Skip to main content

Analysis for FlexE control model
draft-wang-ccamp-flexe-control-analysis-00

The information below is for an old version of the document.
Document Type
This is an older version of an Internet-Draft whose latest revision state is "Expired".
Authors Qilei Wang , xiaobingNIU , Yunbin Xu
Last updated 2019-03-11
RFC stream (None)
Formats
Stream Stream state (No stream defined)
Consensus boilerplate Unknown
RFC Editor Note (None)
IESG IESG state I-D Exists
Telechat date (None)
Responsible AD (None)
Send notices to (None)
draft-wang-ccamp-flexe-control-analysis-00
Internet Engineering Task Force                             Q. Wang, Ed.
Internet-Draft                                               X. Niu, Ed.
Intended status: Informational                           ZTE Corporation
Expires: September 12, 2019                                        Y. Xu
                                                                   CAICT
                                                          March 11, 2019

                    Analysis for FlexE control model
               draft-wang-ccamp-flexe-control-analysis-00

Abstract

   This document gives some analysis about the control of FlexE.

Status of This Memo

   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 https://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 September 12, 2019.

Copyright Notice

   Copyright (c) 2019 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
   (https://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.

Wang, et al.           Expires September 12, 2019               [Page 1]
Internet-Draft                FlexE control                   March 2019

Table of Contents

   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . .   2
   2.  Terminology . . . . . . . . . . . . . . . . . . . . . . . . .   3
     2.1.  Requirements Language . . . . . . . . . . . . . . . . . .   3
   3.  Analysis  . . . . . . . . . . . . . . . . . . . . . . . . . .   3
     3.1.  General Introduction of FlexE . . . . . . . . . . . . . .   3
       3.1.1.  FlexE Group . . . . . . . . . . . . . . . . . . . . .   3
       3.1.2.  FlexE Client  . . . . . . . . . . . . . . . . . . . .   3
       3.1.3.  Adaptation function between FlexE Client and FlexE
               Group . . . . . . . . . . . . . . . . . . . . . . . .   4
       3.1.4.  MAC Frame . . . . . . . . . . . . . . . . . . . . . .   5
       3.1.5.  Adaptation between MAC frames and FlexE Client  . . .   5
     3.2.  General requirements  . . . . . . . . . . . . . . . . . .   5
       3.2.1.  Configuration of FlexE group  . . . . . . . . . . . .   5
       3.2.2.  Allocate Resources for Client MAC flows . . . . . . .   6
     3.3.  FlexE Client Configuration Procedures . . . . . . . . . .   6
     3.4.  Control Requirements Derived  . . . . . . . . . . . . . .   8
   4.  Summary . . . . . . . . . . . . . . . . . . . . . . . . . . .   9
   5.  Acknowledgements  . . . . . . . . . . . . . . . . . . . . . .   9
   6.  IANA Considerations . . . . . . . . . . . . . . . . . . . . .   9
   7.  Security Considerations . . . . . . . . . . . . . . . . . . .   9
   8.  References  . . . . . . . . . . . . . . . . . . . . . . . . .   9
     8.1.  Normative References  . . . . . . . . . . . . . . . . . .   9
     8.2.  Informative References  . . . . . . . . . . . . . . . . .  10
   Authors' Addresses  . . . . . . . . . . . . . . . . . . . . . . .  10

1.  Introduction

   OIF published the first version of FlexE Implementation Agreement in
   March 2016, aiming to provide a generic mechanism for supporting a
   variety of Ethernet MAC rates that may or may not correspond to any
   existing Ethernet PHY rate.  SG15 in ITU-T has endorsed the OIF FlexE
   data plane and parts of [ITU-T G.872], [ITU-T G.709], [ITU-T G.798]
   and [ITU-T G.8023].  The Recommendations depend on or are based on
   the FlexE data plane.

   This draft is intended to trigger discussion of the FlexE control
   architecture according to the analysis in section 2.  What kind of
   architecture should we employed when configuring FlexE equipments,
   how to configure the FlexE group and FlexE client, and what kind of
   parameters do we need to take into consideration?  The analysis is
   mainly based on the description in section 7 and 8 of [ITU-T G.8023],
   which is "Characteristics of equipment functional blocks supporting
   Flex Ethernet interfaces".

Wang, et al.           Expires September 12, 2019               [Page 2]
Internet-Draft                FlexE control                   March 2019

2.  Terminology

2.1.  Requirements Language

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
   document are to be interpreted as described in RFC 2119 [RFC2119].

3.  Analysis

3.1.  General Introduction of FlexE

   The FlexE shim is built into the Ethernet PCS (physical coding
   sublayer).  If a FlexE group is set up, a corresponding n*100G (or
   n*200G, n*400G) PCS module is created as well.

   The difference between the FlexE and the traditional 100G Ethernet is
   that the traditional Ethernet PCS has a 1:1 relationship with the
   client MAC flow, while with FlexE one bonded huge PCS module can be
   used to transport more than one client MAC flow i.e., the
   relationship is 1:n.

3.1.1.  FlexE Group

   A FlexE Group is consisted of 1 to m bonded Ethernet PHYs.  The rate
   of the Ethernet PHY could be 100G, 200G or 400G.  All PHYs in the
   group must operate at the same rate.

   FlexE group is consisted of a number of PHYs, and each PHY is
   consisted of 66B blocks stream.  Section monitoring overhead is
   added/extracted as one 66B block at the FlexE group source and
   destination (i.e., trail termination) to determine the status of the
   FlexE group (i.e., FlexE trail).  Currently, only RPF (Remote PHY
   Fault) indication is used to report the state of FlexE group.

   One FlexE group exists between two FlexE shim, there is no switching
   defined in FlexE.  In addition, only one fault indication is defined,
   there is no other OAM function developed yet.  Based on these
   analysis, we may understand FlexE is just an interface technology,
   and once a FlexE group is configured, it only functions as one
   Ethernet link, same as Ethernet PHY.

3.1.2.  FlexE Client

   A FlexE Client is an Ethernet flow based on a MAC data rate that may
   or may not correspond to any Ethernet PHY rate.  The FlexE Client MAC
   rates supported by a FlexE Groups could be 10, 40, and m*25 Gb/s.
   The FlexE Client MAC rates supported by FlexE Groups may support all,

Wang, et al.           Expires September 12, 2019               [Page 3]
Internet-Draft                FlexE control                   March 2019

   or only a subset of these FlexE Client rates, e.g., m*25 Gb/s.  Each
   FlexE Client is presented to the FlexE Shim as a 64B/66B encoded bit-
   stream according to clause 82 of [IEEE 802.3].

   According to the description in clause 8.1 of [ITU-T G.8023], there
   is no overhead defined for monitoring a FlexE client, so the trail
   for FlexE client in the equipment does not exist.  The FlexE client
   trail termination function is a null function.  Therefore, there is
   no need to model FlexE client as a layer.

3.1.3.  Adaptation function between FlexE Client and FlexE Group

   In order to distribute the FlexE client over PHYs of one FlexE group,
   a number of management information command should be sent to the
   adaptation function which performs the mapping of FlexE client over
   FlexE group.

   According to the description in clause 7.2 of [ITU-T G.8023], the
   external management information command sent to the source adaptation
   function is listed below:

      TxCC, TxCCA, TxCCB, TxCR, TxCA

      TxGID, TxPHYMAP

   The TxCC, TxCCA and TxCCB are used to configure the calendar for use,
   which could be type A or type B calendar configuration, and FlexE
   client number.

   TxCR and TxCA are used to coordinate the switch of calendar
   configuration between the FlexE source and destination node.

   The TxGID is used to configure the FlexE group identifier.  The
   TxPHYMAP is used to configure the set of PHYs in the FlexE group.

   The built-in function multiplexer performs the action of assigning
   the individual FlexE Client to specific calendar slots of the FlexE
   group.

   At the destination side, the overhead should be extracted first to
   compare the GID and PID.  The Demultiplexer function activates the
   FlexE Client and assigns the calendar slots of the FlexE group
   payload area to the individual FlexE client as defined by the client
   calendar information carried in the overhead.

Wang, et al.           Expires September 12, 2019               [Page 4]
Internet-Draft                FlexE control                   March 2019

3.1.4.  MAC Frame

   Defined in IEEE.

3.1.5.  Adaptation between MAC frames and FlexE Client

   It can be seen from the Figure 8-6 of [ITU-T G.8023] that the
   external management information commands used as input to the
   adaptation function are defined by [IEEE 802.3].  The [IEEE 802.3]
   process mainly includes the 64B/66B encoding, as well as MAC frame
   check sequence generation and frame counting.  The FlexE client
   stream is generated at the determined FlexE Client MAC rate and
   64B/66B encoded.

3.2.  General requirements

   It can be inferred from section 2.1.2 and section 2.1.5 that process
   mainly involved when producing the FlexE Client from MAC frames is
   64b/66b encoding, and this encoding has already been defined by [IEEE
   802.3].  No extra overhead is added.  Therefore, configuration for
   FlexE client layer is needed.  Based on the above analysis, two
   general requirements for control/management of FlexE are considered
   in this draft.

      Configuration of FlexE group

      Allocation of one or more FlexE group calendar slot resources to a
      client MAC flow.

3.2.1.  Configuration of FlexE group

   It can be concluded from the above analysis that external
   configuration tools should be involved to bring one FlexE group into
   service.  The initial configuration commands can come from external
   management system, SDN controller etc.

   A FlexE group must be configured first before any client signals are
   carried over it.  When a new FlexE Group is brought into service, the
   initial configuration must be provisioned from both ends, and the
   initial configuration must be the same.  The group is configured to
   consist of from 1 to n 100G FlexE Instances carried over from 1 to m
   PHYs of the same rate (100GBASE-R, 200GBASE-R, or 400GBASE-R).  Each
   PHY tunnel may consist of multiple hops.

Wang, et al.           Expires September 12, 2019               [Page 5]
Internet-Draft                FlexE control                   March 2019

3.2.2.  Allocate Resources for Client MAC flows

   The FlexE client MAC flows are encapsulated in one or more FlexE
   calendar slots.  Questions that may be raised when considering
   whether external control/management tools are needed.

      How is the bandwidth (number of calendar slots) allocated to a MAC
      client?

      How are the calendar slots assigned to each FlexE clients?

      Does the external management/control system need to be involved?

   According to the analysis in section 2.1.3, it can be inferred that
   the built-in multiplexer and demultiplexer function can work together
   to insert/extract FlexE Client stream from some calendar slots
   correctly, as well as the trail termination function for FlexE client
   is a null function.  Therefore, only the bandwidth information of the
   FlexE Client is needed, the number of calendar slots required can be
   derived from the bandwidth information.

   The FlexE client physical port is an internal port which only perform
   the function of encapsulating upper layer packets into MAC frames,
   64b/66b encoding.  The bandwidth capability of these internal ports
   should be known by external management/control tools in order to
   multiplex and demultiplex the upper layer flow correctly.

3.3.  FlexE Client Configuration Procedures

   Example below is depicted to set up a 25G MPLS-TP P2P path (from
   MPLS-TP-1 to MPLS-TP-2) over one FlexE link[two FlexE Groups] between
   Route-1 and Router-2.  This is to help understand why control of
   FlexE client is not needed.  FlexE group is assumed configured in
   this case.

Wang, et al.           Expires September 12, 2019               [Page 6]
Internet-Draft                FlexE control                   March 2019

    ==================================================================

MPLS-TP-1
+-----+                                                           MPLS-TP-2
|     |                                                           +-----+
|     |                        XXXXXX                             |     |
+-----+                     XXXX    XXXXXX                        |     |
      |                    X             XXXX                     +-----+
      |             OTN   X                 XX   OTN              |
      +-+-----+   +-----+X                    X+-----+    +-----+-+
        |     +---+     +----------------------+     +----+     |
        |     +---+     +----------------------+     +----+     |
        +-----+   +-----+ XX                 X +-----+    +-----+
        Router-1           XXX             XXX            Router-2
                              XXXXXX XXXX XX

    ==================================================================

                        Figure 1: Network Scenario

   Configuration needed for each node when setting up MPLS-TP path:

      Outgoing MPLS label and output port, 25G --> MPLS-TP-1;

      Incoming MPLS label and input port, 25G --> Router-1;

      MPLS traffic over FlexE link, from Router-1 to Router-2;

      Outgoing MPLS label and outgoing port, 25G --> Routing-2;

      Incoming label and input port, 25G --> MPLS-TP-2

   What else would be done by Router-1 and Router-2 if we want to put
   MPLS packets over FlexE link?

   Router-1, perform the generally used data link (i.e., MAC)
   encapsulation procedures to transmit the packet to next hop Router-2,
   which includes:

      encapsulation of the packets into MAC frames.  The source and
      destination MAC address can be derived by searching Router-1's
      internal mapping table, i.e., (outgoing label, outgoing port, MAC
      address) table, then do the 64b/66b encoding for these MAC frames.

Wang, et al.           Expires September 12, 2019               [Page 7]
Internet-Draft                FlexE control                   March 2019

      in addition, Router-1 would announce the required bandwidth
      required from Router-1 to Router-2 to local FlexE multiplexer, the
      local FlexE multiplexer then allocates 5*5G calendar slots to the
      MPLS-TP client signal, and insert FlexE overhead onto each PHY.

      Router-1's configuration is finished.

   Router-2's configuration would be different, as its function is to
   extract the FlexE client signal from the calendar slots according to
   the information carried in the FlexE overhead and route the packets
   based on the MPLS label.  As there is no switching of calendar slots
   in FlexE, the calendar slots allocated to certain client signal would
   not change.

   It can be inferred from the above procedures that the configuration
   of a MPLS client is not impacted by the use of FlexE comparing to
   traditional Ethernet.

3.4.  Control Requirements Derived

   a.  Using control plane method to configure FlexE group, which may
       include the configuration of group number, PHY number and
       correlation between logical PHY number and physical port number.

   b.  Configuration of "logical" PHY, which includes initial
       configuration of bandwidth (number of calendar slots); change of
       bandwidth while in service.

   c.  External control command should be provide to trigger the switch
       of calendar slots.

   d.  Interworking between 5G slot granularity capable node and 25G
       slot granularity node.

   e.  Configuration of aware case.  As calendar slots is not visible to
       external management/control tools.  Bandwidth information of the
       Ethernet PHYs selected can be used to infer the unavailable slot
       information, as the unavailable slots are placed at the end of
       each relevant sub-calendar (the highest numbered slots).
       [RFC6002] defines a new switching type DCSC (Data Channel
       Switching Capable) to describe the switching of whole digital
       channel presented on single channel interface, which can describe
       the Ethernet terminated at the physical (port) level and all
       traffic received on a port is switched to a physical port at the
       LSP egress.  The FlexE digital channel presented at each PHY
       interface can be described with DCSC switching type.  However,
       FlexE aware case may not be depicted with DCSC, as described in
       section 17.12 of [ITU-T G.709], each FlexE (sub) groups (i.e.,

Wang, et al.           Expires September 12, 2019               [Page 8]
Internet-Draft                FlexE control                   March 2019

       each 100G FlexE signals) are crunched, padded and interleaved in
       FlexE, the bandwidth of digital channel links presented in one
       tunnel may varies.  Therefore, a new switching type DCSC-flexe is
       needed in aware case to depict the end-to-end tunnel with
       bandwidth varies at each digital channel links.  The
       configuration of FlexE instance and unavailable slot information
       can be derived through the bandwidth of each hop.

   Different kinds of alarms should be taken into consideration when
   modelling FlexE technology, which may include PHY failed, skew exceed
   threshold, inconsistent configuration between two ends.

4.  Summary

   According to the analysis in section 2, the main control/management
   requirement for FlexE technology is to configure the FlexE group.
   Once a FlexE group is configured and the capability information of
   the internal FlexE client ports associated with this FlexE group is
   known, use of the FlexE technology is the same as that in traditional
   Ethernet.

5.  Acknowledgements

6.  IANA Considerations

   This memo includes no request to IANA.

7.  Security Considerations

   None.

8.  References

8.1.  Normative References

   [ITU-T_G709]
              ITU-T, "ITU-T G.709: Optical Transport Network Interfaces;
              07/2016",  http://www.itu.int/rec/T-REC-
              G..709-201606-P/en, July 2016.

   [ITU-T_G798]
              ITU-T, "ITU-T G.798: Characteristics of optical transport
              network hierarchy equipment functional blocks", August
              2018.

Wang, et al.           Expires September 12, 2019               [Page 9]
Internet-Draft                FlexE control                   March 2019

   [ITU-T_G8023]
              ITU-T, "ITU-T G.8023: Characteristics of equipment
              functional blocks supporting Ethernet physical layer and
              Flex Ethernet interfaces",  , 2016.

   [ITU-T_G872]
              ITU-T, "ITU-T G.872: The Architecture of Optical Transport
              Networks; 2017",  http://www.itu.int/rec/T-REC-G.872/en,
              January 2017.

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

8.2.  Informative References

   [I-D.izh-ccamp-flexe-fwk]
              Hussain, I., Valiveti, R., Pithewan, K., Wang, Q.,
              Andersson, L., Zhang, F., Chen, M., Dong, J., Du, Z.,
              zhenghaomian@huawei.com, z., Zhang, X., Huang, J., and Q.
              Zhong, "GMPLS Routing and Signaling Framework for Flexible
              Ethernet (FlexE)", draft-izh-ccamp-flexe-fwk-00 (work in
              progress), October 2016.

Authors' Addresses

   Qilei Wang (editor)
   ZTE Corporation
   Nanjing
   CN

   Email: wang.qilei@zte.com.cn

   Xiaobing Niu (editor)
   ZTE Corporation
   Beijing
   CN

   Email: niu.xiaobing@zte.com.cn

Wang, et al.           Expires September 12, 2019              [Page 10]
Internet-Draft                FlexE control                   March 2019

   Yunbin Xu
   CAICT
   Beijing
   CN

   Email: xuyunbin@caict.ac.cn

Wang, et al.           Expires September 12, 2019              [Page 11]