Minutes IETF103: sfc
minutes-103-sfc-01

Meeting Minutes Service Function Chaining (sfc) WG
Title Minutes IETF103: sfc
State Active
Other versions plain text
Last updated 2018-11-10

Meeting Minutes
minutes-103-sfc

   
===============================
Service Function Chaining (SFC)
IETF 103 - Bangkok
Thursday, November 8, 2018
13:50-15:50 (UTC+07:00)
Meeting Minutes
===============================

SFC WG chairs: Joel Halpern, Jim Guichard
SFC secretary: Tal Mizrahi

Meeting minutes: Tal Mizrahi

Chair Slides
------------
Presenter: Joel Halpern
Slides:
https://datatracker.ietf.org/meeting/103/materials/slides-103-sfc-sfc-chairs-slides-00

Summary:
- Note well applies.
- The agenda for the current session was presented.
- WG progress was presented.
- RFC 8459 was published.
- The multi-layer OAM document was adopted. That work needs to be aligned with
the OAM framework.

NSH Encapsulation for In-situ OAM Data (Frank Brockners)
--------------------------------------------------------
Drafts:
        https://tools.ietf.org/html/draft-ietf-sfc-ioam-nsh-00
        https://tools.ietf.org/html/draft-ietf-sfc-proof-of-transit-01

Presentation:
https://datatracker.ietf.org/meeting/103/materials/slides-103-sfc-nsh-encapsulation-for-in-situ-oam-data-00

Summary:
- IOAM over NSH: the draft is pretty stable.
- Looking for any last comments before requesting WG LC.
- Proof of transit: draft changed based on discussion in IETF 102. Now includes
in-order preservation. - We have the SSSS with order preservation, and we have
the HMAC. There is a question of whether we should have both, or define only
one. - We will take it to the list.

Discussion:
- Greg mirsky: no point in having both options.
- Kent liang: what is the value of being in order. This will help understand
what we want. Home much value is there in option 2? - Frank: with SSSS
(Shamir's Secret) you are not forced to do the order-preservation trick. In
some cases we do not need order preservation, and can save the trouble. - Joel:
if it is not expensive, you can receive order preservation.

Active OAM for Service Function Chains in Networks (Greg Mirsky)
----------------------------------------------------------------
Draft: https://tools.ietf.org/html/draft-wang-sfc-multi-layer-oam-12

Presentation:
https://datatracker.ietf.org/meeting/103/materials/slides-103-sfc-active-oam-for-service-function-chains-in-networks-00

Summary:
- The name of the draft has changed, since it was adopted.
- An overview of the changes was presented.

Discussion:
- Adrian Farrel: two scenarios: O-bit is clear and the protocol is OAM, or
O-bit is set and protocol is not OAM. Which of these scenarios is an error? -
Greg: we need to consider for each case whether we consider this an error or
not. - Adrian: I believe the second should be considered an error. - Frank
Brockners: if the next protocl tells you that you have OAM, what is the value
of setting the O bit or not? - Joel: the way we define it, SFF should not look
beyond the network service header. That is what the O bit is for. - Frank: got
it.

Performance Measurement with Alternate Marking (Greg Mirsky)
------------------------------------------------------------
Draft: https://tools.ietf.org/html/draft-mirsky-sfc-pmamm-06

Presentation:
https://datatracker.ietf.org/meeting/103/materials/slides-103-sfc-performance-measurement-with-alternate-marking-00

Summary:
- Draft now suggests to use a single marking bit, replacing a bit that is
currently not used in the NSH. - Would like to request working group adoption.

Hybrid Two-Step Performance Measurement Method (Greg Mirsky)
------------------------------------------------------------
Draft: https://tools.ietf.org/html/draft-mirsky-ippm-hybrid-two-step-02

Presentation:
https://datatracker.ietf.org/meeting/103/materials/slides-103-sfc-hybrid-two-step-performance-measurement-method-00

Discussion:
- Sumandra Majee: follow-up packets - does it mean the packet has to be kept by
the devices? - Greg: there can be only one outstanding packet at any given
time. - Sumandra: the follow-up must follow the trigger packet immediately? -
Greg: you cannot guarantee that it is immediately, but shortly afterwards.

Network Service Header (NSH) Explicit Congestion Notification (ECN) Support
(Donald Eastlake)
---------------------------------------------------------------------------------------------
Draft: https://tools.ietf.org/html/draft-eastlake-sfc-nsh-ecn-support-01

Presentation:
https://datatracker.ietf.org/meeting/103/materials/slides-103-sfc-network-service-header-nsh-explicit-congestion-notification-ecn-support-00

Summary:
- Quick overview, since this was presnted in the last IETF meeting.

Discussion:
- Jeff Tantsura: I support adoption.
- Joel: authors should send an email to the list and request adoption.

Geneve applicability for service function chaining (Sami Boutros)
-----------------------------------------------------------------
Draft:
https://tools.ietf.org/html/draft-boutros-nvo3-geneve-applicability-for-sfc-02

Presentation:
https://datatracker.ietf.org/meeting/103/materials/slides-103-sfc-geneve-applicability-for-service-function-chaining-00

Summary:
- The draft was presented.

Discussion:
- Kent Liang: you mentioned two options. One of the options was to function as
SFF. - Sami: actually NVEs will act as SFF in both options. In a service
topology in NVO3 domain. NVE is actually an SFF. NVE is mapped to a service
function. Two levels of encapsulation. - Jim: first option you use the NSH
header for forwarding decision. Second option - not. - Joel: same tradeoff as
always: state is in the device, or state is in the header. From and SFC
perspective we do not care how the table gets populated. - Kent: you have hop
by hop information. But you don't have to use SPI/SI. - Joel: if the transport
has enough information, you can use the transport header to take the forwarding
decision.

SFC Path OAM (Ting Ao)
----------------------
Drafts:
        https://tools.ietf.org/html/draft-ao-sfc-oam-return-path-specified-02
        https://tools.ietf.org/html/draft-ao-sfc-oam-path-consistency-03

Presentation:
https://datatracker.ietf.org/meeting/103/materials/slides-103-sfc-sfc-path-oam-00

Summary:
- Authors are requesting WG adoption.

Discussion:
- Joel: send your request to the list.


Service Function Chaining: Subscriber and Policy Identification (Dirk Von-Hugo)
-------------------------------------------------------------------------------
Draft: https://tools.ietf.org/html/draft-sfc-serviceid-header-01

Presentation:
https://datatracker.ietf.org/meeting/103/materials/slides-103-sfc-service-function-chaining-subscriber-and-policy-identification-00

Discussion:
- Joel: I thank the authors for this significantly improved version of the
document. I would like to see more discussion on the mailing list.

Name-Based Service Function Forwarder (nSFF) component within SFC framework
(Debashish Purkayastha)
---------------------------------------------------------------------------------------------------
Draft: https://tools.ietf.org/html/draft-trossen-sfc-name-based-sff-01

Presentation:
https://datatracker.ietf.org/meeting/103/materials/slides-103-sfc-name-based-service-function-forwarder-nsff-component-within-sfc-framework-01

Discussion:
- Joel: the interaction between nSFF and NR is a new protocol that you are
inventing? - Debashish: yes.

Adjourned at 15:14.