Dear all, draft-ietf-6tisch-msf-02.txt is just published.
The main changes made in this version are - Resolve the issues in Pending Elements section. - Using autonomous SHARED and non-SHARED cells. - Overall revised the specification. Additional information about how the issues in the pending elecments section are resolved is written below. (issues are separated by "=") Any reviews or comments are welcome! Tengfei Security issue on autonomous cell installation ==================== The autonomous cell to the pledge is installed on the join proxy before the pledge is authorized, which is a security issue. Related discussion: ttps:// mailarchive.ietf.org/arch/msg/6tisch/9jcaTddi6vLO5zHqTDNk6yqrNao [Resolved. During join process, only pledge required to install autonomous SHARED cell to JP, no action required from JP side.] Handling the case when bandwidth allocation exceeds available capacity ==================== When node A initializes a 6top ADD transaction to node B, but node B does NOT have enough bandwidth to allocate. What can node B do to indicate this case in its 6top response, and how should node A handle the packet after receiving the response? Related discussion: https://www.ietf.org/mail-archive/web/6tisch/current/msg06095.html [ToDo. According to 6top RFC8480, SUCCESS return code will be in the 6top response with an empty cellList possibly. For this case, I don't have a prefect solution. One solution I am using is marking the node B as 'no_resource', then updating it's routing table and filtering out the neighbor marked as 'no_resource'. However, this is a layer-violation design.] joining traffic MUST be sent in minimal cell ==================== Joining traffic MUST be sent in minimal cell. Issues link: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/11 [Resolved. Joining traffic will be sent on autonomous cell.] NumCellsElapsed shouldn't update on shared dedicated cell ==================== NumCellsElapsed is used to track under/overuse of cells. If we are backing off, i.e. skipping cells, should those skipped cells count towards NumCellsElapsed ? Issues link: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/13 [Resolved. The counters are only applied on managed unicast cell.] non-trusted packet shouldn't be accounted for for adapting traffic ==================== Mention that the untrusted packet (e.g. join request/response) shouldn't be counted for adapting the traffic. Issues link: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/15 [Resolved. clarified in adapting traffic section.] Customize the backoff exponent on dedicated shared cell. ==================== Use small backoff exponent on dedicated cell since it's only shared by two nodes. Discussion: do we still have this case with the merger of ASF? Issues link: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/17 [NotApplied. After ASF is merged, the dedicated shared cell is not used anymore. The autonomous SHARED cell is shared among a set of neighbors hence no need to customize the backoff exponent.] Adapting 6P Timeout ==================== After 6P TIMEOUT, the next 6P transaction should be configured using a larger TIMEOUT. Issues link: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/19 [ToDo. I am not sure this is an issue or not with current version. Need to discuss with Malisa and Simon.] Timeout calculation is wrong ==================== The 6P Timeout is calculated as : list style: symbols (1/(C+1))*(1/PDR)*SIXP_TIMEOUT_SEC_FACTOR It should be list style: symbols (1/(C+1))*(1/PDR)*SIXP_TIMEOUT_SEC_FACTOR according to the paper: https://arxiv.org/pdf/1507.05810.pdf Related link is at: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/22 [Resolved. The comments are taken and applied in the text.] Separate Cell Counters for TX and RX ==================== NumCellsUsed and NumCellsElapsed counter should be separated between TX and RX. Question: still applies? Related Link is at: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/20 [Resolved. The comments are taken and applied in the text.] Two slotframes ==================== Why are they two slotframes in same length? Issues link: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/25 [Resolved. The length of two slotframes are SHOULD be the same, instead of MUST.] Parameters for SAX is missing ==================== It'd be nice to have an example or a test vector in the draft in order to validate a SAX implementation used for MSF. This is critical for interoperability. Related discussion: https://mailarchive.ietf.org/arch/msg/6tisch/9jcaTddi6vLO5zHqTDNk6yqrNao [Resolved. An appredix of how to implement the sax is added.] Wrong end state statement ==================== In Section 1, one of the end states of the join process is having "one default unicast cell to/from each of its neighbors". It should be an autonomous TX cells to its Join Proxy. Related discussion: hhttps:// mailarchive.ietf.org/arch/msg/6tisch/9jcaTddi6vLO5zHqTDNk6yqrNao [Resolved.Comments are taken and applied in v2.] Dependency on RFC8180 shouldn't be a MUST ==================== Section 2 states "A node implementing MSF MUST implement the Minimal 6TiSCH Configuration [RFC8180], which defines the "minimal cell", a single shared cell providing minimal connectivity between the nodes in the network". MUST here is too strong. Some may want to use MSF with a base schedule other than the one defined in RFC8180, with full understanding on the implications by not following RFC8180. Proposal is to replace MUST by SHOULD. Related discussion: https://mailarchive.ietf.org/arch/msg/6tisch/9jcaTddi6vLO5zHqTDNk6yqrNao [Resolved. Comments are taken and applied in v2.] DIO can be unicast packet as indicated in RFC6550 ==================== Section 2 states "DODAG Information Objects (DIOs), defined by [RFC6550]. These are broadcast frames." However, the DIO as a response to a DIS is a unicast frame. Related discussion: hhttps:// mailarchive.ietf.org/arch/msg/6tisch/9jcaTddi6vLO5zHqTDNk6yqrNao [Resolved. Comments are taken and applied in v2.] Rules for broadcast frames is out of scope of MSF ==================== In Section 2, the rules for broadcast frames on the minimal cell seem to be an implementation-specific optimization. And it is beyond of the scope of this draft. This idea is about how to use the minimal cell efficiently; it's not directly related to how to use cells scheduled by MSF. Related discussion: https://mailarchive.ietf.org/arch/msg/6tisch/9jcaTddi6vLO5zHqTDNk6yqrNao [ToDo. I agree with this comment. Should we remove those text or keep them in the draft but mentioned those rules are not a MUST but a RECOMMENDED?] Wrong statement in Step 4 ==================== In Section 4.4 Step 3, the statement "After joining" in the explanation of step 3 should be "After having chosen a JP"? At this step 3, "joining" is not done. Related discussion: https://mailarchive.ietf.org/arch/msg/6tisch/9jcaTddi6vLO5zHqTDNk6yqrNao [Resolved. Comments are taken and applied in v2.] Rephrase NumCellsElapsed to NumCellsElapsed ==================== Rephrase NumCellsElapsed to NumCellsElapsed Issues link: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/14 [Resolved. Comments are taken and applied in v2.] Create a list of packets that be able to be sent on minimal cell ==================== The current text enumerates explicitly EBs and DIOs. The other broadcast packets are not mentioned. For example: RPL DIS or application-layer broadcast? It should mention EBs and DIOs but also say "any other broadcast packet". Issues link: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/24 [Resolved. Comments are taken and applied in v2.] Clarify term 'dedicated cell' ==================== Terminology: "dedicated" needs to be clarified in MSF. Issues link: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/26 [Resolved. dedicated cells are not used in v2. The type of cell used in the specification are autonomous non-SHARED cell, autonomous SHARED cell and managed unicast cell.] text is missing ==================== The text that describes the 6P command with cell options etc. was accidentally removed. Needs adding back somewhere. The text was: Step 5 - 6P ADD to Preferred Parent After having selected a preferred parent, the joined node MUST issue a 6P ADD command to that parent, with the following fields: list style: symbols CellOptions: set to TX=1,RX=1,SHARED=1 CellOptions: set to TX=1,RX=1,SHARED=1 NumCells: set to 1 CellList: at least 5 cells, chosen according to Section 7 /list> After the 6P Transaction is finished, the node MUST synchronize only to its preferred parent. At this point, the node has a dedicated cell which allows for bidirectional communication with its preferred parent. Issues link: https://github.com/twatteyne/draft-ietf-6tisch-msf/issues/29 [NotApplied. This comments is referring to the version before ASF is merged.] -- Chang Tengfei, Pre-Postdoctoral Research Engineer, Inria
_______________________________________________ 6tisch mailing list [email protected] https://www.ietf.org/mailman/listinfo/6tisch
