[alto] Meeting minutes for Oct 20, 2020

kaigao@scu.edu.cn Wed, 21 October 2020 03:15 UTC

Date: Wed, 21 Oct 2020 11:15:18 +0800
From: kaigao@scu.edu.cn
To: alto@ietf.org, "alto-weekly-meeting@googlegroups.com" <alto-weekly-meeting@googlegroups.com>
Content-Type: multipart/alternative; boundary="----=_Part_198683_1165483374.1603250118365"
MIME-Version: 1.0
Message-ID: <6248dc88.da5c.17549275ede.Coremail.kaigao@scu.edu.cn>
Archived-At: <https://mailarchive.ietf.org/arch/msg/alto/bwCWOKjLa_s7vS0OzmCAIgsKL_A>
Subject: [alto] Meeting minutes for Oct 20, 2020
Precedence: list

Hi everyone,

This is a quick summary of the meeting on Oct 20, 2020.

The authors of the WG documents are urged to reply to Vijay's email on plans to respond to WG reviews.

Richard mentioned the application of serving as the 3rd WG chair should be sent to Martin (our AD) by Oct 20.

On the recharter discussion, Chunshan gives a short presentation on the IB/OB issue. In particular, he mentioned that 3GPP has defined IB+OB for the edge computing case, where UE and UPF exchange information using a special 3GPP tunnel header (in release 16), and UPF exchanges information with local NEF using a new interface defined in release 17.

Richard asked whether is security/privacy issue. Chunshan replied that the information exposed by 3GPP standards are trustful. Also 3GPP has identified some metrics that are not sensitive and can be safely distributed to 3rd parities.

Sabine asked whether the two transport methods should be integrated or should there be a single point to get information of both RAN and edge information so that a query contains both time-sensitive and insensitive information. Chunshan replied that local NEF can reroute the request to the central NEF is the information is not time sensitive. Luis mentioned it is also possible to carry the information in packets, and there are existing efforts as in IFIT which piggyback telemetry information in packets.

Sabine suggested we should start a discussion on the mailing list and gather feedback on use cases which combine both RAN and edge information. She mentioned that many people are working on this issue and such a discussion can help getting more attraction from people outside the WG.

Luis shared his thoughts on treating ALTO/3GPP standards as underlay/overlay. He mentioned there is existing work on making data centers as overlay, and ALTO aggregates the information and answers queries from applications. Similarly, the 3GPP entities have only information of the edge and can use ALTO to learn information about the network.

Chunshan, Luis, Sabine and Richard, please correct me if some points are missed or misinterpreted. More detailed meeting minutes can be found at [1].

[1] https://docs.google.com/document/d/1H5HvSbOT54IPCniWi8Xrxive8q1UpV1ZyYda46P0-b0/edit?usp=sharing

Best,

Kai

[alto] Meeting minutes for Oct 20, 2020 kaigao