Transmission of SCHC-compressed packets over IEEE 802.15.4 networks

Transmission of SCHC-compressed packets over IEEE 802.15.4 networks UPC

C/Esteve Terradas, 7 Castelldefels 08860 Spain carles.gomez@upc.edu

Consultant

Rue de Rennes Cesson-Sevigne 35510 France anaminaburo@gmail.com

INT 6lo Working Group A framework called Static Context Header Compression and fragmentation (SCHC) has been designed with the primary goal of supporting IPv6 over Low Power Wide Area Network (LPWAN) technologies . One of the SCHC components is a header compression mechanism. If used properly, SCHC header compression allows a greater compression ratio than that achievable with traditional 6LoWPAN header compression . For this reason, it may make sense to use SCHC header compression in some 6LoWPAN environments, including IEEE 802.15.4 networks. This document specifies how a SCHC-compressed packet can be carried over IEEE 802.15.4 networks. The document also enables the transmission of SCHC-compressed UDP/CoAP headers over 6LoWPAN-compressed IPv6 packets.

RFC 6282 is the main specification for IPv6 over Low power Wireless Personal Area Network (6LoWPAN) IPv6 header compression . That RFC was designed assuming IEEE 802.15.4 as the layer below the 6LoWPAN adaptation layer, and it has also been reused by the IPv6 over Networks of Resource-constrained Nodes (6lo) working group (with proper adaptations) for IPv6 header compression over many other technologies relatively similar to IEEE 802.15.4 in terms of characteristics such as physical layer bit rate, layer 2 maximum payload size, etc. Examples of such technologies comprise BLE, DECT-ULE, ITU G.9959, MS/TP, NFC, and PLC. RFC 6282 provides additional functionality, such as a mechanism for UDP header compression. In the best cases, RFC 6282 allows to compress a 40-byte IPv6 header down to a 2-byte compressed header (for link-local interactions) or a 3-byte compressed header (when global IPv6 addresses are used). On the other hand, RFC 6282 typically compresses a UDP header to a size of 2 to 4 bytes. Therefore, in advantageous conditions, a 48-byte uncompressed IPv6/UDP header may be compressed down to a 4- to 6-byte format (when using link-local addresses) or a 5- to 7-byte format (for global interactions) by using RFC 6282. Recently, a framework called Static Context Header Compression (SCHC) has been designed with the primary goal of supporting IPv6 over Low Power Wide Area Network (LPWAN) technologies . SCHC comprises header compression and fragmentation functionality tailored to the extraordinary constraints of LPWAN technologies, which are more severe than those exhibited by IEEE 802.15.4 or other relatively similar technologies. SCHC header compression allows a greater compression ratio than that of RFC 6282. If used properly, SCHC allows to compress an IPv6/UDP header down to e.g. a single byte. In addition, SCHC can be used to compress Constrained Application Protocol (CoAP) headers , which further increases the achievable performance improvement of using SCHC header compression, since there is no 6LoWPAN header compression mechanism defined for CoAP. Therefore, it may make sense to use SCHC header compression in some 6LoWPAN environments, including IEEE 802.15.4 networks, considering its greater efficiency. This document specifies how a SCHC-compressed packet can be carried over IEEE 802.15.4 networks. In order to ease a transition from existing 6LoWPAN/6Lo implementations to support SCHC header compression, the document also enables the transmission of SCHC-compressed UDP/CoAP headers over 6LoWPAN-compressed IPv6 packets. Further transition approaches are also described. The mechanism to be used to provide the SCHC header compression context to the nodes in an IEEE 802.15.4 network is out of the scope of this document. Note that, as per this document, and while SCHC defines fragmentation mechanisms as well, 6LoWPAN/6lo fragmentation is used when necessary to transport SCHC-compressed packets over IEEE 802.15.4 networks . This specification updates RFC 8138, RFC 8724, and RFC 9008.

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 BCP14 , , when, and only when, they appear in all capitals, as shown here.

The reader is expected to be familiar with the terms and concepts defined in specifications of 6LoWPAN frame formats , RPL and companion documents , 6LoWPAN Routing Header , SCHC , SCHC for CoAP , and SCHC architecture . RFC 8724 defines the Rule concept, whereby a Rule may be used to support header compression or fragmentation functionality. In the present document, Rules are only used for header compression. RFC 6775 defines the term 6LoWPAN Node (6LN) as the following: "A 6LoWPAN node is any host or router participating in a LoWPAN. This term is used when referring to situations in which either a host or router can play the role described." In this document, as in RFC 9008, 6LN acts as a leaf.

The traditional 6LoWPAN-based protocol stack for constrained devices (Figure 1, left) places the 6LoWPAN adaptation layer between IPv6 and an underlying technology such as IEEE 802.15.4. Suitable upper layer protocols include CoAP and UDP. (Note that, while CoAP has also been specified over TCP, and TCP may play a significant role in IoT environments , 6LoWPAN header compression has not been defined for TCP, as of the writing.) 6LoWPAN can be envisioned as a set of two main sublayers, where the upper one provides header compression, while the lower one offers fragmentation. This document defines an alternative approach for packet header compression over IEEE 802.15.4, which leads to a modified protocol stack (Figure 1, right). Fragmentation functionality remains the one defined by 6LoWPAN and 6lo .

SCHC header compression may be applied to the headers of different protocols or sets of protocols. Some examples include: i) IPv6 packet headers, ii) joint IPv6 and UDP packet headers, iii) joint IPv6, UDP and CoAP packet headers, etc.

In order to ease a transition from existing 6LoWPAN implementations to support SCHC header compression, the present document also: i) illustrates two possible protocol stacks, where 6LoWPAN header compression is used to compress IPv6/UDP headers while SCHC compresses CoAP headers (see Figure 2 and Section 5.1), and ii) enables the transmission of SCHC-compressed UDP/CoAP headers over 6LoWPAN-compressed IPv6 packets (see Figure 3 and Section 5.2). However, note that the greatest header compression performance can be achieved by using SCHC to also compress the UDP header. RFC 8824 defines how SCHC can be used to compress CoAP headers, including Object Security for Constrained RESTful Environments (OSCORE)-protected messages . On the other hand, it is possible to carry SCHC-compressed CoAP headers over UDP by means of using SCHC UDP ports . Figure 2 (left) shows the resulting protocol stack, where 6LoWPAN header compression is applied to UDP and IPv6. When Datagram Transport Layer Security (DTLS) is preferred to protect SCHC-compressed CoAP messages, the DTLS layer sits between the SCHC and UDP layers (Figure 2, right).

Finally, the "transition" protocol stack enabled by this document, which allows the transmission of 6LoWPAN-compressed IPv6 packets containing SCHC-compressed UDP/CoAP data units, is shown in Figure 3 (rightmost).

This section describes how SCHC architecture concepts (such as "SCHC Stratum", "Discriminator", "SCHC Header Instance", "SCHC Packet Instance", and "Set of Rules" (SoR)) [draft-ietf-schc-architecture] are applied when SCHC is used to compress IPv6 packet headers over IEEE 802.15.4 networks. In addition, the concepts of Single-instance networks and Multiple-instance networks are introduced.

When SCHC is used to compress IPv6 packets over IEEE 802.15.4 networks, the SCHC Stratum is located on top of layer 2 and below layer 3 (that is, at layer 2.5). Note that the compressed data of the SCHC Stratum may also comprise upper layer packet headers. For example, SCHC may be used to compress IP headers, IP/UDP headers or IP/UDP/CoAP headers (all at once). In both Single-instance and Multiple-instance networks, the Discriminator is a 6LoWPAN Dispatch Type set to the SCHC Dispatch or to the SCHC Pointer Dispatch (see Section 4).

In Single-instance networks, all network nodes that use SCHC for C/D have a single SCHC Packet Instance, and thus a single SoR for SCHC Packet C/D. For this reason, in Single-instance networks, the SCHC Header is fully compressed (i.e., the SCHC Header requires 0 bits to be transmitted over the air). In Single-instance networks, all network nodes that use SCHC for C/D have a single SCHC Header Instance, and therefore a single SoR for SCHC Header C/D, which in this case comprises a single, implicit Rule for SCHC Header C/D.

In Multiple-instance networks, at least some of the network nodes that use SCHC for C/D have more than one SCHC Packet Instances, and thus one SoR associated to each SCHC Packet Instance. Therefore, in Multiple-instance networks, the SCHC Header cannot be fully compressed (i.e., in compressed form, the SCHC Header requires more than 0 bits to be transmitted over the air). In Multiple-instance networks, all network nodes that use SCHC for C/D have a single SCHC Header Instance, and therefore a single SoR for SCHC Header C/D, which may comprise several Rules for SCHC Header C/D.

IEEE 802.15.4 supports two main network topologies: the star topology, and the peer-to-peer (i.e., mesh) topology. SCHC has been designed for LPWAN technologies, which are typically based on a star topology where constrained devices (e.g., sensors) communicate with a less constrained, central network gateway [RFC 8376]. However, as stated in [draft-ietf-schc-architecture], SCHC is generic and it can also be used in networking environments beyond the ones originally considered for SCHC. SCHC compression is applicable to both star topology and mesh topology IEEE 802.15.4 networks. The mechanism to be used to provide the SCHC header compression context to the nodes in an IEEE 802.15.4 network is out of the scope of this document.

In order to support the transmission of SCHC-compressed packets between two endpoints that are single-hop neighbors, both endpoints MUST store the Rules intended for the communication between those two endpoints. The frame format to be used to carry a SCHC-compressed packet in single-hop communication is described in Section 4.1.

6LoWPAN defines two approaches for multihop communication: Route-Over and Mesh-Under . In Route-Over, routing is performed at the IP layer. In Mesh-Under, routing functionality is located at the adaptation layer, below IP. This section describes how SCHC-compressed packets are transmitted over a multihop IEEE 802.15.4 network, for both Route-Over and Mesh-Under.

SCHC header compression MAY be used in a Route-Over network in a straightforward approach, whereby all routers (i.e., all 6LRs and 6LBRs) MUST store all the Rules in use by any nodes in the network, whereas a host MUST store the Rules defined for its communication with other endpoints. This approach is called Straightforward Route-Over (SRO). In this case, 6LoWPAN routers are able to decompress (if needed) received packet headers and compress packet headers before being forwarded. In SRO, in Single-instance networks, a RuleID and the Rule it identifies MUST be unique network-wide (note: the means to ensure so are out of the scope of this document). In this case, in order to simplify the management of RuleIDs in the network, in SRO, all nodes in the 6LoWPAN network MAY share the same SoR. In SRO, in Multiple-instance networks, a not fully compressed SCHC Header MUST be used. Figure 4 illustrates an example Single-instance network with the Rules that need to be stored by the nodes in SRO. In this example, RuleID 1 is intended for communication between Host A and Host B, RuleID 2 is intended for communication between Host A and Host C, and RuleID 3 is used for the communication between Host A and an external node called Host E.

Figure 5 illustrates an example Multiple-instance network with the Rules that need to be stored by the nodes in SRO. In this example, in addition to the Rules used in Figure 4, which correspond to a SCHC Packet Instance called I1 in this example, there is a second RuleID 2, which corresponds to communication between A and B, in a second SCHC Packet Instance (I2).

The frame format to be used to carry a SCHC-compressed packet in SRO is described in Section 4.1.

In a Route-Over network that uses the IPv6 Routing Protocol for Low-Power and Lossy Networks (RPL) [RFC6550], the RPL non-storing mode [RFC6550, RFC 6554] and [RFC8138] MAY be exploited in order to efficiently transmit SCHC-compressed packets. In this approach, packets sent by a 6LN are tunneled to the root, and packets intended for 6LNs are tunneled from the root (note: a tunnel is not needed when the root itself is the source). Traffic between two 6LNs traverses an Upward tunnel to the root and a Downward tunnel from the root. The present document defines the described approach as Tunneled, RPL-based Route-Over approach (TRO). In TRO, each 6LoWPAN node (i.e., a host, a 6LR or a 6LBR) MUST store the Rules defined for its communication with other endpoints. A 6LR is relieved to store Rules used by pairs of endpoints that do not include the 6LR itself. A 6LBR MUST store all the Rules used by all nodes in the network. In a TRO Single-instance network, a RuleID and the Rule it identifies MUST be unique network-wide (note: the means to ensure so are out of the scope of this document). In a TRO Multiple-instance network, a not fully compressed SCHC Header MUST be used. Figure 6 illustrates the Rules that need to be stored by the nodes in TRO, based on the same example Single-instance network and endpoint pairs shown in Figure 4.

Figure 7 illustrates an example Multiple-instance network with the Rules that need to be stored by the nodes in TRO. In this example, in addition to the Rules used in Figure 6, which correspond to a SCHC Packet Instance called I1 in this example, there is a second RuleID 2, which corresponds to communication between A and B, in a second SCHC Packet Instance (I2).

RFC 9008 describes how the communication between a 6LN and another endpoint (another 6LN or the root of the same RPL domain, or an external node, e.g., on the Internet) is performed. For the sake of description clarity, Figure 6 (adapted from Figure 3 in RFC 9008) provides a reference topology including nodes referred to in the remainder of this subsection.

In RPL non- storing mode, for Downward traffic, the root adds a source-routing header. The root also performs IPv6-in-IPv6 encapsulation, except when the root itself is the packet source. The IPv6-in-IPv6 encapsulation terminates at the 6LN (if it is a RAL, e.g., U, S or R) or at the last 6LR, e.g., V or X, (if the 6LN is a RUL, e.g., T or Q). For Upward traffic, IPv6-in-IPv6 encapsulation is performed by the first 6LR, e.g. V or X, when the 6LN is a RUL, e.g., T or Q, that sends a packet to an external node or to another 6LN in the same RPL domain, but not to the root. When the 6LN is a RAL (e.g., U, S or R) that sends packets to the same destinations, IPv6-in-IPv6 encapsulation may be performed (by the RAL itself). The destination in the outer header of the IPv6-in-IPv6 encapsulation for Upward traffic is the root. This document updates RFC 9008 by specifying that, in TRO, when a 6LN transmits an IPv6 packet whose header is compressed by means of SCHC instead of 6LoWPAN header compression (RFC 6282), the SCHC-compressed packet MUST be tunneled by means of IPv6-in-IPv6 encapsulation up to the root. This applies regardless of the inner, SCHC-compressed packet destination. For Upward traffic, when the 6LN is a RAL (e.g., U, S or R), the 6LN itself performs the IPv6-in-IPv6 encapsulation. However, if the 6LN is a RUL (e.g., T or Q), IPv6-in-IPv6 encapsulation is performed by the first 6LR (e.g., E or C, respectively). In the latter case, in order to enable efficient packet transmission in the first hop from the 6LN, the first 6LR SHOULD be provided with SCHC Rules allowing efficient header compression of packets sent by that 6LN. For Downward traffic, when the 6LN is a RUL (e.g., G or J), in order to enable efficient packet transmission in the last hop to the 6LN, the last 6LR (e.g., V or X, respectively) SHOULD be provided with SCHC Rules allowing efficient header compression of packets sent to that 6LN. Not providing such SCHC Rules to the first or last 6LR (for Upward or Downward traffic, respectively) should only happen if it is not practical or possible to do so (e.g., due to lack of available memory at the 6LR). For the sake of efficiency, RFC 8138 MUST be used to compress IPv6-in-IPv6 headers, the RPL Option (RFC 6553) and the source routing header (RPL Routing Header type 3, RFC 6554). The frame format to be used to carry a SCHC-compressed packet in TRO is described in Section 4.3.

In the previous approach, TRO, intermediate nodes do not have to know the IPv6 destination address of a SCHC-compressed IPv6 packet to be able to forward it. Another approach where intermediate nodes do not have to store the compression/decompression Rules used by the endpoints, which in addition does not require IPv6-in-IPv6 encapsulation, non-storing mode RPL and RFC 8138 compression, is called Pointer-based Route-Over (PRO). In PRO, a pointer (called "SCHC Pointer") is prepended to the SCHC-compressed packet, in order to indicate the location and length of the Hop Limit and the destination address residues in the SCHC-compressed header. Therefore, a 6LR is able to determine the IPv6 destination address of a SCHC-compressed packet, decrement its Hop Limit and route the packet, without the need to store the corresponding Rules. Note that, in PRO, each 6LoWPAN node (i.e., a host, a 6LR, or a 6LBR) MUST store the Rules defined for its communication as an endpoint with other endpoints. A 6LBR MUST store the Rules used by any network node for communication with external nodes. In a PRO Single-instance network, a RuleID MAY be used to identify different Rules used by different pairs of endpoints. In a PRO Multiple-instance network, a not fully compressed SCHC Header MUST be used. Figure 9 illustrates the Rules that are stored by the nodes in an example Single-instance network based using PRO. Note that, in this example, the network exploits the fact that PRO allows a given RuleID to be used by different pairs of endpoints.

PRO is compatible with RPL storing mode, as well as with other routing protocols. Figure 10 illustrates an example Multiple-instance network with the Rules that need to be stored by the nodes in PRO. In this example, in addition to the Rules used in Figure 9, which correspond to a SCHC Packet Instance called I1 in this example, there is an additional RuleID 2, which corresponds to communication between A and D, in a second SCHC Packet Instance (I2).

When SCHC header compression is used in a Mesh-Under network, Mesh-Under operates as described in RFC 4944. The frame format to be used to carry a SCHC-compressed packet in the Mesh-Under approach is described in Section 4.3. For header compression in a Mesh-Under network, a network node MUST store the Rules defined for its communication with other endpoints. In this case, a RuleID MAY be reused across disjoint pairs of endpoints, to identify different Rules used by such disjoint pairs of endpoints. In Mesh-Under, in a Single-instance network, a RuleID MAY be used to identify different Rules used by different pairs of endpoints. In a Mesh-Under Multiple-instance network, a fully compressed SCHC Header MAY be used as long as it is possible to determine the SCHC Packet Instance needed to decompress a SCHC-compressed packet based on the packet source identifier (which in Mesh-Under is present in the Mesh-Under header [RFC 4944]). Figure 11 illustrates the Rules that need to be stored by the nodes when SCHC is used for header compression in a Single-Instance Mesh-Under network, based on the same example network and endpoint pairs shown in Figure 9. Note that, in this example, the network exploits the fact that Mesh-under allows a given RuleID to be reused by different pairs of endpoints, even if the Rules sharing the same RuleID are different. Nodes denoted "m" in Figure 8 correspond to Mesh-Under forwarders [RFC 6606].

Figure 12 illustrates an example Multiple-instance network with the Rules that need to be stored by the nodes in PRO. In this example, in addition to the Rules used in Figure 9, which correspond to a SCHC Packet Instance called I1 in this example, there is an additional RuleID 2, which corresponds to communication between A and D, in a second SCHC Packet Instance (I2).

This section defines the frame formats that can be used when a SCHC-compressed packet is carried over IEEE 802.15.4. Such formats are carried as IEEE 802.15.4 frame payload.

This subsection defines the frame format for carrying SCHC-compressed packets over IEEE 802.15.4 for single-hop communication (see 3.3) or when SRO is used for multihop communication (see 3.4.1). This format comprises a SCHC Dispatch Type, a SCHC Header, a SCHC Packet (i.e. a SCHC-compressed packet (RFC 8724), and Padding bits, if any). illustrates the described frame format.

Adding SCHC header compression to the panoply of header compression mechanisms used in 6LoWPAN/6Lo environments creates the need to signal when a packet header has been compressed by using SCHC. To this end, the present document specifies the SCHC Dispatch. The SCHC Dispatch indicates that the next field in the frame format is a SCHC-compressed header (SCHC Header in , see 4.1.2)). This document defines the SCHC Dispatch as a 6LoWPAN Dispatch Type for SCHC header compression . With the aim to minimize overhead, the present document allocates a 1-byte pattern in Page 0 for the SCHC Dispatch Type: SCHC Dispatch Type bit pattern: 01000100 (Page 0) (Note: to be confirmed by IANA))

The SCHC Header ("SCHC Hdr" in Figure 9) determines the SCHC Packet Instance to be used to decompress the next field (SCHC-compressed header, see 4.1.3). As described in the SCHC architecture draft, the SCHC Header comprises a RuleID and a compression residue [draft-ietf-schc-architecture]. In Single-instance networks, the SCHC Header MUST be fully compressed, i.e., its size in compressed form is 0 bits. In Multiple-instance networks, the SCHC Header cannot be fully compressed; in this case, the RuleID size (of the Rule used to compress the SCHC Header) is RECOMMENDED to be between 1 and 8 bits.

The SCHC-compressed Header ("Cmprd Header" in ) corresponds to a packet header that has been compressed by using SCHC. As defined in , a SCHC-compressed header comprises a RuleID, and a compression residue. As per the present specification, a RuleID size between 1 and 16 bits is RECOMMENDED. In order to decide the RuleID size to be used in a network, the trade-off between (compressed) header overhead and the number of Rules needs to be carefully assessed.

If SCHC header compression leads to a SCHC Packet size of a non-integer number of bytes, padding bits of value equal to zero MUST be appended to the SCHC Packet as appropriate to align to an octet boundary.

This subsection defines the frame formats for carrying SCHC-compressed packets over IEEE 802.15.4 in TRO (see 3.3.2). Such formats are based on RFC 8138; however, instead of RFC 6282 header compression, this specification uses SCHC header compression. Accordingly, this specification updates RFC 8138 by stating that a 6LoRH header MUST always be placed before the LOWPAN_IPHC as defined in RFC 6282 [RFC6282] or the SCHC Dispatch, followed by the SCHC-compressed packet, as defined in the present specification. Since 6LoRH uses Dispatch Types in Page 1, the present specification also defines a SCHC Dispatch Type in Page 1, with the same bit pattern as the one in Page 0: 01000100 (to be confirmed by IANA). In the TRO frame formats, the SCHC Header is preceded by the SCHC Dispatch (in this case, in Page 1). The frame format for Downward transmission, except when the SCHC-compressed packet source is a RPL root, is shown in :

This subsection describes the frame format for carrying SCHC-compressed packets over IEEE 802.15.4 in PRO (see 3.3.3). Such format is shown in :

The first field in is defined as the SCHC Pointer Dispatch, which signals the start of a PRO Header format. This document defines the SCHC Pointer Dispatch as a 6LoWPAN Dispatch Type for SCHC header compression. With the aim to minimize header overhead, the present document allocates a 1-byte pattern in the 6LoWPAN Dispatch Type Page 0 for the SCHC Pointer Dispatch Type: SCHC Pointer Dispatch Type bit pattern: 01000101 (Page 0) (Note: to be confirmed by IANA)) The next field in the PRO Header is the Context IDentifier (CID) flag, which is set to 1 to signal that the Destination Context Identifier (DCI) field (see PRO_header_format) is present in the frame. When CID is set to 0, the DCI field is not present. The DCI field is optional. When present, it has a size of 4 bits. Similarly to RFC 6282, this field identifies the prefix of the IPv6 destination address. How such prefix context is distributed and maintained is out of the scope of the present document. The Bit pointer gives the starting position of the Hop Limit followed by the IPv6 destination address in the SCHC residue of the SCHC-compressed IPv6 header (in bits), starting after the Address Length field and before the first field of the SCHC-compressed IPv6 header (i.e., the RuleID). For example, if the Hop Limit and the IPv6 destination address residue are the only residues in a SCHC-compressed IPv6 packet header (i.e., such residue starts right after the RuleID in the SCHC-compressed header), then the Bit pointer will have a value of RuleID length in bits. The Hop Limit (HLM) flag is 1 bit that indicates the length of the Hop Limit field residue in the SCHC-compressed IPv6 header. When HLM equals 0, the Hop Limit compression residue has a size of 4 bits. In this case, the 4 most significant bits of the uncompressed Hop Limit field are equal to 0. Therefore, Hop Limit compression applies only to Hop Limit values between 15 and 0. When HLM is set to 1, the Hop Limit compression residue has a size of 8 bits (i.e., it is uncompressed). Address Length indicates the size of the IPv6 destination address residue (in bits). It can be up to 128 bits to allow representing the complete destination address, if needed. PRO requires a special SCHC Rule design where the FIDs of the IPv6 Destination and Source addresses are swapped (see 6.1.1).

This subsection describes the frame formats for carrying SCHC-compressed packets over IEEE 802.15.4 in the Mesh-Under approach (see 3.3.3). Note that the formats are provided in this section for the sake of clarity and completeness, since they are the same as those in RFC 4944, except for the fact that SCHC-compressed packets are carried. The frame format for a SCHC-compressed packet to be sent by means of Mesh-Under, when fragmentation is not needed, is shown in :

The different transmission alternatives enabled by the present document are shown in :

In order to enable the transition protocol stack, (i.e., supporting SCHC-compressed UDP/CoAP headers over 6LoWPAN-compressed IPv6 packets), the present document exploits the work that is being done by the INTAREA WG, to define a new Internet Protocol Number for SCHC . In this approach, the NH field of the RFC 6282-compressed IPv6 header format is set to 0. The Next Header field of the IPv6 header remains an 8-bit (uncompressed) field carrying the SCHC Internet Protocol Number. The resulting protocol encapsulation and corresponding format for an unfragmented packet, which is carried as IEEE 802.15.4 frame payload, is shown in . Padding is added as needed to align the format to an octet boundary.

+-----------------------+------------------+--------------+ - - + | RFC6282-compressed | | | | | IPv6 header | SCHC-compressed | CoAP Payload | Pad | |(NH=0,Next Header=SCHC)| UDP/CoAP headers | | | +-----------------------+------------------+--------------+- - -+ ]]> For networks using the transition protocol stack based on RPL routing, the formats defined in RFC 8138 may also be used for the sake of efficiency, as shown in . In this figure, the first field is the Page switch with value 1, followed by RFC 8138-compressed routing artifacts, then followed by the RFC 6282-compressed IPv6 header (which indicates that the next header data unit is a SCHC Packet).

IPv6 and UDP header fields MUST be compressed as per Section 10 of RFC 8724. IPv6 addresses are split into two 64-bit-long fields; one for the prefix and one for the Interface Identifier (IID). To allow for a single Rule being used for both directions, RFC 8724 identifies IPv6 addresses and UDP ports by their role (Dev or App) and not by their position in the header (source or destination). This optimization can be used as is in some IEEE 802.15.4 networks (e.g., an IEEE 802.15.4 star topology where the peripheral devices (Devs) send/receive packets to/from a network-side entity (App)). However, in some types of 6LoWPAN environments (e.g., when a sender and its destination are both peer nodes in a mesh topology network), additional functionality is needed to allow use of the Dev and App roles for C/D. In this case, each SCHC C/D entity needs to know its role (Dev or App) in addition to the Rule(s), and corresponding RuleIDs, for each endpoint it communicates with before such communication occurs . In such cases, the terms Uplink and Downlink that have been defined in RFC 8724 need to be understood in the context of each specific pair of endpoints. RFC 8724 (Section 7.1) states that "In a Rule, the Field Descriptors are listed in the order in which the fields appear in the packet header". The present specification updates RFC 8724 by stating that, in order to allow IPv6 header compression in PRO, the Field Descriptors of the IPv6 destination address (i.e., IPv6 DevPrefix and IPv6 DevIID) MUST appear before the Field Descriptors of the IPv6 source address (i.e., IPv6 AppPrefix and IPv6 AppIID), while the rest of fields appear in the same order as in the IPv6 packet header. In PRO, in order to support IPv6 header compression, one Rule MUST be defined for each direction between the two involved C/D endpoints. In such a Rule, the IPv6 DevPrefix and IPv6 DevIID FIDs MUST refer to the destination address (i.e., the destination endpoint takes the "Dev" role) of the SCHC-compressed IPv6 header. This allows a 6LR to read the compression residue of the Hop Limit and IPv6 destination address fields of the SCHC-compressed header by means of the Bit Pointer.

Compression of IPv6 source and destination prefixes MUST be performed as per Section 10.7.1 of RFC 8724. Additional guidance is given in the present section. Compression of IPv6 source and destination IIDs MUST be performed as per Section 10.7.2 of RFC 8724. One particular consideration when SCHC C/D is used in IEEE 802.15.4 networks is that, in contrast with some LPWAN technologies, IEEE 802.15.4 data frame headers include both source and destination fields. If the Dev or App IID are based on an L2 address, in some cases the IID can be reconstructed with information coming from the L2 header. Therefore, in those cases, DevIID and AppIID CDAs can be used. RFC 8724 states that "If the Rule is intended to compress packets with different prefix values, match-mapping SHOULD be used" (Section 10.7.1 of RFC 8724) and "If several IIDs are possible, then the TV contains the list of possible IIDs, the MO is set to "match-mapping" and the CDA is set to "mapping-sent"" (Section 10.7.2 of RFC 8724). However, the present specification updates RFC 8724 by stating that, in PRO, a source node MUST NOT use the match-mapping operator or the "mapping-sent" CDA to compress the IPv6 destination address prefix or the IPv6 destination IID, because 6LRs do not store SCHC context, and therefore do not have the match-mapping index meaning information.

RFC 8724 states that "a SCHC compressor MAY elide the UDP checksum when another layer guarantees at least equal integrity protection for the UDP payload and the pseudo-header". IEEE 802.15.4 frames carry a 16-bit Frame Check Sequence (FCS), which is computed by means of a 16-bit ITU-T CRC algorithm. Considering the FCS size, the greater error detection capabilities of CRC compared with checksum, and the fact that the IEEE 802.15.4 FCS will be checked at each hop in an IEEE 802.15.4 multihop network, the UDP checksum MUST be elided when using SCHC to compress UDP headers.

CoAP header fields MUST be compressed as per Sections 4 to 6 of RFC 8824. Additional guidance is given in this section. For CoAP header compression/decompression, the SCHC Rules description uses direction information in order to reduce the number of Rules needed to compress headers. As stated in 5.1, in some types of 6LoWPAN environments (e.g., when a sender and its destination are both peer nodes in a mesh topology network), each SCHC C/D entity needs to know its role (Dev or App), in addition to the Rule(s), and corresponding RuleIDs, for each endpoint it communicates with before such communication occurs . Therefore, in such cases, direction information will be specific to each pair of endpoints.

A number of optimizations have been developed in order to efficiently support IPv6 Neighbor Discovery (ND) in 6LoWPAN environments (6LoWPAN ND) [RFC 6775][RFC 8505]. SCHC can also be used to compress 6LoWPAN ND packets. At the time of this writing, compression of ICMPv6 headers is being specified in the SCHC WG [draft-barthel-schc-oam-schc-03]. Thus, it will be possible to compress the IPv6 header and the ICMPv6 headers of a packet carrying a 6LoWPAN ND message.

After applying SCHC header compression to a packet intended for transmission, if the size of the resulting SCHC Packet (Section 4) exceeds the IEEE 802.15.4 frame payload space available, such SCHC Packet MUST be fragmented, carried and reassembled by means of the fragmentation and reassembly functionality defined by 6LoWPAN or 6Lo . In a Route-Over multihop network, the 6LoWPAN fragment forwarding technique called Virtual Reassembly Buffer (VRB) SHOULD be used. However, VRB might not be the best approach for a particular network, e.g., if at least one of the caveats described in Section 6 of RFC 8930 is unacceptable or cannot be addressed.

This document requests the allocation of the 6LoWPAN Dispatch Type Field Bit Patterns, on the Pages and with the Header Types shown next:

This document does not define SCHC header compression functionality beyond the one defined in RFC 8724. Therefore, the security considerations in section 12.1 of RFC 8724 and in section 9 of RFC 8824 apply. As a safety measure, a SCHC decompressor implementing the present specification MUST NOT reconstruct a packet larger than 1500 bytes . IEEE 802.15.4 networks support link-layer security mechanisms such as encryption and authentication. As in RFC 8824, the use of a cryptographic integrity-protection mechanism to protect the SCHC-compressed headers is REQUIRED. The addition of the pointer used in PRO creates new attack opportunities. A malicious node might be able to modify the related fields (i.e., Bit Pointer or Address Length) to avoid a router to correctly reconstruct the IPv6 destination field of a SCHC-compressed IPv6 packet, thus avoiding delivery of the packet to its intended destination. Appropriate use of link-layer security should significantly reduce the probability of the described threat.

Ana Minaburo and Laurent Toutain suggested for the first time the use of SCHC in environments where 6LoWPAN has traditionally been used. Flavien Moullec is a contributor to this document. Laurent Toutain, Pascal Thubert, Dominique Barthel, Guangpeng Li, Carsten Bormann, Nathan Lecorchet, Stuart Cheshire, Kiran Makhijani, and Georgios Z. Papadopoulos made comments that helped shape this document. Carles Gomez has been funded in part by the Spanish Government through project PID2019-106808RA-I00, and by Secretaria d'Universitats i Recerca del Departament d'Empresa i Coneixement de la Generalitat de Catalunya 2017 through grant SGR 376 and 2021 throught grant SGR 00330.

Uplink packet

In this example, for SCHC compression of IPv6/UDP headers, RuleID 0x20 is used. The Rule corresponding to RuleID 0x20 is shown in .

TO-DO

Uplink packet

In this example, for SCHC compression of UDP/CoAP headers, RuleID 0x22 is used. The Rule corresponding to RuleID 0x22 is shown in .

This section provides an analysis of the features, pros and cons of the route-over multihop approaches defined in this document: i) SRO, ii) TRO, and iii) PRO. TO-DO: align with latest descriptions of SRO, TRO and PRO.

SRO incurs the lowest header overhead among the considered Route-Over approaches, as it only requires the SCHC Dispatch (1 byte). However, it is the most demanding approach in terms of memory usage, since all network nodes (including intermediate nodes) need to store all the Rules in use in the network. Therefore, it will be suitable for rather small networks and/or where nodes have sufficient memory. Also, SCHC context should be ideally and actually be as static as possible, in order to avoid frequent network- wide stored SCHC context updates.

TRO incurs a header overhead that includes a fixed part (a Page Switch plus the SCHC Dispatch, of 1 byte each), plus a variable part that comprises RFC 8138-compressed routing artifacts. Regarding the latter, in a Downward transmission, it would include the SRH-6LoRH (of variable size, of 4 bytes in the best case, or e.g., 8 bytes as in Fig. 20 of RFC 8138), the RPI-6LoRH (3 bytes in the best case) and the IP-in-IP header (not present if the source is the Root, at least 3 bytes otherwise). In the cases considered, and when the Root is not the packet source, the total header overhead of this approach would be of at least 12-16 bytes. For upward transmission, the variable part of the header overhead for this approach would include only the RPI-6LoRH (at least, 3 bytes) and the IP-in-IP header (at least, 3 bytes). Therefore, in the cases considered, the total header overhead of this approach would be of at least 8 bytes. Note that, while the overhead of TRO may appear to be relatively high, tunnel-based structures like the one assumed in TRO may exist already in a network deployment. Therefore, in such cases, the additional overhead of TRO may be actually lower. An advantage of this approach is that a node only has to store the Rules for the communications it is involved in as an endpoint, which minimizes memory requirements and the impact of potential SCHC context updates. For example, pure intermediate nodes do not have to store SCHC context. Note that this approach requires the network to use RPL, non-storing mode. Furthermore, the paths for communication between two nodes in the same network or with external nodes will need to traverse the Root. For communication with external nodes, traversing the Root will be needed anyway, therefore this feature does not pose any issue. However, this constraint will preclude the usage of optimal routes in some cases.

PRO incurs the PRO header overhead (i.e., between 3 and 3.5 bytes). In addition, with PRO, the Hop Limit field will have to be carried fully inline (1 byte) or compressed down to a minimum size of 4 bits. Furthermore, PRO introduces a limit to the achievable IPv6 destination address compression performance, as described next (note that the size of the destination address compression residue will depend on and will need to be planned for the intended use case of the network): A.- In special cases (e.g., if there is only one possible destination that is known beforehand), there will not be a destination address residue. B.- For a given destination prefix known by the network nodes (e.g., when prefix contexts are used, or if there can only be one destination prefix), if there can be several possible destinations in that network, the destination address residue will be up to 8 bytes (it could be less depending on how the addresses in that network are built, for example, it could be just 2 bytes). C.- For destination prefixes not covered by prefix contexts or a priori knowledge by the nodes, the destination address residue will have to be the whole address (16 bytes), since an intermediate node does not know which is the destination prefix. An advantage of PRO, as in TRO, is that a node only has to store the Rules for the communications it is involved in as an endpoint, which minimizes memory requirements and the impact of potential SCHC context updates. For example, pure intermediate nodes do not have to store SCHC context. A potential advantage of PRO is that, in contrast with TRO, paths for intranetwork communication are not necessarily constrained to traversing a root node. Therefore, for intranetwork communication, the chances of using optimal paths are greater. Another feature is that the routing solution to be used is not tied to RPL non-storing mode.

Assessing the suitability of the different approaches requires considering the following dimensions: network size, node memory capabilities, header overhead, routing constraints / path optimality, intra- or inter-network communication. TO-DO: to be completed.