Re: [tsvwg] DTLS 1.3 over SCTP

[Claudio Porfiri <[e-mail]>]

Hi Michael,
Unfortunately message rate is not known.

BR,
Claudio

-----Original Message-----
From: tsvwg <[e-mail]> On Behalf Of Michael Tuexen
Sent: Friday, 14 July 2023 12:24
To: Claudio Porfiri <claudio.porfiri=[e-mail]>
Cc: tsvwg IETF list <[e-mail]>
Subject: Re: [tsvwg] DTLS 1.3 over SCTP

> On 14. Jul 2023, at 12:16, Claudio Porfiri <claudio.porfiri=[e-mail]> wrote:
> 
> Hi Michael,
> Yes, they send atomic messages. No fragmentation is foreseen.
OK, great.

Any prediction/requirements on the message rates and/or bandwidths?

Best regards
Michael
> 
> BR,
> Claudio
> 
> -----Original Message-----
> From: tsvwg <[e-mail]> On Behalf Of Michael Tuexen
> Sent: Friday, 14 July 2023 11:42
> To: Claudio Porfiri <claudio.porfiri=[e-mail]>
> Cc: tsvwg IETF list <[e-mail]>
> Subject: Re: [tsvwg] DTLS 1.3 over SCTP
> 
>> On 14. Jul 2023, at 11:24, Claudio Porfiri <claudio.porfiri=[e-mail]> wrote:
>> 
>> Hi Michael,
>> As an example, S1-AP protocol that exploits SCTP and is described in TS 36.413 (the protocol) and TS 36.412 (the transport) can send a single message that contains all the features from a User Equipment and the total maximum size of this signal may grow up to 142k.
> I see.
>> S1-AP is not the most demanding though, there are signals in Xn protocol specified in TS 48.423 that can grow up to more than 500k
> 500KB in a single message. What message rates are expected? What is the expected bandwidth of links carrying this
> traffic.
> 
> I assume that sending/receiving the messages are atomic operations from an application point of view. Is that correct?
> 
> Best regards
> Michael
>> 
>> BR,
>> Claudio.
>> 
>> -----Original Message-----
>> From: Michael Tuexen <[e-mail]> 
>> Sent: Friday, 14 July 2023 11:14
>> To: Claudio Porfiri <[e-mail]>
>> Cc: tsvwg IETF list <[e-mail]>
>> Subject: Re: [tsvwg] DTLS 1.3 over SCTP
>> 
>>> On 14. Jul 2023, at 08:04, Claudio Porfiri <claudio.porfiri=[e-mail]> wrote:
>>> 
>>> Hi Michael,
>>> I am reading this new draft but I have a quick comment at once.
>>> The requirements from 3GPP towards SCTP are far beyond the limit of 64k, for instance S1-AP needs up to 142k.
>> Hi Claudio,
>> 
>> could you provide some insight what kind of signalling message needs up to 142KB?
>> Just wondering what kind of information needs 142 KB.
>>> This new draft improves the situation but doesn't solve it.
>> The focus is not limited to 3GPP. We are trying to improve the situation
>> for RFC 6083 with minimizing the changes.
>> 
>> Best regards
>> Michael
>>> 
>>> Best regards,
>>> Claudio
>>> 
>>> -----Original Message-----
>>> From: tsvwg <[e-mail]> On Behalf Of Michael Tuexen
>>> Sent: Thursday, 13 July 2023 16:35
>>> To: tsvwg IETF list <[e-mail]>
>>> Subject: [tsvwg] DTLS 1.3 over SCTP
>>> 
>>> Dear all,
>>> 
>>> Hannes Tschofenig and myself have submitted an ID for using DTLS 1.3 over SCTP:
>>> https://www.ietf.org/archive/id/draft-tuexen-tsvwg-rfc6083-bis-02.html
>>> 
>>> This is an alternative to
>>> https://www.ietf.org/archive/id/draft-ietf-tsvwg-dtls-over-sctp-bis-06.html
>>> 
>>> Our document is based on RFC 6083. The major differences are:
>>> * Use DTLS 1.3 instead of DTLS 1.0
>>> * Use key updates instead of renegotiation. This limits the number of
>>> rekeyings to 2^64, but that should not limit in real world scenarios.
>>> * Bump the maximum user message size to 64KB by using RFC 8449.
>>> 
>>> Any comments welcome.
>>> 
>>> Best regards
>>> Michael