]> Huawei Technologies Duesseldorf GmbH

Riesstrasse 25 Munich D-80992 Germany cenk.gundogan@huawei.com

HAW Hamburg

Berliner Tor 7 Hamburg D-20099 Germany t.schmidt@haw-hamburg.de http://inet.haw-hamburg.de/members/schmidt

Network Systems Research and Design

4 Shady Hill Square Cambridge MA 02138 USA daveoran@orandom.net

TUD Dresden University of Technology

Helmholtzstr. 10 Dresden D-01069 Germany m.waehlisch@tu-dresden.de

IRTF ICNRG CCNx constrained networks compressed delta time IoT CCNx utilizes delta time for a number of functions. When using CCNx in environments with constrained nodes or bandwidth constrained networks, it is valuable to have a compressed representation of delta time. In order to do so, either accuracy or dynamic range has to be sacrificed. Since the current uses of delta time do not require both simultaneously, one can consider a logarithmic encoding. This document updates RFC 8609 (CCNx messages in TLV Format) to specify this alternative encoding. This document is a product of the IRTF Information-Centric Networking Research Group (ICNRG).

CCNx is well suited for Internet of Things (IoT) applications . LoWPAN adaptation layers (e.g., ) confirm the advantages of a space-efficient packet encoding for low-power and lossy networks. CCNx utilizes absolute and delta time values for a number of functions. When using CCNx in resource-constrained environments, it is valuable to have a compact representation of time values. However, any compact time representation has to sacrifice accuracy or dynamic range. For some time uses this is relatively straightforward to achieve, for other uses, it is not. As a result of experiments in bandwidth-constrained IEEE 802.15.4 deployments , this document discusses the various cases of time values, proposes a compact encoding for delta times, and updates to utilize this encoding format in CCNx messages. This document has received fruitful reviews by the members of the research group (see the Acknowledgments section). It is the consensus of ICNRG that this document should be published in the IRTF Stream of the RFC series. This document does not constitute an IETF standard.

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 when, and only when, they appear in all capitals, as shown here. This document uses the terminology of and for CCNx entities. The following terms are used in the document and defined as follows: synonym for octet a time offset measured in seconds an 8-bit encoded time value as defined in

CCNx, as currently specified in , utilizes delta time for only the lifetime of an Interest message (see sections 2.1, 2.2, 2.4.2, 10.3 of ). It is a hop-by-hop header value, and is currently encoded via the T_INTLIFE TLV as a 64-bit integer ( section 3.4.1). While formally an optional TLV, in all but some corner cases every Interest message is expected to carry the Interest Lifetime TLV, and hence having compact encoding is particularly valuable for keeping Interest messages short. Since the current uses of delta time do not require both accuracy and dynamic range simultaneously, one can consider a logarithmic encoding such as that specified in and outlined in . This document updates CCNx messages in TLV Format to permit this alternative encoding for selected time values.

CCNx, as currently specified in , utilizes absolute time for various important functions. Each of these absolute time usages poses a different challenge for a compact representation. These are discussed in the following subsections.

Signature Time is the time the signature of a content object was generated (sections 8.2-8.4). Expiry Time indicates the expiry time of a content object (section 4). Both values are content message TLVs and represent absolute timestamps in milliseconds since the POSIX epoch. They are currently encoded via the T_SIGTIME and T_EXPIRY TLVs as 64-bit unsigned integers (see section 3.6.4.1.4.5 and 3.6.2.2.2). Both time values could be in the past, or in the future, potentially by a large delta. They are also included in the security envelope of the message. Therefore, it seems there is no practical way to define an alternative compact encoding that preserves its semantics and security properties; hence we don't consider it further as a candidate.

Recommended Cache Time (RCT) for a content object (see section 4) is a hop-by-hop header stating the expiration time for a cached content object in milliseconds since the POSIX epoch. It is currently encoded via the T_CACHETIME TLV as a 64-bit unsigned integer (see section 3.4.2). A recommended cache time could be far in the future, but cannot be in the past and is likely to be a reasonably short offset from the current time. Therefore, this document allows the recommended cache time to be interpreted as a relative time value rather than an absolute time, since the semantics associated with an absolute time value do not seem to be critical to the utility of this value. This document therefore updates the recommended cache time with the following rule set: Use absolute time as per Use relative time, if the compact time representation is used (see and ) If relative time is used, the time offset recorded in a message will typically not account for residence times on lower layers (e.g., for processing, queuing) and link delays for every hop. The recommended cache time can thus not be expressed as accurate as with absolute time. This document targets low-power networks, where delay bounds are rather loose, or do not exist. An accumulated error due to transmission delays in the range of milliseconds and seconds for the recommended cache time is still tolerable in these networks, and does not impact the protocol performance. Networks with tight latency bounds use dedicated hardware, optimized software routines, and traffic engineering to reduce latency variations. Time offsets can then be corrected on every hop to yield exact cache times.

This document uses the compact time representation of ICNLoWPAN (see section 7 of ) that is inspired by and . Its logarithmic encoding supports a representation ranging from milliseconds to years. depicts the logarithmic nature of this time representation.

Time codes encode exponent and mantissa values in a single byte, but in contrast to the representation in , time codes only encode non-negative numbers and hence do not include a separate sign bit. shows the configuration of a time code: an exponent width of 5 bits, and a mantissa width of 3 bits.

+----+----+----+----+----+----+----+----+ | exponent (b) | mantissa (a) | +----+----+----+----+----+----+----+----+ 0 1 2 3 4 5 6 7 ]]>

The base unit for time values are seconds. A time value is calculated using the following formula (adopted from and ), where (a) represents the mantissa, (b) the exponent, and (C) a constant factor set to C := 1/32. (0 + a/8) * 2 * C (1 + a/8) * 2^b * C The subnormal form provides a gradual underflow between zero and the smallest normalized number. Eight time values exist in the subnormal range [0s,~0.0546875s] with a step size of ~0.0078125s between each time value. This configuration also encodes the following convenient numbers in seconds: [1, 2, 4, 8, 16, 32, 64, ...]. further includes test vectors to illustrate the logarithmic range. An example algorithm to encode a time value into the corresponding exponent and mantissa is given as pseudo code in . Not all time values can be represented by a time code. For these instances, the closest time code is chosen that is smaller than the value to encode.

(1 + 7/8) * 2^31 * C) // check bounds return (7, 31) // return maximum else b = floor (log2(v / C)) // round down a = floor ((v / (2^b * C) - 1) * 8) // round down return (a, b) ]]>

As an example: No specific time code for 0.063 exists, but this algorithm maps to the closest valid time code that is smaller, i.e., exponent 1 and mantissa 0 (the same as for time value 0.0625).

A straightforward way to accommodate the compact time approach is to use a 1-byte length field to indicate this alternative encoding while retaining the existing TLV registry entries. This approach has backward compatibility problems, but is still considered for the following reasons: Both CCNx RFCs are experimental and not Standards Track, hence expectations for forward and backward compatibility are not as stringent. "Flag day" upgrades of deployed CCNx networks, while inconvenient, are still feasible. The major use case for these compressed encodings are smaller-scale IoT and/or sensor networks where the population of consumers, producers, and forwarders is reasonably small. Since the current TLVs have hop-by-hop semantics, they are not covered by any signed hash and hence may be freely re-encoded by any forwarder. That means a forwarder supporting the new encoding can translate freely between the two encodings. The alternative of assigning new TLV registry values does not substantially mitigate the interoperability problems anyway.

The Interest Lifetime definition in allows for a variable-length lifetime representation, where a length of 1 encodes the linear range [0,255] in milliseconds. This document changes the definition to always encode 1-byte Interest lifetime values in the compact time value representation ().

1 | +---------------+---------------+---------------+---------------+ / / / Lifetime (Length octets) / / / +---------------+---------------+---------------+---------------+ ]]>

The Recommended Cache Time definition in specifies an absolute time representation that is of a length fixed to 8 bytes. This document changes the definition to always encode 1-byte Recommended Cache Time values in the compact relative time value representation ().

The packet processing is adapted to calculate an absolute time from the relative time code based on the absolute reception time. On transmission, a new relative time code is calculated based on the current system time.

This document has no IANA actions.

This document makes no semantic changes to , nor to any of the security properties of the message encodings of , and hence has the same security considerations as those two existing documents. Additional considerations need to be made in networks that deploy forwarders with support (updated forwarder) and without support (legacy forwarder) for this compact time representation: A legacy forwarder interprets a time code as an Interest lifetime between 0 and 255 milliseconds. This may lead to a degradation of network performance, since Pending Interest Table (PIT) entries timeout quicker than expected. An updated forwarder interprets short lifetimes set by a legacy forwarder as time codes, thus configuring timeouts of up to 4 years. This leads to an inefficient occupation of state space. A legacy forwarder considers a Recommended Cache Time with length 1 as a structural or syntactical error and it SHOULD discard the packet (section 10.3.9). Otherwise, a legacy forwarder interprets time codes as absolute time values far in the past, which impacts cache utilization.

&RFC2119; &RFC8174; &RFC8569; &RFC8609; &RFC5497; &RFC7927; &RFC9139; HAW Hamburg HAW Hamburg HAW Hamburg FU Berlin Computer Communications, Vol. 164, No. 1, p. 114–123, Elsevier

The test vectors in show sample time codes and their corresponding time values according to the algorithm outlined in . Time Code Time Value (seconds) 0x00 0.0000000 0x01 0.0078125 0x04 0.0312500 0x08 0.0625000 0x15 0.2031250 0x28 1.0000000 0x30 2.0000000 0xF8 67108863.0000000 0xFF 125829120.0000000

A forwarder frequently converts compact time into milliseconds to compare Interest lifetimes and the duration of cache entries. On many architectures, multiplication and division perform slower than addition, subtraction, and bit shifts. The following equations approximate the formulas in , and scale from seconds into the milliseconds range by applying a factor of 2^10 instead of 10^3. This results in an error of 2.4%. 2^10 * a * 2^-3 * 2^1 * 2^-5
= a << 3 (2^10 + a * 2^-3 * 2^10) * 2^b * 2^-5
= (1 << 5 + a << 2) << b

We would like to thank the active members of the ICNRG research group for constructive thoughts. In particular, we thank Marc Mosko and Ken Calvert for their valuable feedback on the encoding scheme and the protocol integration. Special thanks also to Carsten Bormann for his constructive in-depth comments during the IRSG review.