| 6tisch | S. Anamalamudi |
| Internet-Draft | M. Zhang |
| Intended status: Standards Track | AR. Sangi |
| Expires: February 4, 2017 | Huawei Technologies |
| C. Perkins | |
| Futurewei | |
| S.V.R.Anand | |
| Indian Institute of Science | |
| August 03, 2016 |
Scheduling Function One (SF1) for hop-by-hop Scheduling in 6tisch Networks
draft-satish-6tisch-6top-sf1-02
This document defines a 6top Scheduling Function called "Scheduling Function One" (SF1) to reserve, label and schedule the end-to-end resources hop-by-hop through distributed Resource Reservation Protocol(RSVP). SF1 uses the 6P signaling messages with a global TrackID to add/delete cells in end-to-end L2-bundles of isolated instance.
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 February 4, 2017.
Copyright (c) 2016 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.
L3-bundle(Instance-1,Instance-2,...Instance-n)
------------------------------------------------->
nodeA<------------------------------------------------- nodeB
L3-bundle(Instance-1,Instance-2,...Instance-n)
Figure 1: L3-bundle for aggregated traffic flows in 1-hop with SF0.
L2-bundle(Instance-1) L2-bundle(Instance-1)
-----------------------> ------------------>
<------------------------ <-------------------
L2-bundle(Instance-1) L2-bundle(Instance-1)
L2-bundle(Instance-2) L2-bundle(Instance-2)
----------------------> ----------------->
Sender<-----------------------nodeB <----------------- Receiver
L2-bundle(Instance-2) L2-bundle(Instance-2)
. .
. .
L2-bundle(Instance-n) L2-bundle(Instance-n)
-----------------------> -------------------->
<------------------------ <--------------------
L2-bundle(Instance-n) L2-bundle(Instance-n)
Figure 2: Dedicated L2-bundles for end-to-end isolated traffic flows with SF1
With Scheduling Function Zero (SF0) [I-D.dujovne-6tisch-6top-sf0], on-the-fly cell scheduling (ADD/DELETE) to 1-hop neighbors can be achieved for aggregated (best-effort) traffic flows. In other words, all the instances from nodeA to nodeB in Fig. 1 are scheduled in a single L3-bundle (IP link). [I-D.ietf-detnet-use-cases]. For such applications, per-instance based L2-bundles need to be scheduled hop-by-hop in between application sender and receiver nodes [I-D.ietf-6tisch-architecture]. In addition, cells in the scheduled end-to-end L2-bundles of each instance have to be dynamically adapted for bursty time-critical traffic flows. To achieve, end-to-end track has to be installed with a global TrackID that is associated with the L2-bundles of each instance. With 1-hop based SF0 cell scheduling, it is difficult to schedule dedicated end-to-end cells for isolated traffic flows. In addition, global bandwidth estimation through Resource Reservation protocol is required for bandwidth allocation in multi-hop cell scheduling. This draft proposes a Scheduling Function One (SF1) to schedule end-to-end dedicated L2-bundles for each instance, and to dynamically adapt the cells in scheduled L2-bundles of ongoing instance through RSVP protocol(see Fig. 2).
With SF1, Sender determines when to reserve the end-to-end resources, support implicit label switching(GMPLS), schedule the labeled L2-bundles hop-by-hop, associate the global TrackID for labeled L2-bundles, and dynamically adapt the cells in ongoing instance through distributed Resource Reservation Protocol (RSVP-lite). The triggering events in SF1 are as follows :
In both cases, distributed RSVP-lite (explained in Section .2.1) is triggered to provide end-to-end resource reservations along with scheduling operations.
In this specification, an end-to-end route path is assumed to be available with reactive P2P-RPL (Storing or non-storing mode) protocols. A distributed Resource Reservation Protocol (RSVP-lite) with 6tisch scheduling capability is designed to schedule the labeled reserved resources hop-by-hop for isolated instance. SF1 of application sender will trigger the RSVP-lite operation, whenever it has time critical traffic flow towards the receiver. The RSVP-lite has two messages namely (1) RSVP-PATH message(Sender to Receiver) and (2) RSVP-RESV message(Receiver to Sender).
The basic RSVP-PATH message [RFC2205] is used to carry the "Sender Traffic Specification" along with "characterization parameters" from sender to receiver. Since RSVP treat objects as opaque data, it is valid to assume other protocol(eg., GMPLS, 6P) as an object in RSVP- PATH messages.
The format of PATH message with the support of 6tisch scheduling capabilities (6P and SF1) is as follows :
<Path Message> ::= <Common Header> [ <INTEGRITY> ]
[ [<MESSAGE_ID_ACK> | <MESSAGE_ID_NACK>] ... ]
[ <MESSAGE_ID> ]
<SESSION> <RSVP_HOP>
<TIME_VALUES>
[ <EXPLICIT_ROUTE> ]
<LABEL_REQUEST>
[ <PROTECTION> ]
[ <LABEL_SET> ... ]
[<SF1 OPERATION REQUEST>]
[<6P OPERATION REQUEST>]
[ <SESSION_ATTRIBUTE> ]
[ <NOTIFY_REQUEST> ]
[ <ADMIN_STATUS> ]
[ <POLICY_DATA> ... ]
<sender descriptor>
"SF1 OPERATION REQUEST" and "6P OPERATION REQUEST" are added in the PATH message to check for 6tisch scheduling capabilities within the intermediate nodes from sender to receiver. "Timeslot Switching Capability(TSC)" is used as an implicit labels to switch the cell at intermediate nodes [RFC3473]. The message format of the "TSC" is out-of-scope in this specification. "LABEL_REQUEST" in path message should set to "Timeslot Switching Capability". "RPLInstanceID" is added in the "SENDER_TEMPLATE" to create Global TrackID during 6P transactions of RSVP-RESV message. Whenever the intermediate node won't support the "Timeslot switching Capability" or "6P transactions" or "SF1 operation" then it needs to send a "PathErr" message back to application sender.
The basic RSVP-RESV messages [RFC2205] are transmitted upstream from receiver to sender to provide resource reservation along with "Label Distribution". In this specification, hop-by-hop scheduling is extended to support both resource reservation and label distribution. The current specification is only defined for unicast point-to-point traffic flows, i.e., Fixed Filter (FF) reservation style.
The format of RESV message with the support of 6tisch scheduling capabilities (6P and SF1) is as follows :
<Resv Message> ::= <Common Header> [ <INTEGRITY> ]
[ [<MESSAGE_ID_ACK> | <MESSAGE_ID_NACK>] ... ]
[ <MESSAGE_ID> ]
<SESSION> <RSVP_HOP>
<TIME_VALUES>
<LABEL>
[<SF1 OPERATION>]
[<6P OPERATION>]
[ <RESV_CONFIRM> ] [ <SCOPE> ]
[ <NOTIFY_REQUEST> ]
[ <ADMIN_STATUS> ]
[ <POLICY_DATA> ... ]
<STYLE> <flow descriptor list>
Upon arrival of the PATH message at an application receiver, the SENDER_TSPEC and ADSPEC objects are interpreted to select the resource reservation parameters. Since the RSVP provides receiver initiated resource reservation setup, the scheduling operation needs to perform upstream from receiver to sender. Subsequently, the reserved resources (bandwidth) are mapped into 6tisch cells through Scheduling Function and corresponding L2-bundle is created. An aggregation of cells is called "bundle"(the directional link to a next-hop neighbor). Every L2-bundle is associated with a global trackID to dynamically adapt the cells "hop-by-hop" to an scheduled instance. In addition, the TrackID is used as a "packet filter" to switch the incoming tracks to outgoing tracks. The receiver will generate the TrackID with the combination of "Source/Destination IP address" and "RPLInstanceID" that is obtained from "SENDER_TEMPLATE/FILTER_SPEC".
next-hop node Receiver
+--------------+ +--------------+
| IPv6 | | IPv6 |
+--------------+ +--------------+
| 6LoWPAN | | 6LoWPAN |
+--------------+ +--------------+
| 6top | | 6top |
+--------------+ +--------------+
| TSCH MAC | | TSCH MAC |
+--------------+ +--------------+
| LLN PHY | | LLN PHY |
+--------------+ +--------------+
| |
| | Rspec:Reserves
| | bandwith
| |
| | SF1:Maps bandwidth to
| | cells
| RESV + 6P Request(TrackID)|
|<------------------------- |
Rspec:Reserves | |
bandwith | |
| |
SF1:Maps bandwidth to| |
cells |6P Response (CellList[..]) |
|-------------------------->|
| |
| |
| |
| 6P confirmation |LABEL SET
| CellList[..]+ Label |label=Channel+Slot
|<--------------------------|
Resv state:"Cell| |
label" | |
| |
| |
| |
Figure 3: Operation of RSVP-RESV message with 6P transactions.
+--------------+ <-Data transmission in end-to-end Track->
| IPv6 | Sender Receiver
+--------------+ | |
| 6LoWPAN | | |
+--------------+ | nodeB |
| 6top | | +----+ |
+--------------+ | | | |
| TSCH MAC | | | | |
+--------------+ | | | |
| LLN PHY | | L2-Bundle | | L2-Bundle |
+--------------+ +----------------+ +---------------+
<--Dedicated cells for each Instance-->
Figure 4: End-to-end cell scheduling with SF1 Scheduling
From RFC[6997], it is noteworthy that application sender that initiates the point-to-point (P2P) traffic is called "Parent node" and application receiver that receives the data is called "Child node". Since the receiver (child node) performs the Scheduling operation upstream towards the sender, "3-step transaction" of 6P protocol needs to be triggered hop-by-hop to schedule the reserved resources(see Fig. 3). Hence, "6P Request" with associated TrackID in metadata field is transmitted in "RESV" message from Receiver to next-hop node. The "NumCells" field in the 6P Request is set to required number of cells whereas receive "CellList" should be empty. Once the outgoing interface of next-hop node receive the "RESV" message, it checks the service request specification(Rspec) and perform the resource reservation. Subsequently, Scheduling Function of next-hop neighbor map the reserved resources into transmit cells. Later, "6P Response" with transmit "CellList"(slotOffset, channelOffset) is downstream to receiver. When the receiver has cells (to receive data) available with the "CellList" in the "6P Response" then "6P Confirmation" with "IANA_6TOP_RC_SUCCESS" is upstream towards next-hop node. Otherwise, "ResvErr" message should send back to the receiver with specific error. Since the cell characteristics(slotOffset,channelOffset)is available in the 6P transactions, the next-hop node will store the "SlotOffset (Timeslot)" as a label to switch the traffic flow to receiver. For the multiple cells(Bundle), a generalized label set is created where each label represents one cell to forward data to receiver. Once the 6P transaction is successful in between next-hop node and receiver, a labeled L2-bundle is created with the associated TrackID. Subsequently, "cell label set" is stored in the Resv state block at the next-hop node. Later, SF1 of "next-hop node" maps the reserved bandwidth to the "receiving cells" to receive the data from its upstream node. The "RESV" message with "6P Request" along with TrackID is transmitted upstream towards sender node. With this, end-to-end Track is installed with a succession of paired L2-bundles(a receive bundle from the previous hop and a transmit bundle to the next hop) for a specific instance from sender to receiver.(See Fig. 4). [RFC3473], cells in paired L2-bundles are used as an implicit labels to label switch the data from Sender to Receiver at the 6top sub-layer.
Whenever, the sender needs to setup a new tunnel that is capable of maintaining resource reservations without double counting(at same intermediate node) the resources with existing tunnel then "RSVP reroute mechanism" need to be initiated [RFC3209]. With this operation, bandwidth can increase/decrease end-to-end in the ongoing tunnel. The detailed explanation of "Reroute mechanism" is explained in [RFC3209].
The detailed explanation of PathErr and ResvErr with different ERROR_SPEC to handle Scheduling and 6P operation errors will be described in later specification.
The Scheduling Function Identifier (SFID) of SF1 is IANA_SFID_SF1(TBD).
IANA is requested to allocate a new Scheduling Function (IANA_SFID_SF1) from the SF space of Scheduling Functions defined in [I-D.dujovne-6tisch-6top-sf0]
TODO
| [RFC3209] | Awduche, D., Berger, L., Gan, D., Li, T., Srinivasan, V. and G. Swallow, "RSVP-TE: Extensions to RSVP for LSP Tunnels", RFC 3209, DOI 10.17487/RFC3209, December 2001. |
| [RFC3473] | Berger, L., "Generalized Multi-Protocol Label Switching (GMPLS) Signaling Resource ReserVation Protocol-Traffic Engineering (RSVP-TE) Extensions", RFC 3473, DOI 10.17487/RFC3473, January 2003. |
| [RFC3945] | Mannie, E., "Generalized Multi-Protocol Label Switching (GMPLS) Architecture", RFC 3945, DOI 10.17487/RFC3945, October 2004. |
| [RFC6553] | Hui, J. and JP. Vasseur, "The Routing Protocol for Low-Power and Lossy Networks (RPL) Option for Carrying RPL Information in Data-Plane Datagrams", RFC 6553, DOI 10.17487/RFC6553, March 2012. |
| [I-D.dujovne-6tisch-6top-sf0] | Dujovne, D., Grieco, L., Palattella, M. and N. Accettura, "6TiSCH 6top Scheduling Function Zero (SF0)", Internet-Draft draft-dujovne-6tisch-6top-sf0-01, March 2016. |
| [I-D.ietf-6tisch-6top-protocol] | Wang, Q. and X. Vilajosana, "6top Protocol (6P)", Internet-Draft draft-ietf-6tisch-6top-protocol-02, July 2016. |
| [I-D.ietf-6tisch-architecture] | Thubert, P., "An Architecture for IPv6 over the TSCH mode of IEEE 802.15.4", Internet-Draft draft-ietf-6tisch-architecture-10, June 2016. |
| [I-D.ietf-detnet-use-cases] | Grossman, E., Gunther, C., Thubert, P., Wetterwald, P., Raymond, J., Korhonen, J., Kaneko, Y., Das, S., Zha, Y., Varga, B., Farkas, J., Goetz, F. and J. Schmitt, "Deterministic Networking Use Cases", Internet-Draft draft-ietf-detnet-use-cases-10, July 2016. |