| Network Working Group | J. Mattsson |
| Internet-Draft | Ericsson AB |
| Intended status: Informational | March 13, 2017 |
| Expires: September 14, 2017 |
Message Size Overhead of CoAP Security Protocols
draft-mattsson-core-security-overhead-00
This document analyzes and compares per-packet message size overheads when using different security protocols to secure CoAP. The analyzed security protocols are DTLS 1.2, DTLS 1.3, TLS 1.2, TLS 1.3, and OSCOAP. DTLS and TLS are analyzed with and without compression. DTLS are analyzed with two different alternatives for header compression.
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at http://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."
This Internet-Draft will expire on September 14, 2017.
Copyright (c) 2017 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 (http://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. Code Components extracted from this document must include Simplified BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Simplified BSD License.
This document analyzes and compares per-packet message size overheads when using different security protocols to secure CoAP over UPD [RFC7252] and TCP [I-D.ietf-core-coap-tcp-tls]. The analyzed security protocols are DTLS 1.2 [RFC6347], DTLS 1.3 [I-D.rescorla-tls-dtls13], TLS 1.2 [RFC5246], TLS 1.3 [I-D.ietf-tls-tls13], and OSCOAP [I-D.ietf-core-object-security]. The DTLS and TLS record layers are analyzed with and without compression. DTLS are analyzed with two different alternatives ([RFC7400] and [raza-6lo-compressed-dtls]) for header compression.
To enable comparison, all the overhead calculations in this section use AES-CCM with a tag length of 8 bytes, a plaintext of 6 bytes, and the sequence number ‘05’. This follows the example in [RFC7400], Figure 16.
This example is taken directly from [RFC7400], Figure 16. The nonce follow the strict profiling given in [RFC7925].
DTLS 1.2 Record Layer (35 bytes, 29 bytes overhead): 17 fe fd 00 01 00 00 00 00 00 05 00 16 00 01 00 00 00 00 00 05 ae a0 15 56 67 92 4d ff 8a 24 e4 cb 35 b9 Content type: 17 Version: fe fd Epoch: 00 01 Sequence number: 00 00 00 00 00 05 Length: 00 16 Nonce: 00 01 00 00 00 00 00 05 Ciphertext: ae a0 15 56 67 92 ICV: 4d ff 8a 24 e4 cb 35 b9
DTLS 1.2 gives 29 bytes overhead.
Note that the compressed overhead is dependent on the parameters epoch, sequence number, and length. The following is only an example.
Note that the sequence number ‘01’ used in [RFC7400], Figure 15 gives an exceptionally small overhead that is not representative at all.
Note that this header compression is not available when DTLS is exchanged over transports that do not use 6LoWPAN together with 6LoWPAN-GHC.
Compressed DTLS 1.2 Record Layer (22 bytes, 16 bytes overhead): b0 c3 03 05 00 16 f2 0e ae a0 15 56 67 92 4d ff 8a 24 e4 cb 35 b9 Compressed DTLS 1.2 Record Layer Header and Nonce: b0 c3 03 05 00 16 f2 0e Ciphertext: ae a0 15 56 67 92 ICV: 4d ff 8a 24 e4 cb 35 b9
When compressed with 6LoWPAN-GHC, DTLS 1.2 with the above parameters (epoch, sequence number, length) gives 16 bytes overhead.
Note that the compressed overhead is dependent on the parameters epoch and sequence number. The following is only an example.
Note that this header compression is not available when DTLS is exchanged over transports that do not use 6LoWPAN together with raza-6lo-compressed-dtls.
Compressed DTLS 1.2 Record Layer (19 bytes, 13 bytes overhead): 90 17 01 00 05 ae a0 15 56 67 92 4d ff 8a 24 e4 cb 35 b9 NHC 90 Compressed DTLS 1.2 Record Layer Header and Nonce: 17 01 00 05 Ciphertext: ae a0 15 56 67 92 ICV: 4d ff 8a 24 e4 cb 35 b9
When compressed with raza-6lo-compressed-dtls, DTLS 1.2 with the above parameters (epoch, sequence number) gives 13 bytes overhead.
The only change compared to DTLS 1.2 is that the DTLS 1.3 record layer does not have an explicit nonce.
DTLS 1.3 Record Layer (27 bytes, 21 bytes overhead): 17 fe fd 00 01 00 00 00 00 00 05 00 0e ae a0 15 56 67 92 4d ff 8a 24 e4 cb 35 b9 Content type: 17 Version: fe fd Epoch: 00 01 Sequence number: 00 00 00 00 00 05 Length: 00 0e Ciphertext: ae a0 15 56 67 92 ICV: 4d ff 8a 24 e4 cb 35 b9
DTLS 1.3 gives 21 bytes overhead.
Note that the overhead is dependent on the parameters epoch, sequence number, and length. The following is only an example.
Note that this header compression is not available when DTLS is exchanged over transports that do not use 6LoWPAN together with 6LoWPAN-GHC.
Compressed DTLS 1.3 Record Layer (20 bytes, 14 bytes overhead):
b0 c3 11 05 00 0e ae a0 15 56 67 92 4d ff 8a 24
e4 cb 35 b9
Compressed DTLS 1.3 Record Layer Header and Nonce:
b0 c3 11 05 00 0e
Ciphertext:
ae a0 15 56 67 92
ICV:
4d ff 8a 24 e4 cb 35 b9
When compressed with 6LoWPAN-GHC, DTLS 1.3 with the above parameters (epoch, sequence number, length) gives 14 bytes overhead.
Note that the compressed overhead is dependent on the parameters epoch and sequence number. The following is only an example.
Note that this header compression is not available when DTLS is exchanged over transports that do not use 6LoWPAN together with raza-6lo-compressed-dtls.
Note that this header compression is not available when DTLS is exchanged over transports that do not use 6LoWPAN together with raza-6lo-compressed-dtls.
Compressed DTLS 1.3 Record Layer (19 bytes, 13 bytes overhead): 90 17 01 00 05 ae a0 15 56 67 92 4d ff 8a 24 e4 cb 35 b9 NHC 90 Compressed DTLS 1.3 Record Layer Header and Nonce: 17 01 00 05 c3 03 05 00 16 f2 0e Ciphertext: ae a0 15 56 67 92 ICV: 4d ff 8a 24 e4 cb 35 b9
When compressed with raza-6lo-compressed-dtls, DTLS 1.3 with the above parameters (epoch, sequence number) gives 13 bytes overhead.
The changes compared to DTLS 1.2 is that the TLS 1.2 record layer does not have epoch and sequence number, and that the version is different.
TLS 1.2 Record Layer (27 bytes, 21 byte overhead): 17 03 03 00 16 00 00 00 00 00 00 00 05 ae a0 15 56 67 92 4d ff 8a 24 e4 cb 35 b9 Content type: 17 Version: 03 03 Length: 00 16 Nonce: 00 00 00 00 00 00 00 05 Ciphertext: ae a0 15 56 67 92 ICV: 4d ff 8a 24 e4 cb 35 b9
TLS 1.2 gives 21 bytes overhead.
Note that the overhead is dependent on the parameters epoch, sequence number, and length. The following is only an example.
Note that this header compression is not available when TLS is exchanged over transports that do not use 6LoWPAN together with 6LoWPAN-GHC.
Compressed TLS 1.2 Record Layer (23 bytes, 17 bytes overhead):
05 17 03 03 00 16 85 0f 05 ae a0 15 56 67 92 4d
ff 8a 24 e4 cb 35 b9
Compressed TLS 1.2 Record Layer Header and Nonce:
05 17 03 03 00 16 85 0f 05
Ciphertext:
ae a0 15 56 67 92
ICV:
4d ff 8a 24 e4 cb 35 b9
When compressed with 6LoWPAN-GHC, TLS 1.2 with the above parameters (epoch, sequence number, length) gives 17 bytes overhead.
The change compared to TLS 1.2 is that the TLS 1.3 record layer uses a different version.
TLS 1.3 Record Layer (27 bytes, 21 byte overhead): 17 03 01 00 16 00 00 00 00 00 00 00 05 ae a0 15 56 67 92 4d ff 8a 24 e4 cb 35 b9 Content type: 17 Version: 03 01 Length: 00 16 Nonce: 00 00 00 00 00 00 00 05 Ciphertext: ae a0 15 56 67 92 ICV: 4d ff 8a 24 e4 cb 35 b9
TLS 1.3 gives 21 bytes overhead.
Note that the overhead is dependent on the parameters epoch, sequence number, and length. The following is only an example.
Note that this header compression is not available when TLS is exchanged over transports that do not use 6LoWPAN together with 6LoWPAN-GHC.
Compressed TLS 1.3 Record Layer (23 bytes, 17 bytes overhead):
02 17 03 c3 01 16 85 0f 05 ae a0 15 56 67 92 4d
ff 8a 24 e4 cb 35 b9
Compressed TLS 1.3 Record Layer Header and Nonce:
02 17 03 c3 01 16 85 0f 05
Ciphertext:
ae a0 15 56 67 92
ICV:
4d ff 8a 24 e4 cb 35 b9
When compressed with 6LoWPAN-GHC, TLS 1.3 with the above parameters (epoch, sequence number, length) gives 17 bytes overhead.
Note that the overhead is dependent on the included CoAP Option numbers, if the CoAP method allows payload, as well as the length of the OSCOAP parameters Sender ID and sequence number. The below calculation uses Method = POST, Option Delta = ‘9’, and Sender ID = ‘25’, and is only an example.
OSCOAP Request (19 bytes, 13 bytes overhead): 90 19 05 41 25 ae a0 15 56 67 92 4d ff 8a 24 e4 cb 35 b9 CoAP Delta and Option Length: 90 Compressed COSE Header: 19 05 41 25 Ciphertext: ae a0 15 56 67 92 ICV: 4d ff 8a 24 e4 cb 35 b9
OSCOAP Response (15 bytes, 9 bytes overhead): 90 ae a0 15 56 67 92 4d ff 8a 24 e4 cb 35 b9 CoAP Delta and Option Length: 90 Ciphertext: ae a0 15 56 67 92 ICV: 4d ff 8a 24 e4 cb 35 b9
OSCOAP with the above parameters gives 13 bytes overhead for requests and 9 bytes overhead for responses.
Unlike DTLS and TLS, OSCOAP has much smaller overhead for responses than requests.
The compression overhead (GHC) is dependent on the parameters epoch, sequence number, and length. The following overheads should be representative for sequence numbers with the same length.
The compression overhead (raza-6lo-compressed-dtls) is dependent on the length of the parameters epoch and sequence number. The following overheads apply for all sequence numbers with the same length.
The OSCOAP overhead is dependent on the included CoAP Option numbers, if the CoAP method allows payload, as well as the length of the OSCOAP parameters Sender ID and sequence number.
Sequence Number '05' '1005' '100005' ---------------------------------------------------------- DTLS 1.2 29 29 29 DTLS 1.3 21 21 21 TLS 1.2 21 21 21 TLS 1.3 21 21 21 ---------------------------------------------------------- DTLS 1.2 (GHC) 16 16 17 DTLS 1.2 (Raza) 13 13 14 DTLS 1.3 (GHC) 14 14 15 DTLS 1.3 (Raza) 13 13 14 TLS 1.2 (GHC) 17 18 19 TLS 1.3 (GHC) 17 18 19 ---------------------------------------------------------- OSCOAP Request 13 14 15 OSCOAP Response 9 9 9
Figure 1: Overhead as a function of sequence number
DTLS 1.2 has quite a large overhead as it uses an explicit sequence number and an explicit nonce. DTLS 1.3, TLS 1.2, and TLS 1.3 have significantly less overhead.
Both DTLS compression methods provides very good compression. raza-6lo-compressed-dtls achieves slightly better compression but requires state. GHC is stateless but provides slightly worse compression. As DTLS 1.3 uses the same version number as DTLS 1.2, both GHC and raza-6lo-compressed-dtls works well also for DTLS 1.3.
The Generic Header Compression (6LoWPAN-GHC) is not very generic (the static dictionary is more or less a DTLS record layer) and the compression of TLS is significantly worse than the compression of DTLS. Similar compression levels as for DTLS could be achieved also for TLS, but this would require different static dictionaries for each version of TLS (as TLS 1.2 and TLS 1.3 uses different version numbers).
The header compression is not available when (D)TLS is exchanged over transports that do not use 6LoWPAN together with 6LoWPAN-GHC or raza-6lo-compressed-dtls.
OSCOAP has much lower overhead than DTLS and TLS. The overhead of OSCOAP is smaller than DTLS over 6LoWPAN with compression, and this small overhead is achieved even on deployments without 6LoWPAN or 6LoWPAN without DTLS compression. OSCOAP is lightweight because it makes use of some excellent features in CoAP, CBOR, and COSE.
This document is purely informational.
The authors want to thank Ari Keränen for reviewing previous versions of the draft.