XCAST team report Yuji IMAI (WIDE Project) 1.RFC5058 issued on Nov 2007. “Explicit Multicast...

6
XCAST team report Yuji IMAI (WIDE Project) 1. RFC5058 issued on Nov 2007. “Explicit Multicast (Xcast) Conce pts and Options” issued.” What we’ve achieve, at last? 2. Next step XCAST as a building block of SAM XCAST 2.0 IRTF SAM RG @71st IETF

Transcript of XCAST team report Yuji IMAI (WIDE Project) 1.RFC5058 issued on Nov 2007. “Explicit Multicast...

Page 1: XCAST team report Yuji IMAI (WIDE Project) 1.RFC5058 issued on Nov 2007. “Explicit Multicast (Xcast) Concepts and Options” issued.” What we’ve achieve,

XCAST team reportYuji IMAI (WIDE Project)

1. RFC5058 issued on Nov 2007.• “Explicit Multicast (Xcast) Concepts and Opti

ons” issued.”• What we’ve achieve, at last?

2. Next step• XCAST as a building block of SAM• XCAST 2.0

IRTF SAM RG@71st IETF

IRTF SAM RG@71st IETF

Page 2: XCAST team report Yuji IMAI (WIDE Project) 1.RFC5058 issued on Nov 2007. “Explicit Multicast (Xcast) Concepts and Options” issued.” What we’ve achieve,

RFC5058What we’ve achieved, at last?

• Not achieved– Good usage, applications– Industrial products & services– Broad deployment

• Achievement– During discussion in SAM-RG

• There are requirements, XCAST will not solve solely.

– Get rough consensus discussing w/ RFC-editors & IESG• This is excellent and potentially very useful work. However, I would l

ike to discuss how to move forward with regards to the proposed allocation of ... [ID-tracker, June 2007]

• Design became old style for modern hardware routers. [RFC-editorial board]

Page 3: XCAST team report Yuji IMAI (WIDE Project) 1.RFC5058 issued on Nov 2007. “Explicit Multicast (Xcast) Concepts and Options” issued.” What we’ve achieve,

XCAST as a building block of SAMPros. & Cons.

• XCAST isn’t good– No built-in signaling– No bridging for other

mcast variations, yet.– No traffic adaptation m

echanism

• XCAST is good– Keeping very huge

number of group– Transmitter can select

arbitrary sub-set of the group members of individual datagram

– providing basic communication mechanisms necessary to realize ALM, OLM.

Page 4: XCAST team report Yuji IMAI (WIDE Project) 1.RFC5058 issued on Nov 2007. “Explicit Multicast (Xcast) Concepts and Options” issued.” What we’ve achieve,

OS (Windows / Mac / Linux / FreeBSD.. )

SAM Applications

XCAST

SAMTK Group Web Server(Apache / PHP)

SAMTK Core Module

XCAST in SAMTK Architecture

Protocol Interface

Group InterfaceGroup Management

Module

Application Interface

HTTP/XML

Hybrid OverlayPlugin

XCAST Plugin ALM

Plugin

Native mcast UnicastAMT

Page 5: XCAST team report Yuji IMAI (WIDE Project) 1.RFC5058 issued on Nov 2007. “Explicit Multicast (Xcast) Concepts and Options” issued.” What we’ve achieve,

XCAST 2.0<draft-ug-xcast20-protocol-spec-00.txt >

• XCAST6 (version 2.0) Protocol Specification, Feb 2008

– Submitted after RFC5058 issued– Modified points

i. Move semi-permeable tunnel info from Hop-by-hop option to DSCP field of IPv6– Option header becomes performance neck w/ modern

hardware routes.

ii. Fixed length list of destinationiii. Describe to clarify the detailed info for interoperation more

than RFC5058.

– Requirement for additional functions to support Hybrid Overlay framework is welcome!

Page 6: XCAST team report Yuji IMAI (WIDE Project) 1.RFC5058 issued on Nov 2007. “Explicit Multicast (Xcast) Concepts and Options” issued.” What we’ve achieve,

XCAST 2.0Implementation

• Operate system first, not spec first!!

• Firstly, with OSS– FreeBSD 6.X, Linux 2.6.XX– XCAST fan club will change the base from

XCAST6-1.0 to 2.0, soon.

• XCAST 2.0 code will release integrated with SAM-TK

“Code then spec not Spec then Code”

Motto of Itojun