RFC 9188




Independent Submission                                            J. Zhu
Request for Comments: 9188                                         Intel
Category: Experimental                                       S. Kanugovi
ISSN: 2070-1721                                                    Nokia
                                                           February 2022


           Generic Multi-Access (GMA) Encapsulation Protocol

Abstract



   A device can be simultaneously connected to multiple networks, e.g.,
   Wi-Fi, LTE, 5G, and DSL.  It is desirable to seamlessly combine the
   connectivity over these networks below the transport layer (L4) to
   improve the quality of experience for applications that do not have
   built-in multi-path capabilities.  Such optimization requires
   additional control information, e.g., a sequence number, in each
   packet.  This document presents a new lightweight and flexible
   encapsulation protocol for this need.  The solution has been
   developed by the authors based on their experiences in multiple
   standards bodies including the IETF and 3GPP.  However, this document
   is not an Internet Standard and does not represent the consensus
   opinion of the IETF.  This document will enable other developers to
   build interoperable implementations in order to experiment with the
   protocol.

Status of This Memo



   This document is not an Internet Standards Track specification; it is
   published for examination, experimental implementation, and
   evaluation.

   This document defines an Experimental Protocol for the Internet
   community.  This is a contribution to the RFC Series, independently
   of any other RFC stream.  The RFC Editor has chosen to publish this
   document at its discretion and makes no statement about its value for
   implementation or deployment.  Documents approved for publication by
   the RFC Editor are not candidates for any level of Internet Standard;
   see Section 2 of RFC 7841.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at
   https://www.rfc-editor.org/info/rfc9188.

Copyright Notice



   Copyright (c) 2022 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (https://trustee.ietf.org/license-info) in effect on the date of
   publication of this document.  Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.

Table of Contents



   1.  Introduction
     1.1.  Scope of Experiment
   2.  Conventions Used in This Document
   3.  Use Case
   4.  GMA Encapsulation Methods
     4.1.  Trailer-Based IP Encapsulation
     4.2.  Header-Based IP Encapsulation
     4.3.  Header-Based Non-IP Encapsulation
     4.4.  IP Protocol Identifier
   5.  Fragmentation
   6.  Concatenation
   7.  Security Considerations
   8.  IANA Considerations
   9.  References
     9.1.  Normative References
     9.2.  Informative References

   Authors' Addresses



1.  Introduction



   A device can be simultaneously connected to multiple networks, e.g.,
   Wi-Fi, LTE, 5G, and DSL.  It is desirable to seamlessly combine the
   connectivity over these networks below the transport layer (L4) to
   improve the quality of experience for applications that do not have
   built-in multi-path capabilities.

   Figure 1 shows the Multi-Access Management Service (MAMS) user-plane
   protocol stack [MAMS], which has been used in today's multi-access
   solutions [ATSSS] [LWIPEP] [GRE1] [GRE2].  It consists of two layers:
   convergence and adaptation.

   The convergence layer is responsible for multi-access operations,
   including multi-link (path) aggregation, splitting/reordering,
   lossless switching/retransmission, fragmentation, concatenation, etc.
   It operates on top of the adaptation layer in the protocol stack.
   From the perspective of a transmitter, a User Payload (e.g., IP
   packet) is processed by the convergence layer first and then by the
   adaptation layer before being transported over a delivery connection;
   from the receiver's perspective, an IP packet received over a
   delivery connection is processed by the adaptation layer first and
   then by the convergence layer.

          +-----------------------------------------------------+
          |   User Payload, e.g., IP Protocol Data Unit (PDU)   |
          +-----------------------------------------------------+
       +-----------------------------------------------------------+
       |  +-----------------------------------------------------+  |
       |  | Multi-Access (MX) Convergence Layer                 |  |
       |  +-----------------------------------------------------+  |
       |  +-----------------------------------------------------+  |
       |  | MX Adaptation   | MX Adaptation   | MX Adaptation   |  |
       |  | Layer           | Layer           | Layer           |  |
       |  +-----------------+-----------------+-----------------+  |
       |  | Access #1 IP    | Access #2 IP    | Access #3 IP    |  |
       |  +-----------------------------------------------------+  |
       |                            MAMS User-Plane Protocol Stack |
       +-----------------------------------------------------------+

                  Figure 1: MAMS User-Plane Protocol Stack

   GRE (Generic Routing Encapsulation) [LWIPEP] [GRE1] [GRE2] can be
   used as the encapsulation protocol at the convergence layer to encode
   additional control information, e.g., key and sequence number.
   However, there are two main drawbacks with this approach:

   *  It is difficult to introduce new control fields because the GRE
      header formats are already defined, and

   *  IP-over-IP tunneling (required for GRE) leads to higher overhead
      especially for small packets.

   For example, the overhead of IP-over-IP/GRE tunneling with both key
   and sequence Number is 32 bytes (20-byte IP header + 12-byte GRE
   header), which is 80% of a 40-byte TCP ACK packet.

   This document presents a lightweight Generic Multi-Access (GMA)
   encapsulation protocol for the convergence layer.  It supports three
   encapsulation methods: trailer-based IP encapsulation, header-based
   IP encapsulation, and non-IP encapsulation.  Particularly, the IP
   encapsulation methods avoid IP-over-IP tunneling overhead (20 bytes),
   which is 50% of a 40-byte TCP ACK packet.  Moreover, it introduces
   new control fields to support fragmentation and concatenation, which
   are not available in GRE-based solutions [LWIPEP] [GRE1] [GRE2].

   The GMA protocol only operates between endpoints that have been
   configured to use GMA.  This configuration can be through any control
   messages and procedures, including, for example, Multi-Access
   Management Services [MAMS].  Moreover, UDP or IPsec tunneling can be
   used at the adaptation sublayer to protect GMA operation from
   intermediate nodes.

   The solution described in this document was developed by the authors
   based on their experiences in multiple standards bodies including the
   IETF and 3GPP.  However, this document is not an Internet Standard
   and does not represent the consensus opinion of the IETF.  This
   document presents the protocol specification to enable
   experimentation as described in Section 1.1 and to facilitate other
   interoperable implementations.

1.1.  Scope of Experiment



   The protocol described in this document is an experiment.  The
   objective of the experiment is to determine whether the protocol
   meets the requirements, can be safely used, and has support for
   deployment.

   Section 4 describes three possible encapsulation methods that are
   enabled by this protocol.  Part of this experiment is to assess
   whether all three mechanisms are necessary or whether, for example,
   all implementations are able to support the main "trailer-based" IP
   encapsulation method.  Similarly, the experiment will investigate the
   relative merits of the IP and non-IP encapsulation methods.

   It is expected that this protocol experiment can be conducted on the
   Internet since the GMA packets are identified by an IP protocol
   number and the protocol is intended for single-hop propagation;
   devices should not be forwarding packets, and if they do, they will
   not need to examine the payload, while destination systems that do
   not support this protocol should not receive such packets and will
   handle them as unknown payloads according to normal IP processing.
   Thus, experimentation is conducted between consenting end systems
   that have been mutually configured to participate in the experiment
   as described in Section 7.

   Note that this experiment "reuses" the IP protocol identifier 114 as
   described in Section 4.4.  Part of this experiment is to assess the
   safety of doing this.  The experiment should consider the following
   safety mechanisms:

   *  GMA endpoints SHOULD detect non-GMA IP packets that also use 114
      and log an error to report the situation (although such error
      logging MUST be subject to rate limits).

   *  GMA endpoints SHOULD stop using 114 and switch to non-IP
      encapsulation, i.e., UDP encapsulation (Figure 7), after detecting
      any non-GMA usage of 114.

   The experiment SHOULD use a packet tracing tool, e.g., WireShark or
   TCPDUMP, to monitor both ingress and egress traffic at GMA endpoints
   and ensure the above safety mechanisms are implemented.

   Path quality measurements (one-way delay, loss, etc.) and congestion
   detection are performed by the receiver based on the GMA control
   fields, e.g., Sequence Number, Timestamp, etc.  The receiver will
   then dynamically control how to split or steer traffic over multiple
   delivery connections accordingly.  The GMA control protocol [GMAC]
   MAY be used for signaling between GMA endpoints.  Another objective
   of the experiment is to evaluate the usage of various receiver-based
   algorithms [GCC] [MPIP] in multi-path traffic management and the
   impact on the End-to-End (E2E) performance (throughput, delay, etc.)
   of higher-layer (transport) protocols, e.g., TCP, QUIC, WebRTC, etc.

   The authors will continually assess the progress of this experiment
   and encourage other implementers to contact them to report the status
   of their implementations and their experiences with the protocol.

2.  Conventions Used in This Document



   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
   "OPTIONAL" in this document are to be interpreted as described in
   BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all
   capitals, as shown here.

3.  Use Case



   As shown in Figure 2, a client device (e.g., smartphone, laptop,
   etc.) may connect to the Internet via both Wi-Fi and LTE connections,
   one of which (e.g., LTE) may operate as the anchor connection, and
   the other (e.g., Wi-Fi) may operate as the delivery connection.  The
   anchor connection provides the IP address and connectivity for end-
   to-end Internet access, and the delivery connection provides an
   additional path between the client and Multi-Access Gateway for
   multi-access optimizations.

                        Multi-Access Aggregation

                    +---+                        +---+
                    | |A|--- LTE Connection -----|C| |
                    |U|-|                        |-|S| Internet
                    | |B|--- Wi-Fi Connection ---|D| |
                    +---+                        +---+
                   client                Multi-Access Gateway

                Figure 2: GMA-Based Multi-Access Aggregation

   A:  The adaptation-layer endpoint of the LTE connection resides in
       the client.



   B:  The adaptation-layer endpoint of the Wi-Fi connection resides in
       the client.



   C:  The adaptation-layer endpoint of the LTE connection resides in
       the Multi-Access Gateway, aka N-MADP (Network Multi-Access Data
       Proxy) in [MAMS].



   D:  The adaptation-layer endpoint of the Wi-Fi connection resides in
       the Multi-Access Gateway.



   U:  The convergence-layer endpoint resides in the client.



   S:  The convergence-layer endpoint resides in the Multi-Access
       Gateway.



   For example, per-packet aggregation allows a single IP flow to use
   the combined bandwidth of the two connections.  In another example,
   packets lost due to a temporary link outage may be retransmitted.
   Moreover, packets may be duplicated over multiple connections to
   achieve high reliability and low latency, where duplicated packets
   are eliminated by the receiving side.  Such multi-access optimization
   requires additional control information, e.g., a sequence number, in
   each packet, which can be supported by the GMA encapsulation protocol
   described in this document.

   The GMA protocol described in this document is designed for multiple
   connections, but it may also be used when there is only one
   connection between two endpoints.  For example, it may be used for
   loss detection and recovery.  In another example, it may be used to
   concatenate multiple small packets and reduce per-packet overhead.

4.  GMA Encapsulation Methods



   The GMA encapsulation protocol supports the following three methods:

   *  Trailer-based IP Encapsulation (Section 4.1)

   *  Header-based IP Encapsulation (Section 4.2)

   *  Header-based non-IP Encapsulation (Section 4.3)

   Non-IP encapsulation MUST be used if the original IP packet is IPv6.

   Trailer-based IP encapsulation MUST be used if it is supported by GMA
   endpoints and the original IP packet is IPv4.

   Header-based encapsulation MUST be used if the trailer-based method
   is not supported by either the client or Multi-Access Gateway.  In
   this case, if the adaptation layer, e.g., UDP tunneling, supports
   non-IP packet format, non-IP encapsulation MUST be used; otherwise,
   header-based IP encapsulation MUST be used.

   If non-IP encapsulation is configured, a GMA header MUST be present
   in every packet.  In comparison, if IP encapsulation is configured, a
   GMA header or trailer may be added dynamically on a per-packet basis,
   and it indicates the presence of a GMA header (or trailer) to set the
   protocol type of the GMA PDU to "114" (see Section 4.4).

   The GMA endpoints MAY configure the GMA encapsulation method through
   control signaling or pre-configuration.  For example, the "MX UP
   Setup Configuration Request" message as specified in Multi-Access
   Management Service [MAMS] includes "MX Convergence Method
   Parameters", which provides the list of parameters to configure the
   convergence layer, and can be extended to indicate the GMA
   encapsulation method.

   GMA endpoint MUST discard a received packet and MAY log an error to
   report the situation (although such error logging MUST be subject to
   rate limits) under any of the following conditions:

   *  The GMA version number in the GMA header (or trailer) is not
      understood or supported by the GMA endpoint.

   *  A flag bit in the GMA header (or trailer) not understood or
      supported by the GMA endpoint is set to "1".

4.1.  Trailer-Based IP Encapsulation



          |<---------------------GMA PDU ----------------------->|
          +------------------------------------------------------+
          | IP hdr |        IP payload             | GMA Trailer |
          +------------------------------------------------------+
          |<------- GMA SDU (user payload)-------->|

        Figure 3: GMA PDU Format with Trailer-based IP Encapsulation

   This method SHALL NOT be used if the original IP packet (GMA service
   data unit (GMA SDU)) is IPv6.

   Figure 3 shows the trailer-based IP encapsulation GMA protocol data
   unit (GMA PDU) format.  A (GMA) PDU may carry one or multiple IP
   packets, aka (GMA) SDUs, in the payload, or a fragment of the SDU.

   The protocol type field in the IP header of the GMA PDU MUST be
   changed to 114 (Any 0-Hop Protocol) (see Section 4.4) to indicate the
   presence of the GMA trailer.

   The following three IP header fields MUST be changed:

   IP Length:  Add the length of "GMA Trailer" to the length of the
      original IP packet.

   Time To Live (TTL):  Set to "1".

   IP checksum:  Recalculate after changing "protocol type", "TTL", and
      "IP Length".

   The GMA (Generic Multi-Access) trailer MUST consist of two mandatory
   fields (the last 3 bytes): Next Header and Flags.

   This is defined as follows:

   Next Header (1 byte):  This is the IP protocol type of the (first)
      SDU in a PDU; it stores the value before it was overwritten to
      114.

   Flags (2 bytes):  Bit 0 is the most significant bit (MSB), and bit 15
      is the least significant bit (LSB).

      Checksum Present (bit 0):  If the Checksum Present bit is set to
         1, then the Checksum field is present.

      Concatenation Present (bit 1):  If the Concatenation Present bit
         is set to 1, then the PDU carries multiple SDUs, and the First
         SDU Length field is present.

      Connection ID Present (bit 2):  If the Connection ID Present bit
         is set to 1, then the Connection ID field is present.

      Flow ID Present (bit 3):  If the Flow ID Present bit is set to 1,
         then the Flow ID field is present.

      Fragmentation Present (bit 4):  If the Fragmentation Present bit
         is set to 1, then the PDU carry a fragment of the SDU and the
         Fragmentation Control field is present.

      Delivery SN Present (bit 5):  If the Delivery SN (Sequence Number)
         Present bit is set to 1, then the Delivery SN field is present
         and contains the valid information.

      Flow SN Present (bit 6):  If the Flow SN Present bit is set to 1,
         then the Sequence Number field is present.

      Timestamp Present (bit 7):  If the Timestamp Present bit is set to
         1, then the Timestamp field is present.

      TTL Present (bit 8):  If the TTL Present bit is set to 1, then the
         TTL field is present.

      Reserved (bit 9-12):  This is set to "0" and ignored on receipt.

      Version (bit 13~15):  This is the GMA version number; it is set to
         0 for the GMA encapsulation protocol specified in this
         document.

   The Flags field is at the end of the PDU, and the Next Header field
   is the second last.  The receiver SHOULD first decode the Flags field
   to determine the length of the GMA trailer and then decode each
   optional field accordingly.  The Generic Multi-Access (GMA) trailer
   MAY consist of the following optional fields:

   Checksum (1 byte):  This contains the (one's complement) checksum sum
      of all 8 bits in the trailer.  For purposes of computing the
      checksum, the value of the Checksum field is zero.  This field is
      present only if the Checksum Present bit is set to 1.

   First SDU Length (2 bytes):  This is the length of the first IP
      packet in the PDU, only included if a PDU contains multiple IP
      packets.  This field is present only if the Concatenation Present
      bit is set to 1.

   Connection ID (1 byte):  This contains an unsigned integer to
      identify the anchor and delivery connection of the GMA PDU.  This
      field is present only if the Connection ID Present bit is set to
      1.

      Anchor Connection ID (MSB 4 bits):  This contains an unsigned
         integer to identify the anchor connection.

      Delivery Connection ID (LSB 4 bits):  This contains an unsigned
         integer to identify the delivery connection.

   Flow ID (1 byte):  This contains an unsigned integer to identify the
      IP flow that a PDU belongs to, for example Data Radio Bearer (DRB)
      ID [LWIPEP] for a cellular (e.g., LTE) connection.  This field is
      present only if the Flow ID Present bit is set to 1.

   Fragmentation Control (FC) (1 byte):  This provides necessary
      information for reassembly, only needed if a PDU carries
      fragments.  This field is present only if the Fragmentation
      Present bit is set to 1.  Please refer to Section 5 for its
      detailed format and usage.

   Delivery SN (1 byte):  This contains an auto-incremented integer to
      indicate the GMA PDU transmission order on a delivery connection.
      Delivery SN is needed to measure packet loss of each delivery
      connection and therefore generated per delivery connection per
      flow.  This field is present only if the Delivery SN Present bit
      is set to 1.

   Flow SN (3 bytes):  This contains an auto-incremented integer to
      indicate the GMA SDU (IP packet) order of a flow.  Flow SN is
      needed for retransmission, reordering, and fragmentation.  It
      SHALL be generated per flow.  This field is present only if the
      Flow SN Present bit is set to 1.

   Timestamp (4 bytes):  This contains the current value of the
      timestamp clock of the transmitter in the unit of 1 millisecond.
      This field is present only if the Timestamp Present bit is set to
      1.

   TTL (1 byte):  This contains the TTL value of the original IP header
      if the GMA SDU is IPv4, or the Hop-Limit value of the IP header if
      the GMA SDU is IPv6.  This field is present only if the TTL
      Present bit is set to 1.

   Figure 4 shows the GMA trailer format with all the fields present,
   and the order of the GMA control fields SHALL follow the bit order in
   the Flags field.  Note that the bits in the Flags field are ordered
   with the first bit transmitted being bit 0 (MSB).  All fields are
   transmitted in regular network byte order and appear in reverse order
   to their corresponding flag bits.  If a flag bit is clear, the
   corresponding optional field is absent.

   For example, bit 0 (the MSB) of the Flags field is the Checksum
   Present bit, and the Checksum field is the last in the trailer with
   the exception of the two mandatory fields.  Bit 1 is the
   Concatenation Present bit, and the FSL field is the second last.

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |     TTL       |                   Timestamp
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
                   |                   Flow SN                     |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Delivery SN  |    FC         |   Flow ID     | Connection ID |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |      First SDU Length (FSL)   |   Checksum    |  Next Header  |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |         Flags                 |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

       Figure 4: GMA Trailer Format with All Optional Fields Present

4.2.  Header-Based IP Encapsulation



   This method SHALL NOT be used if the original IP packet (GMA SDU) is
   IPv6.

   Figure 5 shows the header-based IP encapsulation format.  Here, the
   GMA header is inserted right after the IP header of the GMA SDU, and
   the IP header fields of the GMA PDU MUST be changed the same way as
   in trailer-based IP encapsulation.

          +-----------------------------------------------+
          |IP hdr | GMA Header  |       IP payload        |
          +-----------------------------------------------+

        Figure 5: GMA PDU Format with Header-Based IP Encapsulation

   Figure 6 shows the GMA header format.  In comparison to the GMA
   trailer, the only difference is that the Flags field is now in the
   front so that the receiver can first decode the Flags field to
   determine the GMA header length.

   The "TTL" field MUST be included and the "TTL" bit in the GMA header
   (or Trailer) MUST be set to 1 if (trailer- or header-based) IP
   encapsulation is used.

       +------------------------------------------------------+
       | Flags | other fields (TTL, Timestamp, Flow SN, etc.) |
       +------------------------------------------------------+

                        Figure 6: GMA Header Format

4.3.  Header-Based Non-IP Encapsulation



   Figure 7 shows the header-based non-IP encapsulation format.  Here,
   "UDP Tunneling" is configured at the MX adaptation layer.  The ports
   for "UDP Tunneling" at the client are chosen from the Dynamic Port
   range, and the ports for "UDP Tunneling" at the Multi-Access Gateway
   are configured and provided to the client through additional control
   messages, e.g., [MAMS].

   "TTL", "FSL", and "Next Header" are no longer needed and MUST not be
   included.  Moreover, the IP header fields of the GMA SDU remain
   unchanged.

    +-------------------------------------------------------------+
    | IP hdr | UDP hdr  | GMA Header | IP hdr |    IP payload     |
    +-------------------------------------------------------------+
                                    |<------- GMA SDU------------>|
                        |<------------------- GMA PDU------------>|

             Figure 7: GMA PDU Format with Non-IP Encapsulation

4.4.  IP Protocol Identifier



   As described in Section 4.1, IP-encapsulated GMA PDUs are indicated
   using the IP protocol type 114.  This is designated and recorded by
   IANA [IANA] to indicate "any 0-Hop Protocol".  No reference is given
   in the IANA registry for the definition of this protocol type, and
   IANA has no record of why the assignment was made or how it is used,
   although it was probably assigned before 1999 [IANA1999].

   There is some risk associated with "reusing" protocol type 114
   because there may be implementations of other protocols also using
   this protocol type.  However, because the protocol described in this
   document is used only between adjacent devices specifically
   configured for this purpose, the use of protocol type 114 should be
   safe.

   As described in Section 1.1, one of the purposes of the experiment
   described in this document is to verify the safety of using this
   protocol type.  Deployments should be aware of the risk of a clash
   with other uses of this protocol type.

5.  Fragmentation



   If the MTU size of the anchor connection (for GMA SDU) is configured
   such that the corresponding GMA PDU length adding the GMA header (or
   trailer) and other overhead (UDP tunneling) MAY exceed the MTU of a
   delivery connection, GMA endpoints MUST be configured to support
   fragmentation through additional control messages [MAMS].

   The fragmentation procedure at the convergence sublayer is similar to
   IP fragmentation [RFC0791] in principle, but with the following two
   differences for less overhead:

   *  The fragment offset field is expressed in number of fragments.

   *  The maximum number of fragments per SDU is 2^7 (=128).

   The Fragmentation Control (FC) field in the GMA trailer (or header)
   contains the following bits:

   Bit 7:  a More Fragment (MF) flag to indicate if the fragment is the
      last one (0) or not (1)

   Bit 0-6:  Fragment Offset (in units of fragments) to specify the
      offset of a particular fragment relative to the beginning of the
      SDU

   A PDU carries a whole SDU without fragmentation if the FC field is
   set to all "0"s or the FC field is not present in the trailer.
   Otherwise, the PDU contains a fragment of the SDU.

   The Flow SN field in the trailer is used to distinguish the fragments
   of one SDU from those of another.  The Fragment Offset (FO) field
   tells the receiver the position of a fragment in the original SDU.
   The More Fragment (MF) flag indicates the last fragment.

   To fragment a long SDU, the transmitter creates n PDUs and copies the
   content of the IP header fields from the long PDU into the IP header
   of all the PDUs.  The length field in the IP header of the PDU SHOULD
   be changed to the length of the PDU, and the protocol type SHOULD be
   changed to 114.

   The data of the long SDU is divided into n portions based on the MTU
   size of the delivery connection.  The first portion of the data is
   placed in the first PDU.  The MF flag is set to "1", and the FO field
   is set to "0".  The i-th portion of the data is placed in the i-th
   PDU.  The MF flag is set to "0" if it is the last fragment and set to
   "1" otherwise.  The FO field is set to i-1.

   To assemble the fragments of an SDU, the receiver combines PDUs that
   all have the same Flow SN.  The combination is done by placing the
   data portion of each fragment in the relative order indicated by the
   Fragment Offset in that fragment's GMA trailer (or header).  The
   first fragment will have the Fragment Offset set to "0", and the last
   fragment will have the More Fragment flag set to "0".

   GMA fragmentation operates above the IP layer of individual access
   connection (Wi-Fi, LTE) and between the two endpoints of convergence
   layer.  The convergence layer endpoints (client, Multi-access
   Gateway) SHOULD obtain the MTU of individual connection through
   either manual configuration or implementing Path MTU Discovery
   (PMTUD) as suggested in [RFC8900].

6.  Concatenation



   The convergence sublayer MAY support concatenation if a delivery
   connection has a larger maximum transmission unit (MTU) than the
   original IP packet (SDU).  Only the SDUs with the same client IP
   address and the same Flow ID MAY be concatenated.

   If the (trailer- or header-based) IP encapsulation method is used,
   the First SDU Length (FSL) field SHOULD be included in the GMA
   trailer (or header) to indicate the length of the first SDU.
   Otherwise, the FSL field SHOULD not be included.

     +-----------------------------------------------------------+
     |IP hdr| IP payload    |IP hdr|   IP payload  | GMA Trailer |
     +-----------------------------------------------------------+

         Figure 8: Example of GMA PDU Format with Concatenation and
                       Trailer-Based IP Encapsulation

   To concatenate two or more SDUs, the transmitter creates one PDU and
   copies the content of the IP header field from the first SDU into the
   IP header of the PDU.  The data of the first SDU is placed in the
   first portion of the data of the PDU.  The whole second SDU is then
   placed in the second portion of the data of the PDU (Figure 8).  The
   procedure continues until the PDU size reaches the MTU of the
   delivery connection.  If the FSL field is present, the IP Length
   field of the PDU SHOULD be updated to include all concatenated SDUs
   and the trailer (or header), and the IP checksum field SHOULD be
   recalculated if the packet is IPv4.

   To disaggregate a PDU, if the (header- or trailer-based) IP
   encapsulation method is used, the receiver first obtains the length
   of the first SDU from the FSL field and decodes the first SDU.  The
   receiver then obtains the length of the second SDU based on the
   length field in the second SDU IP header and decodes the second SDU.
   The procedure continues until no byte is left in the PDU.  If the
   non-IP encapsulation method (Figure 7) is used, the IP header of the
   first SDU will not change during the encapsulation process, and the
   receiver SHOULD obtain the length of the first SDU directly from its
   IP header (Figure 9).

                                    |<-------1st GMA SDU------------
   +---------------------------------------------------------------+
   | IP hdr | UDP hdr  | GMA Header | IP hdr |       IP payload    |
   +---------------------------------------------------------------+
            | IP hdr |           IP payload    |
   +-------------------------------------------+
   -------->|<-------2nd GMA SDU--------------->

         Figure 9: Example of GMA PDU Format with Concatenation and
                  Header-Based Non-IP (UDP) Encapsulation

   If a PDU contains multiple SDUs, the Flow SN field is for the last
   SDU, and the Flow SN of other SDUs carried by the same PDU can be
   obtained according to its order in the PDU.  For example, if the SN
   field is 6 and a PDU contains 3 SDUs (IP packets), the SN is 4, 5,
   and 6 for the first, second, and last SDU, respectively.

   GMA concatenation can be used for packing small packets of a single
   application, e.g., TCP ACKs, or from multiple applications.  Notice
   that a single GMA flow may carry multiple application flows (TCP,
   UDP, etc.).

   GMA endpoints MUST NOT perform concatenation and fragmentation in a
   single PDU.  If a GMA PDU carries a fragmented SDU, it MUST NOT carry
   any other (fragmented or whole) SDU.

7.  Security Considerations



   Security in a network using GMA should be relatively similar to
   security in a normal IP network.  GMA is unaware of IP- or higher-
   layer end-to-end security as it carries the IP packets as opaque
   payload.  Deployers are encouraged to not consider that GMA adds any
   form of security and to continue to use IP- or higher-layer security
   as well as link-layer security.

   The GMA protocol at the convergence sublayer is a 0-hop protocol and
   relies on the security of the underlying network transport paths.
   When this cannot be assumed, appropriate security protocols (IPsec,
   DTLS, etc.)  SHOULD be configured at the adaptation sublayer.  On the
   other hand, packet filtering requires either that a firewall looks
   inside the GMA packet or that the filtering is done on the GMA
   endpoints.  In those environments in which this is considered to be a
   security issue, it may be desirable to terminate the GMA operation at
   the firewall.

   Local-only packet leak prevention (HL=0, TTL=1) SHOULD be on
   preventing the leak of the local-only GMA PDUs outside the "local
   domain" to the Internet or to another domain that could use the same
   IP protocol type, i.e., 114.

8.  IANA Considerations



   This document has no IANA actions.

9.  References



9.1.  Normative References



   [GRE1]     Dommety, G., "Key and Sequence Number Extensions to GRE",
              RFC 2890, DOI 10.17487/RFC2890, September 2000,
              <https://www.rfc-editor.org/info/rfc2890>.

   [GRE2]     Leymann, N., Heidemann, C., Zhang, M., Sarikaya, B., and
              M. Cullen, "Huawei's GRE Tunnel Bonding Protocol",
              RFC 8157, DOI 10.17487/RFC8157, May 2017,
              <https://www.rfc-editor.org/info/rfc8157>.

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

9.2.  Informative References



   [ATSSS]    3GPP, "Study on access traffic steering, switch and
              splitting support in the 5G System (5GS) architecture",
              Release 16, 3GPP TR 23.793, December 2018,
              <https://portal.3gpp.org/desktopmodules/Specifications/
              SpecificationDetails.aspx?specificationId=3254>.

   [GCC]      Holmer, S., Lundin, H., Carlucci, G., De Cicco, L., and S.
              Mascolo, "A Google Congestion Control Algorithm for Real-
              Time Communication", Work in Progress, Internet-Draft,
              draft-ietf-rmcat-gcc-02, 8 July 2016,
              <https://datatracker.ietf.org/doc/html/draft-ietf-rmcat-
              gcc-02>.

   [GMAC]     Zhu, J. and M. Zhang, "UDP-based Generic Multi-Access
              (GMA) Control Protocol", Work in Progress, Internet-Draft,
              draft-zhu-intarea-gma-control-00, 13 October 2021,
              <https://datatracker.ietf.org/doc/html/draft-zhu-intarea-
              gma-control-00>.

   [IANA]     IANA, "Protocol Numbers",
              <https://www.iana.org/assignments/protocol-numbers>.

   [IANA1999] IANA, Wayback Machine archive of "Protocol Numbers",
              February 1999,
              <https://web.archive.org/web/19990203044112/
              http://www.isi.edu:80/in-notes/iana/assignments/protocol-
              numbers>.

   [LWIPEP]   3GPP, "Evolved Universal Terrestrial Radio Access
              (E-UTRA); LTE-WLAN Radio Level Integration Using Ipsec
              Tunnel (LWIP) encapsulation; Protocol specification",
              Release 13, 3GPP TS 36.361, July 2020,
              <https://portal.3gpp.org/desktopmodules/Specifications/
              SpecificationDetails.aspx?specificationId=3037>.

   [MAMS]     Kanugovi, S., Baboescu, F., Zhu, J., and S. Seo, "Multiple
              Access Management Services Multi-Access Management
              Services (MAMS)", RFC 8743, DOI 10.17487/RFC8743, March
              2020, <https://www.rfc-editor.org/info/rfc8743>.

   [MPIP]     Sun, L., Tian, G., Zhu, G., Liu, Y., Shi, H., and D. Dai,
              "Multipath IP Routing on End Devices: Motivation, Design,
              and Performance", 2017,
              <https://eeweb.engineering.nyu.edu/faculty/yongliu/docs/
              MPIP_Tech.pdf>.

   [RFC0791]  Postel, J., "Internet Protocol", STD 5, RFC 791,
              DOI 10.17487/RFC0791, September 1981,
              <https://www.rfc-editor.org/info/rfc791>.

   [RFC8900]  Bonica, R., Baker, F., Huston, G., Hinden, R., Troan, O.,
              and F. Gont, "IP Fragmentation Considered Fragile",
              BCP 230, RFC 8900, DOI 10.17487/RFC8900, September 2020,
              <https://www.rfc-editor.org/info/rfc8900>.

Authors' Addresses



   Jing Zhu
   Intel

   Email: jing.z.zhu@intel.com


   Satish Kanugovi
   Nokia

   Email: satish.k@nokia.com