Re: [core] a question about the Message Format in draft-tschofenig-core-coap-tcp-tls-00

"zhengfei (E)" <zhengfei10@huawei.com> Sat, 21 November 2015 03:20 UTC

From: "zhengfei (E)" <zhengfei10@huawei.com>
To: "cabo@tzi.org" <cabo@tzi.org>
Thread-Topic: Re: a question about the Message Format in draft-tschofenig-core-coap-tcp-tls-00
Thread-Index: AQHRJAuRshW4SsuulU+iuzBYkb95Vw==
Date: Sat, 21 Nov 2015 03:20:27 +0000
Message-ID: <C6223E7F04FAF9418F0C162994D4E8339AE621@szxeml513-mbx.china.huawei.com>
References: <tencent_5D55430907F8140C1F7B7349@qq.com>
In-Reply-To: <tencent_5D55430907F8140C1F7B7349@qq.com>
Accept-Language: zh-CN, en-US
Content-Language: zh-CN
Content-Type: multipart/alternative; boundary="_000_C6223E7F04FAF9418F0C162994D4E8339AE621szxeml513mbxchina_"
MIME-Version: 1.0
Archived-At: <http://mailarchive.ietf.org/arch/msg/core/TbSLYXXr-LlUb8DT_qf-VseztPo>
Cc: "core@ietf.org" <core@ietf.org>
Subject: Re: [core] a question about the Message Format in draft-tschofenig-core-coap-tcp-tls-00
Precedence: list

Hi, Bormann

Thank you very much for your quick updating to -01,
I think a clear uniform scheme(L1 only) could greatly improve the compatibility between different implementations.

The simple 4-byte TCP-CoAP packet head whose length is same to UDP-CoAP really looks good,
especially when CoAP stack will implement to be a dual stack, who supports both over UDP and over TCP simultaneously.
I will modify our implementation from L3 to L1 too, following this updated draft.
-------
Finally, sorry that I have used my private email and an empty name previously, my name is Zheng Fei.
Thanks again.☺

------------------ 原始邮件 ------------------
发件人: Carsten Bormann <cabo@tzi.org<mailto:cabo@tzi.org>>
发送时间: 2015年11月20日 01:13
收件人: 774163 <774163@qq.com<mailto:774163@qq.com>>
抄送: slemay <slemay@zebra.com<mailto:slemay@zebra.com>>, vsolorzanobarboza <vsolorzanobarboza@zebra.com<mailto:vsolorzanobarboza@zebra.com>>, Hannes.tschofenig <Hannes.tschofenig@gmx.net<mailto:Hannes.tschofenig@gmx.net>>
主题: Re: a question about the Message Format in draft-tschofenig-core-coap-tcp-tls-00

Hi 774163,

(how do you pronounce your name? :-)

> I have a question about the latest document now.
> About the Message Format, In section 4, I can see that you choose the L1
> scheme, and no other schemes are mentioned, so if I can draw conclusion
> that L1 is the draft standard, and other schemes are No longer within
> the scope of the consideration?

That is the result of the Yokohama meeting and the consensus call we had
after that.

> But in section 3 --> "unless alternative L3 below is chosen"
> It seems that other schemes are still valid? I'm really confused about this.

Oops.

Failing to remove this remnant was my editorial oversight.
Thank you for alerting us.  I just submitted an update, -01, to remove
this confusing passage.

> I think most users are hesitating between L1 and L3. Actually we have
> implement the L3 scheme but I really like L1 scheme.

Indeed, the WG has been hesitating a while as well.
In the end, the simplicity of L1 won over the generality (and slightly
improved efficiency) of L3.

> Because L1 is
> simple, its length is fixed so it easy to implement. And I think 2-byte
> is enough in most cases. If 2-byte is not enough, 4-byte still can be
> not enough in theory. Besides, we has block protocol to deal the extreme
> case. :)

Exactly!
These are pretty much the thought processes that led to this result.

> so if L1 scheme is the only valid scheme?

Yes, that is the outcome of the consensus process.

Thanks again for reaching out and notifying us of this oversight.

Grüße, Carsten

Re: [core] a question about the Message Format in… zhengfei (E)