>

</section>
<section title="Log Management Guidelines" anchor="section-4.3.3">

<t> Successful delivery (in terms of user experience) of applications or content via multicast
   between pairs of interconnecting AD's can be improved through the ability to
   exchange appropriate logs for various workflows - troubleshooting,
   accounting and billing, traffic and content transmission optimization, 
   content and application development optimization and so on.</t>

<t>The basic model as explained in before is that the content source and
   on its behalf AD-1 take over primary responsibility for customer experience and the 
   AD-2's support this. The application/content owner is the only participant who
   has and needs full insight into the application level and can map the customer application
   experience to the network traffic flows - which it then with the help of AD-2
   or logs from AD-2 can analyze and interpret.</t>

<t>The main difference between unicast delivery and multicast delivery is that
   the content source can infer a lot more about downstream network problems
   from a unicasted stream than from a multicasted stream: The multicasted stream
   is not per-EU except after the last replication, which is in most cases
   not in AD-1. Logs from the application,  including the receiver side at the EU, 
   can provide insight, but can not help to fully isolate network problems because of the IP multicast 
  per-application operational state built across AD-1 and AD-2 (aka: the (S,G) state and any other feature 
   operational state such as DiffServ QoS).</t>

<t>See <xref target="privacy"/> for more discussions about the privacy considerations
   of the model described here.</t>

<t>Different type of logs are known to help support operations in AD-1 when provided by AD-2.
   This could be done as part of AD-1/AD-2 contracts. Note that except for implied multicast specific
   elements, the options listed here are not unique or novel
   for IP multicast, but they are more important for services novel to the
   operators than for operationally well established services (such as unicast). Therefore
   we detail them as follows:</t>

	<t><list style="symbols">
	<t>Usage information logs at aggregate level.</t>

	<t>Usage failure instances at an aggregate level.</t>

	<t>Grouped or sequenced application access.
      performance, behavior and failure at an aggregate level to support
      potential Application Provider-driven strategies. Examples of
      aggregate levels include grouped video clips, web pages, and sets
      of software download.</t>

	<t>Security logs, aggregated or summarized according to agreement
      (with additional detail potentially provided during security
      events, by agreement).</t>

	<t>Access logs (EU), when needed for troubleshooting.</t>

	<t>Application logs (what is the application doing), when needed for
      shared troubleshooting.</t>

	<t>Syslogs (network management), when needed for shared
      troubleshooting.</t>

	</list>
	</t>

	<t>
   The two AD's may supply additional security logs to each other as
   agreed to by contract(s). Examples include the following:</t>

	<t><list style="symbols">
	<t>Information related to general security-relevant activity which
      may be of use from a protective or response perspective, such as
      types and counts of attacks detected, related source information,
      related target information, etc.</t>

	<t>Aggregated or summarized logs according to agreement (with
      additional detail potentially provided during security events, by
      agreement).</t>

	</list>
	</t>

</section>
</section>
<section title="Operations - Service Performance and Monitoring Guidelines" anchor="section-4.4">

	<t>
   Service Performance refers to monitoring metrics related to
   multicast delivery via probes. The focus is on the service provided
   by AD-2 to AD-1 on behalf of all multicast application sources
   (metrics may be specified for SLA use or otherwise). Associated
   guidelines are as follows:</t>

<t><list style="symbols">
    <t>Both AD's are expected to monitor, collect, and analyze service
    performance metrics for multicast applications. AD-2 provides
    relevant performance information to AD-1; this enables AD-1 to
    create an end-to-end performance view on behalf of the
    multicast application source.</t>

    <t>Both AD's are expected to agree on the type of probes to be
    used to monitor multicast delivery performance. For example,
    AD-2 may permit AD-1's probes to be utilized in the AD-2
    multicast service footprint. Alternately, AD-2 may deploy its
    own probes and relay performance information back to AD-1.</t>

</list>
</t>

	<t>
   Service Monitoring generally refers to a service (as a whole)
   provided on behalf of a particular multicast application source
   provider. It thus involves complaints from End Users when service
   problems occur. EUs direct their complaints to the source provider;
   in turn the source provider submits these complaints to AD-1. The
   responsibility for service delivery lies with AD-1; as such AD-1
   will need to determine where the service problem is occurring - its
   own network or in AD-2. It is expected that each AD will have tools
   to monitor multicast service status in its own network.</t>

	<t><list style="symbols"><t>Both AD's will determine how best to deploy multicast service
        monitoring tools. Typically, each AD will deploy its own set of
        monitoring tools; in which case, both AD's are expected to
        inform each other when multicast delivery problems are
        detected.</t>

	<t>AD-2 may experience some problems in its network. For example,
        for the AMT Use Cases, one or more AMT Relays may be
        experiencing difficulties. AD-2 may be able to fix the problem
        by rerouting the multicast streams via alternate AMT Relays. If
        the fix is not successful and multicast service delivery
        degrades, then AD-2 needs to report the issue to AD-1.</t>

	<t>When problem notification is received from a multicast
        application source, AD-1 determines whether the cause of the
        problem is within its own network or within the AD-2 domain. If
        the cause is within the AD-2 domain, then AD-1 supplies all
        necessary information to AD-2. Examples of supporting
        information include the following:<list style="symbols"><t>Kind of problem(s).</t>

	<t>Starting point &amp; duration of problem(s).</t>

	<t>Conditions in which problem(s) occur.</t>

	<t>IP address blocks of affected users.</t>

	<t>ISPs of affected users.</t>

	<t>Type of access e.g., mobile versus desktop.</t>

	<t>Network locations of affected EUs.</t>

	</list>
	</t>

	<t>Both AD's conduct some form of root cause analysis for
        multicast service delivery problems. Examples of various
        factors for consideration include:<list style="symbols"><t>Verification that the service configuration matches the
            product features.</t>

	<t>Correlation and consolidation of the various customer
            problems and resource troubles into a single root service
            problem.</t>

	<t>Prioritization of currently open service problems, giving
            consideration to problem impact, service level agreement,
            etc.</t>

	<t>Conduction of service tests, including one time tests or a
            series of tests over a period of time.</t>

	<t>Analysis of test results.</t>

	<t>Analysis of relevant network fault or performance data.</t>

	<t>Analysis of the problem information provided by the customer
            (CP).</t>

	</list>
	</t>

	<t>Once the cause of the problem has been determined and the
        problem has been fixed, both AD's need to work jointly to
        verify and validate the success of the fix.</t>

	</list>
	</t>

</section>
<section title="Client Reliability Models/Service Assurance Guidelines" anchor="section-4.5">

	<t>
   There are multiple options for instituting reliability
   architectures, most are at the application level. Both AD's should
   work those out with their contract or agreement and with the multicast
   application source providers.</t>

	<t>
   Network reliability can also be enhanced by the two AD's by
   provisioning alternate delivery mechanisms via unicast means.</t>

	</section>

<section title="Application Accounting Guidelines" anchor="section-4.6">

<t>Application level accounting needs to be handled differently in the application
than in IP unicast because the source side does not directly deliver packets to individual
receivers. Instead, this needs to be signalled back by the receiver to
the source.</t>

<t>
For network transport diagnostics, AD-1 and AD-2 should have mechanisms in place
 to ensure proper accounting for the volume of bytes delivered through the peering
point and separately the number of bytes delivered to EUs.</t>

	</section>

	</section>

	<section title="Troubleshooting and Diagnostics" anchor="section-5"><t>
   Any service provider supporting multicast delivery of content should
   have the capability to collect diagnostics as part of multicast
   troubleshooting practices and resolve network issues accordingly.
   Issues may become apparent or identified either through network
   monitoring functions or by customer reported problems as described
   in section 4.4.</t>

	<t>
   It is recommended that multicast diagnostics will be performed
   leveraging established operational practices such as those documented in <xref target="MDH-04"/>.
   However,
   given that inter-domain multicast creates a significant
   interdependence of proper networking functionality between providers
   there does exist a need for providers to be able to signal (or otherwise alert)
   each other if there are any issues noted by either one.</t>

	<t>
   Service providers may also wish to allow limited read-only
   administrative access to their routers to their AD peers for troubleshooting.
   Of specific interest are access to active troubleshooting tools especially
   <xref target="Traceroute"/> and <xref target="I-D.ietf-mboned-mtrace-v2"/>.</t>


   <t>Another option is to include this functionality into the IP multicast receiver application   
   on the EU device and allow for these diagnostics to be remotely used by support operations.
   Note though that AMT does not allow to pass traceroute or mtrace requests, therefore troubleshooting
   in the presence of AMT does not work as well end-to-end as it can with native
   (or even GRE encapsulated) IP multicast, especially wrt. to traceroute and mtrace.
   Instead, troubleshooting directly on the actual network devices is then more likely necessary.</t>

	<t>
   The specifics of the notification and alerts are beyond the scope of
   this document, but general guidelines are similar to those described
   in section 4.4 (Service Performance and Monitoring). Some general
   communications issues are stated as follows.</t>

	<t><list style="symbols"><t>Appropriate communications channels will be established between
        the customer service and operations groups from both AD's to
        facilitate information sharing related to diagnostic
        troubleshooting.</t>

	<t>A default resolution period may be considered to resolve open
        issues. Alternately, mutually acceptable resolution periods
        could be established depending on the severity of the
        identified trouble.</t>

	</list>
	</t>

	</section>

     
	<section title="Security Considerations" anchor="section-6">

<section  title="DoS attacks (against state and bandwidth)" anchor="section-6.1">

<t>Reliable operations of IP multicast requires some basic protection against
DoS (Denial of Service) attacks.</t>

<t>SSM IP multicast is self protecting against attacks from illicit sources.
Their traffic will not be forwarded beyond the first hop router because
that would require (S,G) memership reports from receiver.
Traffic from sources will only be forwarded from the valid
source because RPF ("Reverse Path Forwarding") is part of the protocols.
 One can say that <xref target="BCP38"/>
style protection against spoofed source traffic is therefore built into PIM-SM/PIM-SSM. 
</t>

<t>Receivers can attack SSM IP multicast by originating such (S,G)
membership reports. This can result in a DoS attack against state through the creation
of a large number of (S,G) states that create high control plane load or even inhibit later
creation of valid (S,G). In conjunction with collaborating illicit sources
it can also result in illicit sources traffic being forwarded.</t>

<t>Today, these type of attacks are usually mitigated 
by explicitly defining the set of permissible (S,G) on e.g.: the last
hop routers in replicating IP multicast to EUs; For example via (S,G) Access Control
Lists applied to IGMP/MLD membership state creation. Each AD is expected
to prohibit (S,G) state creation for invalid sources inside their own
AD.</t>

<t>In the peering case, AD-2 is without further information not aware of the set of valid
(S,G) from AD-1, so this set needs to be communicated via operational
procedures from AD-1 to AD-2 to provide protection against this type of DoS attacks.
Future work could signal this information
in an automated way: BGP extensions, DNS Resource Records or backend automation
between AD-1 and AD-2. Backend automation is the short term most viable solution
because it does not require router software extensions like the other two. 
Observation of traffic flowing via (S,G) state could also be used
to automate recognition of invalid (S,G) state created by receivers in the
absence of explicit information from AD-1. </t>

<t>The second DoS attack through (S,G) membership reports is when receivers create too
much valid (S,G) state to attack bandwidth available to other EU.
Consider the uplink into a last-hop-router connecting to 100 EU. If
one EU joins to more multicast content than what fits into this
link, then this would impact also the quality of the same content
for the other 99 EU. If traffic is not rate adaptive, the effects are even
worse.</t>

<t>The mitigation is the same as what is often employed
for unicast: Policing of per-EU total amont of traffic. 
Unlike unicast though, this can not be done anywhere along the path (e.g.:
on an arbitrary bottleneck link), but it has to happen at the point of
last replication to the different EU. Simple solutions such as
limiting the maximum number of joined (S,G) per EU are readily
available, solutions that consider bandwidth consumed exist as
vendor specific feature in routers. Note that this is primarily
a non-peering issue in AD-2, it only becomes a peering issue if
the peering-link itself is not big enough to carry all possible
content from AD-1 or in case 3.4 where the AMT relay in AD-1 is that
last replication point.</t>

<t>Limiting the amount of (S,G) state per EU is also a good
first measure to prohibit too much undesired "empty" state to
be built (state not carrying traffic), but it would not suffice
in case of DDoS attack - viruses that impact a large number of
EU devices.</t>

</section>

<section  title="Content Security" anchor="section-6.2">

<t>Content confidentiality, DRM (Digital Restrictions Management), authentication and
authorization are optional based on the content delivered. 
For content that is "FTA" (Free To Air), the following
considerations can be ignored and content can be sent unencrypted and
without EU authentication and authorization. Note though that the
mechanisms described here may also be desireable by the application source 
to better track users even if the content itself would not require it.</t>

<t>For interdomain content, there are at least two models for content
confidentiality, DRM and end-user authentication and authorization:</t>

<t>In the classical (IP)TV model, responsibility is per-domain and content
is and can be passed on unencrypted. AD-1 delivers content to AD-2,
AD-2 can further process the content including features like 
ad-insertion and AD-2 is the sole point of contact regarding the contact
for its EUs. In this document, we do not consider this case because
it typically involves higher than network layer service aspects operated
by AD-2 and this document focusses on the network layer AD-1/AD-2
peering case, but not the application layer peering case. Nevertheless,
 this model can be derived through additional work from what is describe here.</t>

<t>The other case is the one in which content confidentiality, DRM,
end-user authentication and authorization are end-to-end:
responsibilities of the multicast application source provider and
receiver application.  This is the model assumed here. It is also
the model used in Internet OTT video delivery. We discuss
the threads incurred in this model due to the use of IP multicast
in AD-1/AD-2 and across the peering.</t>

<t>End-to-end encryption enables end-to-end EU authentication and
authorization: The EU may be able to IGMP/MLD join and receive the content,
but it can only decrypt it when it receives the decryption key from the content source
in AD-1. The key is the authorization. Keeping that key to itself and
prohibiting playout of the decrypted content to non-copy-protected
interfaces are typical DRM features
in that receiver application or EU device operating system.</t>

<t>End-to-end ecnryption is continuously attacked. Keys may be subject
to brute force attack so that content can be decrypted potentially
later, or keys are extracted from the EU application/device and shared
with other unauthenticated receivers. One important class of content
is where the value is in live consumption, such as sports
or other event (concert) streaming. Extraction of keying material
from compromised authenticated EU and sharing with unauthenticated
EU is not sufficient. It is also necessary for those unauthenticated
EUs to get a streaming copy of the content itself. In unicast
streaming, they can not get such a copy from the content source
(because they can not authenticate) and because of asymmetric
bandwidths, it is often impossible to get the content from compromised EUs to large
number of unauthenticated EUs. EUs behind classical 16 Mbps down, 1 Mbps up ADSL links are
the best example.  With increasing broadband access speeds unicast peer-to-peer copying
of content becomes easier, but it likely will always be easily detectable by the ADs
because of its traffic patterns and volume.</t>

<t>When IP multicast is being used without additionals security, AD-2
is not aware which EU is authenticated for which content. Any
unauthenticated EU in AD-2 could therefore get a copy of the encrypted
content without suspicion by AD-2 or AD-1 and either live-deode it in the presence of
compromised authenticated EU and key sharing, or later decrypt it in the
presence of federated brute force key cracking.</t>

<t>To mitigate this issue, the last replication point that is
creating (S,G) copies to EUs would need to permit those copies only after
authentication of EUs. This would establish the same authenticated EU only
copy deliver thast is used in unicast.</t>

<t>Schemes for per EU IP multicast authentication/authorization (and
in result non-delivery/copying of per-content IP multicast traffic)
have been built in the past and are deployed in service providers for
intradomain IPTV services, but no standard exist for this. For example,
there is no standardized radius attribute for authenticating the IGMP/MLD
filter set, but implementations of this exist. The authors
are specifically also not aware of schemes where the same authentication
credentials used to get the encryption key from the content source
could also be used to authenticate and authorize the network layer IP
multicast replication for the content. Such schemes are technically not difficult to build 
and would avoid creating and maintaining a separate network forwarding
authentication/authorization scheme decoupled from the end-to-end
authentication/authorization system of the application.</t>

<t>If delivery of such high value content in conjunction with the
peering described here is desired, the short term recommendations are
for sources to clearly isolate the source and group addresses used
for different content bundles, communicate those (S,G) patterns from
AD-1 to the AD-2 and let AD-2 leverage existing per-EU authentication/
authorization mechanisms in network devices to establish filters for
(S,G) sets to each EU.</t>

</section>

<section  title="Peering Encryption" anchor="section-6.3">

<t>Encryption at peering points for multicast delivery may be used per agreement
between AD-1/AD-2.</t>

<t>In the case of a private peering link, IP multicast does not have attack
 vectors on a peering link different from those of IP unicast, but the content owner may have
defined high bars against unauthenticated copying of even the end-to-end encrypted
content, and in this case AD-1/AD-2 can agree on additional transport encryption across
that peering link. In the case of a broadcast peering connection (e.g.: IXP),
transport encryption is also the easiest way to prohibit unauthenticated copies by
other ADs on the same peering point.</t>

<t>If peering is across a tunnel going across intermittent transit ADs
(not discused in detail in this document), then encryption of that tunnel
traffic is recommended. It not only prohibits possible "leakage"
of content, but also to protects the the information what content is
being consumed in AD-2 (aggregated privacy protection).</t>

<t>See the following subsection for reasons why the peering point may
also need to be encrypted for operational reasons.</t>

</section>

<section  title="Operational Aspects" anchor="section-6.4">

<t><xref target="section-4.3.3"/> discusses exchange of log information,
this section discussed exchange of (S,G) information and <xref target="privacy"/>
discusses exhange of program information. All these operational pieces of data 
should by default be exchanged via authenticated and encrypted peer-to-peer communication
protocols between AD-1 and AD-2 so that only the intended recipient in the peers
AD have access to it. Even exposure of the least sensitive information to
third parties opens up attack vectors. Putting for example valid (S,G) information
into DNS (as opposed to passing it via secured channels from AD-1 to AD-2)
 to allow easier filtering of invalid (S,G) would also allow attackers
to easier identify valid (S,G) and change their attack vector.</t>

<t>From the perspective of the ADs, security is most critical for the log
information as it provides operational insight into the originating AD, but
it also contains sensitive user data:</t>

<t>Sensitive user data exported from AD-2 to AD-1 as part of logs could be as much as 
the equivalent of 5-tuple unicast traffic flow accounting (but not more, e.g.: no
application level information). As mentioned in <xref target="privacy"/>,
in unicast, AD-1 could capture these traffic statistics itself because this
is all about AD-1 originated traffic flows to EU receivers in AD-2, and operationally
passing it from AD-2 to AD-1 may be necessary when IP multicast is used because of the replication happening
in AD-2.</t>

<t>Nevertheless, passing such traffic statistics inside AD-1 from a capturing router
to a backend system is likely less subject to third party attacks then passing it interdomain
from AD-2 to AD-1, so more diligence needs to be applied to secure it.</t>

<t>If any protocols used for the operational information exchange are not easily secured at transport layer or
higher (because of the use of legacy products or protocols in the network), then AD-1 and AD-2 can also
consider to ensure that all operational data exchange goes across the
same peering point as the traffic and use network layer encryption of the
peering point as discussed in before to protect it.</t>

<t>
End-to-end authentication and authorization of EU may involve some kind of token authentication and
is done at the application layer independently of the two AD's. If
there are problems related to failure of token authentication when
end-users are supported by AD-2, then some means of validating
proper working of the token authentication process (e.g., back-end
servers querying the multicast application source provider's token
authentication server are communicating properly) should be
considered. Implementation details are beyond the scope of this
document.</t>

<t>Security Breach Mitigation Plan - In the event of a security
breach, the two AD's are expected to have a mitigation plan for
shutting down the peering point and directing multicast traffic
over alternative peering points. It is also expected that
appropriate information will be shared for the purpose of
securing the identified breach.</t>

</section>
</section>

<section title="Privacy Considerations" anchor="privacy">

<t>The described flow of information about content and the end-user described in this
document aims to maintain privacy:</t>

<t>AD-1 is operating on behalf (or owns) the content source and is therefore part of the content-consumption
relationship with the end-user. The privacy considerations between the EU and AD-1
are therefore in general (exception see below) the same as if no IP multicast was used,
 especially because for any privacy conscious content, end-to-end encryption can
and should be used.</t>

<t>Interdomain multicast transport service related information is provided by the
AD-2 operators to AD-1. AD-2 is not required to gain additional insight into the user
behavior through this process that it would not already have without the service collaboration
with AD-1 - unless AD-1 and AD-2 agree on it and get approval from the EU.</t>

<t>For example, if it is deemed beneficial for
EU to directly get support from AD-2 then it would in general be necessary for
AD-2 to be aware of the mapping between content and network (S,G) state so that AD-2
knows which (S,G) to troubleshoot when the EU complains about problems with a specific
content. The degree to which this dissemination is done by AD-1 explicitly to meet
privacy expectations of EUs is typically easy to assess by AD-1. Two
simple examples:</t>

<t>For a sports content bundle, every EU will happily click on the "i approve that
the content program information is shared with your service provider" button, to ensure
best service reliability because service conscious AD-2 would likely also try to
ensure that high value content, such as the (S,G) for SuperBowl like content 
would be the first to receive care in case of network issues.</t>

<t>If the content in question was one where the EU expected more privacy, the EU
should prefer a content bundle that included this content in a large variety of other
content, have all content end-to-end encrypted and the programming information not be
shared with AD-2 to maximize privacy. Nevertheless,
the privacy of the EU against AD-2 observing traffic would still be lower than in the equivalent
setup using unicast, because in unicast, AD-2 could not correlate which EUs are watching
the same content and use that to deduce the content. Note that even
the setup in <xref target="section-3.4"/> where AD-2 is not involved in IP multicast
at all does not provide privacy against this level of analysis by AD-2 because
there is no transport layer encryption in AMT and therefore AD-2 can
correlate by onpath traffic analysis who is consuming the same content from an AMT
relay from both the (S,G) join messages in AMT and the identical content segments (that
where replicated at the AMT relay).</t>

<t>In summary: Because only content to be consumed by multiple EUs is carried via
IP multicast here, and all that content can be end-to-end encrypted, the only IP
multicast specific privacy consideration is for AD-2 to know or reconstruct what content
an EU is consuming. For content for which this is undesirable, some form of protections
as explained above are possible, but ideally, the model of <xref target="section-3.4"/>
could be used in conjunction with future work adding
e.g.: dTLS <xref target="RFC6347"/> encryption between AMT relay and EU.</t>

<t>Note that IP multicast by nature would permit the EU privacy against the
countent source operator because unlike unicast, the content source does not
natively know which EU is consuming which content: In all cases where AD-2 provides
replication, only AD-2 does know this directly. This document does not attempt
to describe a model that does maintain such level of privacy against the content
source but only against exposure to intermediate parties, in this case AD-2.</t>

</section>

	<section title="IANA Considerations" anchor="section-8"><t>
   No considerations identified in this document.</t>

	</section>

<!-- Removed through review by Mirja Kuehlewind on version -11
     Toerless: a wrap up might make sense, but agree that the following
     text is not very helpfull, and not very important to have wrap-up,
     so righ now removing it without replacing it.

	<section title="Conclusions" anchor="section-8"><t>
   This Best Current Practice document provides detailed Use Case
   scenarios for the transmission of applications via multicast across
   peering points between two Administrative Domains. A detailed set of
   guidelines supporting the delivery is provided for all Use Cases.</t>

	<t>
   For Use Cases involving AMT tunnels (cases 3.4 and 3.5), it is
   recommended that proper procedures are implemented such that the
   various AMT Gateways (at the End User devices and the AMT nodes in
   AD-2) are able to find the correct AMT Relay in other AMT nodes as
   appropriate. Section 4.2 provides an overview of one method that
   finds the optimal Relay-Gateway combination via the use of an
   Anycast IP address for AMT Relays.</t>


	</section>
-->

	<section title="Acknowledgments" anchor="acknowledgments"><t>
   The authors would like to thank the following individuals for their
   suggestions, comments, and corrections:</t>

	<t>
   Mikael Abrahamsson</t>

	<t>
   Hitoshi Asaeda</t>

	<t>
   Dale Carder</t>

	<t>
   Tim Chown</t>

	<t>
   Leonard Giuliano</t>

	<t>
   Jake Holland</t>

	<t>
   Joel Jaeggli</t>

	<t>
   Albert Manfredi</t>

	<t>
   Stig Venaas</t>

<t>Henrik Levkowetz</t>

	</section>

<section title="Change log [RFC Editor: Please remove]" anchor="changelog">

<t>Please see discussion on mailing list for changes before -11.</t>
<t>-11: version in IESG review.</t>
<t>-12: XML'ified version of -11, committed solely to make rfcdiff easier. XML versions hosted on https://www.github.com/toerless/peering-bcp</t>
<t>-13:
<list style="symbols">
    <t>IESG feedback. Complete details in: https://raw.githubusercontent.com/toerless/peering-bcp/master/11-iesg-review-reply.txt</t>
    <t>Ben Campbell: Location information about EU (End User) is Network Locatio information</t>
    <t>Ben Campbell: Added explanation of assumption to introduction that traffic is sourced from AD-1 to (one or many) AD-2, mentioned that sourcing from EU is out of scope.</t>
    <t>Introduction: moved up bullet points about exchanges and transit to clean up flow of assumptions.</t>
    <t>Ben Campbell: Added picture for the GRE case, visualized tunnels in all pictures.</t>
    <t>Ben Campbell: See 13-discus.txt on github for more details of changes for this review.</t>
    <t>Alissia Cooper: Added more explanation for Log Management, explained privacy context.</t>
    <t>Alissia Cooper: removed pre pre-RFC5378 disclaimer.</t>
    <t>Alissia Cooper: removed mentioning of potential mutual compensation between domains if the other violates SLA.</t>
    <t>Mirja Kuehlewind: created section 4.1.1 to discuss congestion control more detailled, adding reference to BCP145, removed stub CC paragraphs from section 3.1 (principle applies to every section 3.x, and did not want to duplicate text between 3.x and 4.x).</t>
    <t>Mirja Kuehlewind: removed section 8 (conclusion). Text was not very good, not important to hae conclusion, maybe bring back with better text if strong interest.</t>
    <t>Introduced section about broadcast peering points because there where too many places already where references to that case existed (4.2.4).</t>
    <t>Introduced section about privact considerations because of comment by Ben Campbell and Alissa Cooper.</t>
    <t>Rewrote security considerations and structured it into key aspects: DoS attacks, content protection, peering point encryption and operational aspects.</t>
    <t>Kathleen Moriarty: Added operational aspects to security section (also for Alissia), e.g.: covering securing the exchange of operational data between ADs.</t>
    <t>Spencer Dawkins: Various editorial fixes. Removed BCP38 text from section 3, superceeded be explanation of PIM-SM RPF check to provide equvialent security to BCP38 in security section 7.1).</t>
    <t>Eric Roscorla: (fixed from other reviews already).</t>
    <t>Adam Roach: Fixed up text about MDH-04, added reference to RFC4786.</t>
</list></t>
<t>-13: Fix for Mirja's review on must for congestion control.</t>

	</section>



	</middle>

	<back>
	<references title="Normative References">


      <?rfc include='reference.RFC.2784'?>
      <?rfc include='reference.RFC.3376'?>
      <?rfc include='reference.RFC.3810'?>
      <?rfc include='reference.RFC.4760'?>
      <?rfc include='reference.RFC.4604'?>
      <?rfc include='reference.RFC.4609'?>
      <?rfc include='reference.RFC.7450'?>
      <?rfc include='reference.RFC.7761'?>


<reference  anchor='BCP38' target='https://www.rfc-editor.org/info/rfc2827'>
<front>
<title>Network Ingress Filtering: Defeating Denial of Service Attacks which employ IP Source Address Spoofing</title>
<author initials='P.' surname='Ferguson' fullname='P. Ferguson'><organization /></author>
<author initials='D.' surname='Senie' fullname='D. Senie'><organization /></author>
<date year='2000' month='May' />
</front>
<seriesInfo name='BCP' value='38'/>
<seriesInfo name='RFC' value='2827'/>
<seriesInfo name='DOI' value='10.17487/RFC2827'/>
</reference>


<reference  anchor='BCP41' target='https://www.rfc-editor.org/info/rfc2914'>
<front>
<title>Congestion Control Principles</title>
<author initials='S.' surname='Floyd' fullname='S. Floyd'><organization /></author>
<date year='2000' month='September' />
</front>
<seriesInfo name='BCP' value='41'/>
<seriesInfo name='RFC' value='2914'/>
<seriesInfo name='DOI' value='10.17487/RFC2914'/>
</reference>


<reference  anchor='BCP145' target='https://www.rfc-editor.org/info/rfc8085'>
<front>
<title>UDP Usage Guidelines</title>
<author initials='L.' surname='Eggert' fullname='L. Eggert'><organization /></author>
<author initials='G.' surname='Fairhurst' fullname='G. Fairhurst'><organization /></author>
<author initials='G.' surname='Shepherd' fullname='G. Shepherd'><organization /></author>
<date year='2017' month='March' />
</front>
<seriesInfo name='BCP' value='145'/>
<seriesInfo name='RFC' value='8085'/>
<seriesInfo name='DOI' value='10.17487/RFC8085'/>
</reference>



	</references>

	<references title="Informative References">
      <?rfc include='reference.RFC.4786'?>
      <?rfc include='reference.RFC.6347'?>


	<reference anchor="INF_ATIS_10"><front>
	<title>CDN Interconnection Use Cases and Requirements in a Multi-Party Federation Environment</title>
	<author>
	</author>
	<date month="December" year="2012"/>
	</front>
	<seriesInfo name="ATIS" value="Standard A-0200010"/>
	</reference>

	<reference anchor="MDH-04"><front>
	<title>Multicast Debugging Handbook</title>
	<author fullname="D. Thaler" initials="D." surname="Thaler">
        </author>
	<author fullname="others" surname="others">
	</author>
	<date month="May" year="2000"/>
	</front>
        <seriesInfo name="IETF I-D" value="draft-ietf-mboned-mdh-04.txt"/>
	</reference>

	<reference anchor="Traceroute" target="http://traceroute.org/#source%20code"><front>
	<title/>
	<author>
	</author>

	<date/>
	</front>

	</reference>
  <?rfc include="reference.I-D.ietf-mboned-mtrace-v2"?>
	</references>

	</back>

	</rfc>