Network Working Group G. Pelletier
Request for Comments: 5225 K. Sandlund
Category: Standards Track Ericsson
April 2008 RObust Header Compression Version 2 (ROHCv2):
Profiles for RTP, UDP, IP, ESP and UDP-Lite
Status of This Memo
This document specifies an Internet standards track protocol for the
Internet community, and requests discussion and suggestions for
improvements. Please refer to the current edition of the "Internet
Official Protocol Standards" (STD 1) for the standardization state
and status of this protocol. Distribution of this memo is unlimited.
This document specifies ROHC (Robust Header Compression) profiles
that efficiently compress RTP/UDP/IP (Real-Time Transport Protocol,
User Datagram Protocol, Internet Protocol), RTP/UDP-Lite/IP (User
Datagram Protocol Lite), UDP/IP, UDP-Lite/IP, IP and ESP/IP
(Encapsulating Security Payload) headers.
This specification defines a second version of the profiles found in
RFC 3095, RFC 3843 and RFC 4019; it supersedes their definition, but
does not obsolete them.
The ROHCv2 profiles introduce a number of simplifications to the
rules and algorithms that govern the behavior of the compression
endpoints. It also defines robustness mechanisms that may be used by
a compressor implementation to increase the probability of
decompression success when packets can be lost and/or reordered on
the ROHC channel. Finally, the ROHCv2 profiles define their own
specific set of header formats, using the ROHC formal notation.
The ROHC WG has developed a header compression framework on top of
which various profiles can be defined for different protocol sets or
compression requirements. The ROHC framework was first documented in
[RFC3095], together with profiles for compression of RTP/UDP/IP
(Real-Time Transport Protocol, User Datagram Protocol, Internet
Protocol), UDP/IP, IP and ESP/IP (Encapsulating Security Payload)
headers. Additional profiles for compression of IP headers [RFC3843]
and UDP-Lite (User Datagram Protocol Lite) headers [RFC4019] were
later specified to complete the initial set of ROHC profiles.
This document defines an updated version for each of the above
mentioned profiles, and the definitions depend on the ROHC framework
as found in [RFC4995]. The framework is required reading to
understand the profile definitions, rules, and their role.
Specifically, this document defines header compression schemes for:
o RTP/UDP/IP : profile 0x0101
o UDP/IP : profile 0x0102
o ESP/IP : profile 0x0103
o IP : profile 0x0104
o RTP/UDP-Lite/IP : profile 0x0107
o UDP-Lite/IP : profile 0x0108
Each of the profiles above can compress the following type of
o AH [RFC4302]
o GRE [RFC2784][RFC2890]
o MINE [RFC2004]
o IPv6 Destination Options header[RFC2460]
o IPv6 Hop-by-hop Options header[RFC2460]
o IPv6 Routing header [RFC2460]
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
"SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
document are to be interpreted as described in RFC 2119 [RFC2119].
This document is consistent with the terminology found in the ROHC
framework [RFC4995] and in the formal notation for ROHC [RFC4997].
In addition, this document uses or defines the following terms:
The Acknowledgment Number identifies what packet is being
acknowledged in the RoHCv2 feedback element (See Section 6.9).
The value of this field normally corresponds to the Master
Sequence Number (MSN) of the header that was last successfully
decompressed, for the compression context (CID) for which the
feedback information applies.
Chaining of Items
A chain of items groups fields based on similar characteristics.
ROHCv2 defines chain items for static, dynamic and irregular
fields. Chaining is achieved by appending an item to the chain
for each header in its order of appearance in the uncompressed
packet. Chaining is useful to construct compressed headers from
an arbitrary number of any of the protocol headers for which a
ROHCv2 profile defines a compressed format.
CRC-3 Control Fields Validation
The CRC-3 control fields validation refers to the validation of
the control fields. This validation is performed by the
decompressor when it receives a Compressed (CO) header that
contains a 3-bit Cyclic Redundancy Check (CRC) calculated over
control fields. This 3-bit CRC covers controls fields carried in
the CO header as well as specific control fields in the context.
In the formal definition of the header formats, this 3-bit CRC is
labeled "control_crc3" and uses the control_crc3_encoding (See
also Section 6.6.11).
The delta refers to the difference in the absolute value of a
field between two consecutive packets being processed by the same
The number of packets by which a packet is received late within
its sequence due to reordering between the compressor and the
decompressor, i.e., reordering between packets associated with the
same context (CID). See the definition of sequentially late
ROHCv2 Header Types
ROHCv2 profiles use two different header types: the Initialization
and Refresh (IR) header type, and the Compressed (CO) header type.
Sequentially Early Packet
A packet that reaches the decompressor before one or several
packets that were delayed over the channel, where all of the said
packets belong to the same header-compressed flow and are
associated to the same compression context (CID). At the time of
the arrival of a sequentially early packet, the packet(s) delayed
on the link cannot be differentiated from lost packet(s).
Sequentially Late Packet
A packet is late within its sequence if it reaches the
decompressor after one or several other packets belonging to the
same CID have been received, although the sequentially late packet
was sent from the compressor before the other packet(s). How the
decompressor detects a sequentially late packet is outside the
scope of this specification, but it can for example use the MSN
for this purpose.
Timestamp Stride (ts_stride)
The timestamp stride (ts_stride) is the expected increase in the
timestamp value between two RTP packets with consecutive sequence
numbers. For example, for a media encoding with a sample rate of
8 kHz producing one frame every 20 ms, the RTP timestamp will
typically increase by n * 160 (= 8000 * 0.02), for some integer n.
Time Stride (time_stride)
The time stride (time_stride) is the time interval equivalent to
one ts_stride, e.g., 20 ms in the example for the RTP timestamp
This section lists most acronyms used for reference, in addition to
those defined in [RFC4995].
AH Authentication Header.
ESP Encapsulating Security Payload.
GRE Generic Routing Encapsulation.
FC Full Context state (decompressor).
IP Internet Protocol.
LSB Least Significant Bits.
MINE Minimal Encapsulation in IP.
MSB Most Significant Bits.
MSN Master Sequence Number.
NC No Context state (decompressor).
OA Optimistic Approach.
RC Repair Context state (decompressor).
ROHC Header compression framework (RFC 4995).
ROHCv2 Set of header compression profiles defined in this document.
RTP Real-time Transport Protocol.
SSRC Synchronization source. Field in RTP header.
CSRC Contributing source. The RTP header contains an optional
list of contributing sources.
TC Traffic Class. Field in the IPv6 header. See also TOS.
TOS Type Of Service. Field in the IPv4 header. See also TC.
TS RTP Timestamp.
TTL Time to Live. Field in the IPv4 header.
UDP User Datagram Protocol.
UDP-Lite User Datagram Protocol Lite.
4. Background (Informative)
This section provides background information on the compression
profiles defined in this document. The fundamentals of general
header compression and of the ROHC framework may be found in sections
3 and 4 of [RFC4995], respectively. The fundamentals of the formal
notation for ROHC are defined in [RFC4997]. [RFC4224] describes the
impacts of out-of-order delivery on profiles based on [RFC3095].
4.1. Classification of Header Fields
Section 3.1 of [RFC4995] explains that header compression is possible
due to the fact that there is much redundancy between field values
within the headers of a packet, especially between the headers of
Appendix A lists and classifies in detail all the header fields
relevant to this document. The appendix concludes with
recommendations on how the various fields should be handled by header
The main conclusion is that most of the header fields can easily be
compressed away since they never or seldom change. A small number of
fields however need more sophisticated mechanisms.
These fields are:
- IPv4 Identification (16 bits) - IP-ID
- ESP Sequence Number (32 bits) - ESP SN
- UDP Checksum (16 bits) - Checksum
- UDP-Lite Checksum (16 bits) - Checksum
- UDP-Lite Checksum Coverage (16 bits) - CCov
- RTP Marker ( 1 bit ) - M-bit
- RTP Sequence Number (16 bits) - RTP SN
- RTP Timestamp (32 bits) - TS
In particular, for RTP, the analysis in Appendix A reveals that the
values of the RTP Timestamp (TS) field usually have a strong
correlation to the RTP Sequence Number (SN), which increments by one
for each packet emitted by an RTP source. The RTP M-bit is expected
to have the same value most of the time, but it needs to be
communicated explicitly on occasion.
For UDP, the Checksum field cannot be inferred or recalculated at the
receiving end without violating its end-to-end properties, and is
thus sent as-is when enabled (mandatory with IPv6). The same applies
to the UDP-Lite Checksum (mandatory with both IPv4 and IPv6), while
the UDP-Lite Checksum Coverage may in some cases be compressible.
For IPv4, a similar correlation as that of the RTP TS to the RTP SN
is often observed between the Identifier field (IP-ID) and the master
sequence number (MSN) used for compression (e.g., the RTP SN when
compressing RTP headers).
4.2. Improvements of ROHCv2 over RFC 3095 Profiles
The ROHCv2 profiles can achieve compression efficiency and robustness
that are both at least equivalent to RFC 3095 profiles [RFC3095],
when used under the same operating conditions. In particular, the
size and bit layout of the smallest compressed header (i.e., PT-0
format U/O-0 in RFC 3095, and pt_0_crc3 in ROHCv2) are identical.
There are a number of differences and improvements between profiles
defined in this document and their earlier version defined in RFC
3095. This section provides an overview of some of the most
Tolerance to reordering
Profiles defined in RFC 3095 require that the channel between
compressor and decompressor provide in-order delivery between
compression endpoints. ROHCv2 profiles, however, can handle
robustly and efficiently a limited amount of reordering after the
compression point as part of the compression algorithm itself. In
addition, this improved support for reordering makes it possible
for ROHCv2 profiles to handle prelink reordering more efficiently.
Profiles in RFC 3095 define multiple operational modes, each with
different updating logic and compressed header formats. ROHCv2
profiles operate in unidirectional operation until feedback is
first received for a context (CID), at which point bidirectional
operation is used; the formats are independent of what operational
logic is used.
IP extension header
Profiles in RFC 3095 compress IP Extension headers using list
compression. ROHCv2 profiles instead treat extension headers in
the same manner as other protocol headers, i.e., using the
chaining mechanism; it thus assumes that extension headers are not
added or removed during the lifetime of a context (CID), otherwise
compression has to be restarted for this flow.
Profiles in RFC 3095 can compress at most two levels of IP
headers. ROHCv2 profiles can compress an arbitrary number of IP
ROHCv2 profiles do not support reference-based list compression.
Robustness and repairs
ROHCv2 profiles do not define a format for the IR-DYN packet;
instead, each profile defines a compressed header that can be used
to perform a more robust context repair using a 7-bit CRC
verification. This also implies that only the IR header can
change the association between a CID and the profile it uses.
ROHCv2 profiles mandate a CRC in the format of the FEEDBACK-2,
while this is optional in RFC 3095. A different set of feedback
options is also used in ROHCv2 compared to RFC 3095.
4.3. Operational Characteristics of ROHCv2 Profiles
Robust header compression can be used over different link
technologies. Section 4.4 of [RFC4995] lists the operational
characteristics of the ROHC channel. The ROHCv2 profiles address a
wide range of applications, and this section summarizes some of the
operational characteristics that are specific to these profiles.
ROHCv2 profiles assume that the lower layer indicates the length
of a compressed packet. ROHCv2 compressed headers do not contain
length information for the payload.
Out-of-order delivery between compression endpoints
The definition of the ROHCv2 profiles places no strict requirement
on the delivery sequence between the compression endpoints, i.e.,
packets may be received in a different order than the compressor
has sent them and still have a fair probability of being
However, frequent out-of-order delivery and/or significant
reordering depth will negatively impact the compression
efficiency. More specifically, if the compressor can operate
using a proper estimate of the reordering characteristics of the
path between the compression endpoints, larger headers can be sent
more often to increase the robustness against decompression
failures due to out-of-order delivery. Otherwise, the compression
efficiency will be impaired from an increase in the frequency of
decompression failures and recovery attempts.
5. Overview of the ROHCv2 Profiles (Informative)
This section provides an overview of concepts that are important and
useful to the ROHCv2 profiles. These concepts may be used as
guidelines for implementations but they are not part of the normative
definition of the profiles, as these concepts relate to the
compression efficiency of the protocol without impacting the
interoperability characteristics of an implementation.
5.1. Compressor Concepts
Header compression can be conceptually characterized as the
interaction of a compressor with a decompressor state machine, one
per context. The responsibility of the compressor is to convey the
information needed to successfully decompress a packet, based on a
certain confidence regarding the state of the decompressor context.
This confidence is obtained from the frequency and the type of
information the compressor sends when updating the decompressor
context from the optimistic approach (Section 5.1.1), and optionally
from feedback messages (See Section 6.9), received from the
5.1.1. Optimistic Approach
A compressor always uses the optimistic approach when it performs
context updates. The compressor normally repeats the same type of
update until it is fairly confident that the decompressor has
successfully received the information. If the decompressor
successfully receives any of the headers containing this update, the
state will be available for the decompressor to process smaller
If field X in the uncompressed header changes value, the compressor
uses a header type that contains an encoding of field X until it has
gained confidence that the decompressor has received at least one
packet containing the new value for X. The compressor normally
selects a compressed format with the smallest header that can convey
the changes needed to achieve confidence.
The number of repetitions that is needed to obtain this confidence is
normally related to the packet loss and out-of-order delivery
characteristics of the link where header compression is used; it is
thus not defined in this document. It is outside the scope of this
specification and is left to implementors to decide.
5.1.2. Tradeoff between Robustness to Losses and to Reordering
The ability of a header compression algorithm to handle sequentially
late packets is mainly limited by two factors: the interpretation
interval offset of the sliding window used for lsb encoded fields
[RFC4997], and the optimistic approach (See Section 5.1.1) for seldom
lsb encoded fields:
The interpretation interval offset specifies an upper limit for
the maximum reordering depth, by which is it possible for the
decompressor to recover the original value of a dynamically
changing (i.e., sequentially incrementing) field that is encoded
using a window-based lsb encoding. Its value is typically bound
to the number of lsb compressed bits in the compressed header
format, and thus grows with the number of bits transmitted.
However, the offset and the lsb encoding only provide robustness
for the field that it compresses, and (implicitly) for other
sequentially changing fields that are derived from that field.
This is shown in the figure below:
<--- interpretation interval (size is 2^k) ---->
v_ref-p v_ref v_ref + (2^k-1) - p
<--- reordering --> <--------- losses --------->
where p is the maximum negative delta, corresponding to the
maximum reordering depth for which the lsb encoding can recover
the original value of the field;
where (2^k-1) - p is the maximum positive delta, corresponding
to the maximum number of consecutive losses for which the lsb
encoding can recover the original value of the field;
where v_ref is the reference value, as defined in the lsb
encoding method in [RFC4997].
There is thus a tradeoff between the robustness against reordering
and the robustness against packet losses, with respect to the
number of MSN bits needed and the distribution of the
interpretation interval between negative and positive deltas in
Seldom changing fields
The optimistic approach (Section 5.1.1) provides the upper limit
for the maximum reordering depth for seldom changing fields.
There is thus a tradeoff between compression efficiency and
robustness. When only information on the MSN needs to be conveyed to
the decompressor, the tradeoff relates to the number of compressed
MSN bits in the compressed header format. Otherwise, the tradeoff
relates to the implementation of the optimistic approach.
In particular, compressor implementations should adjust their
optimistic approach strategy to match both packet loss and reordering
characteristics of the link over which header compression is applied.
For example, the number of repetitions for each update of a non-lsb
encoded field can be increased. The compressor can ensure that each
update is repeated until it is reasonably confident that at least one
packet containing the change has reached the decompressor before the
first packet sent after this sequence.
5.1.3. Interactions with the Decompressor Context
The compressor normally starts compression with the initial
assumption that the decompressor has no useful information to process
the new flow, and sends Initialization and Refresh (IR) packets.
Initially, when sending the first IR packet for a compressed flow,
the compressor does not expect to receive feedback for that flow,
until such feedback is first received. At this point, the compressor
may then assume that the decompressor will continue to send feedback
in order to repair its context when necessary. The former is
referred to as unidirectional operation, while the latter is called
The compressor can then adjust the compression level (i.e., what
header format it selects) based on its confidence that the
decompressor has the necessary information to successfully process
the compressed headers that it selects.
In other words, the responsibilities of the compressor are to ensure
that the decompressor operates with state information that is
sufficient to successfully decompress the type of compressed
header(s) it receives, and to allow the decompressor to successfully
recover that state information as soon as possible otherwise. The
compressor therefore selects the type of compressed header based on
the following factors:
o the outcome of the encoding method applied to each field;
o the optimistic approach, with respect to the characteristics of
o the type of operation (unidirectional or bidirectional), and if in
bidirectional operation, feedback received from the decompressor
(ACKs, NACKs, STATIC-NACK, and options).
Encoding methods normally use previous value(s) from a history of
packets whose headers it has previously compressed. The optimistic
approach is meant to ensure that at least one compressed header
containing the information to update the state for a field is
received. Finally, feedback indicates what actions the decompressor
has taken with respect to its assumptions regarding the validity of
its context (Section 5.2.2); it indicates what type of compressed
header the decompressor can or cannot decompress.
The decompressor has the means to detect decompression failures for
any compressed (CO) header format, using the CRC verification.
Depending on the frequency and/or on the type of the failure, it
might send a negative acknowledgement (NACK) or an explicit request
for a complete context update (STATIC-NACK). However, the
decompressor does not have the means to identify the cause of the
failure, and in particular the decompression of what field(s) is
responsible for the failure. The compressor is thus always
responsible for determining the most suitable response to a negative
acknowledgement, using the confidence it has in the state of the
decompressor context, when selecting the type of compressed header it
will use when compressing a header.
5.2. Decompressor Concepts
The decompressor normally uses the last received and successfully
validated (IR packets) or verified (CO packets) header as the
reference for future decompression.
The decompressor is responsible for verifying the outcome of every
decompression attempt, to update its context when successful, and
finally to request context repairs by making coherent usage of
feedback once it has started using feedback.
Specifically, the outcome of every decompression attempt is verified
using the CRC present in the compressed header; the decompressor
updates the context information when this outcome is successfully
verified; finally, if the decompressor uses feedback once for a
compressed flow, then it will continue to do so for as long as the
corresponding context is associated with the same profile.
5.2.1. Decompressor State Machine
The decompressor operation may be represented as a state machine
defining three states: No Context (NC), Repair Context (RC), and Full
The decompressor starts without a valid context, the NC state. Upon
receiving an IR packet, the decompressor validates the integrity of
its header using the CRC-8 validation. If the IR header is
successfully validated, the decompressor updates the context and uses
this header as the reference header, and moves to the FC state. Once
the decompressor state machine has entered the FC state, it does not
normally leave; only repeated decompression failures will force the
decompressor to transit downwards to a lower state. When context
damage is detected, the decompressor moves to the repair context (RC)
state, where it stays until it successfully verifies a decompression
attempt for a compressed header with a 7-bit CRC or until it
successfully validates an IR header. When static context damage is
detected, the decompressor moves back to the NC state.
Below is the state machine for the decompressor. Details of the
transitions between states and decompression logic are given in the
sub-sections following the figure.
| CRC-8(IR) |
| !CRC-8(IR) or CRC-7(CO) or or CRC-7(CO) |
| PT not allowed CRC-8(IR) or CRC-3(CO) |
| +--->---+ +--->----->----->----->---+ +--->---->---+ |
| | | | | | | |
| | v | v | v v
+-----------------+ +----------------------+ +--------------------+
| No Context (NC) | | Repair Context (RC) | | Full Context (FC) |
+-----------------+ +----------------------+ +--------------------+
^ ^ Static Context | ^ !CRC-7(CO) or | ^ Context Damage | |
| | Damage Detected | | PT not allowed | | Detected | |
| +--<-----<-----<--+ +----<------<----+ +--<-----<-----<--+ |
| Static Context Damage Detected |
CRC-8(IR) : Successful CRC-8 validation for the IR header.
!CRC-8(IR) : Unsuccessful CRC-8 validation for the IR header.
CRC-3(CO) : Successful CRC verification for the decompression
of a CO header, based on the number of CRC bits
carried in the CO header.
!CRC-7(CO) : Failure to CRC verify the decompression of a CO
header carrying a 7-bit CRC.
PT not allowed : The decompressor has received a packet type (PT)
for which the decompressor's current context does
not provide enough valid state information to
decompress the packet.
Static Context Damage Detected: See definition in Section 5.2.2.
Context Damage Detected: See definition in Section 5.2.2.
126.96.36.199. No Context (NC) State
Initially, while working in the No Context (NC) state, the
decompressor has not yet successfully validated an IR header.
In the NC state, only packets carrying sufficient information on
the static fields (i.e., IR packets) can be decompressed.
The decompressor can move to the Full Context (FC) state when the
CRC validation of an 8-bit CRC in an IR header is successful.
In the NC state, the decompressor should send a STATIC-NACK if a
packet of a type other than IR is received, or if an IR header has
failed the CRC-8 validation, subject to the feedback rate
limitation as described in Section 5.2.3.
188.8.131.52. Repair Context (RC) State
In the Repair Context (RC) state, the decompressor has successfully
decompressed packets for this context, but does not have confidence
that the entire context is valid.
In the RC state, only headers covered by an 8-bit CRC (i.e., IR)
or CO headers carrying a 7-bit CRC can be decompressed.
The decompressor can move to the Full Context (FC) state when the
CRC verification succeeds for a CO header carrying a 7-bit CRC or
when validation of an 8-bit CRC in an IR header succeeds.
The decompressor moves back to the NC state if it assumes static
In the RC state, the decompressor should send a STATIC-NACK when
CRC-8 validation of an IR header fails, or when a CO header
carrying a 7-bit CRC fails and static context damage is assumed,
subject to the feedback rate limitation as described in
Section 5.2.3. If any other packet type is received, the
decompressor should treat it as a CRC verification failure to
determine if NACK is to be sent.
184.108.40.206. Full Context (FC) State
In the Full Context (FC) state, the decompressor assumes that its
entire context is valid.
In the FC state, decompression can be attempted regardless of the
type of packet received.
The decompressor moves back to the RC state if it assumes context
damage. If the decompressor assumes static context damage, it
moves directly to the NC state.
In the FC state, the decompressor should send a NACK when CRC-8
validation or CRC verification of any header type fails and if
context damage is assumed, or it should send a STATIC-NACK if
static context damage is assumed; this is subject to the feedback
rate limitation described in Section 220.127.116.11.2.2. Decompressor Context Management
All header formats carry a CRC and are context updating. A packet
for which the CRC succeeds updates the reference values of all header
fields, either explicitly (from the information about a field carried
within the compressed header) or implicitly (fields inferred from
The decompressor may assume that some or the entire context is
invalid, when it fails to validate or to verify one or more headers
using the CRC. Because the decompressor cannot know the exact
reason(s) for a CRC failure or what field caused it, the validity of
the context hence does not refer to what specific part(s) of the
context is deemed valid or not.
Validity of the context rather relates to the detection of a problem
with the context. The decompressor first assumes that the type of
information that most likely caused the failure(s) is the state that
normally changes for each packet, i.e., context damage of the dynamic
part of the context. Upon repeated decompression failures and
unsuccessful repairs, the decompressor then assumes that the entire
context, including the static part, needs to be repaired, i.e.,
static context damage. Failure to validate the 3-bit CRC that
protects control fields should be treated as a decompression failure
when the decompressor asserts the validity of its context.
Context Damage Detection
The assumption of context damage means that the decompressor will
not attempt decompression of a CO header that carries only a 3-bit
CRC, and will only attempt decompression of IR headers or CO
headers protected by a CRC-7.
Static Context Damage Detection
The assumption of static context damage means that the
decompressor refrains from attempting decompression of any type of
header other than the IR header.
How these assumptions are made, i.e., how context damage is detected,
is open to implementations. It can be based on the residual error
rate, where a low error rate makes the decompressor assume damage
more often than on a high rate link.
The decompressor implements these assumptions by selecting the type
of compressed header for which it will attempt decompression. In
other words, validity of the context refers to the ability of a
decompressor to attempt (or not) decompression of specific packet
When ROHCv2 profiles are used over a channel that cannot guarantee
in-order delivery, the decompressor may refrain from updating its
context with the content of a sequentially late packet that is
successfully decompressed. This is to avoid updating the context
with information that is older than what the decompressor already has
in its context.
5.2.3. Feedback Logic
ROHCv2 profiles may be used in environments with or without feedback
capabilities from decompressor to compressor. ROHCv2 however assumes
that if a ROHC feedback channel is available and if this channel is
used at least once by the decompressor for a specific context, this
channel will be used during the entire compression operation for that
context (i.e., bidirectional operation).
The ROHC framework defines 3 types of feedback messages: ACKs, NACKs,
and STATIC-NACKs. The semantics of each message is defined in
Section 18.104.22.168. of [RFC4995]. What feedback to send is coupled with
the context management of the decompressor, i.e., with the
implementation of the context damage detection algorithms as
described in Section 5.2.2.
The decompressor should send a NACK when it assumes context damage,
and it should send a STATIC-NACK when it assumes static context
damage. The decompressor is not strictly expected to send ACK
feedback upon successful decompression, other than for the purpose of
improving compression efficiency.
When ROHCv2 profiles are used over a channel that cannot guarantee
in-order delivery, the decompressor may refrain from sending ACK
feedback for a sequentially late packet that is successfully
The decompressor should limit the rate at which it sends feedback,
for both ACKs and STATIC-NACK/NACKs, and should avoid sending
unnecessary duplicates of the same type of feedback message that may
be associated with the same event.