10
1 draft-pentland-dna-protocol3- draft-pentland-dna-protocol3- 00.txt 00.txt Brett Pentland

11 draft-pentland-dna-protocol3-00.txt Brett Pentland

Embed Size (px)

Citation preview

Page 1: 11 draft-pentland-dna-protocol3-00.txt Brett Pentland

11

draft-pentland-dna-protocol3-00.txtdraft-pentland-dna-protocol3-00.txt

Brett Pentland

Page 2: 11 draft-pentland-dna-protocol3-00.txt Brett Pentland

2

Agenda

Document description Open Issues Consensus call

Page 3: 11 draft-pentland-dna-protocol3-00.txt Brett Pentland

3

Background

Based on two prior proposals draft-pentland-dna-protocol-01.txt draft-jinchoi-dna-protocol2-01.txt

Only one fast RA scheme Link Identification schemes merged

Page 4: 11 draft-pentland-dna-protocol3-00.txt Brett Pentland

4

Fast RA

Taken directly from draft-pentland-dna-protocol-01.txt

Token = A Token = B Token = C

Token = S

Say that: B XOR S < A XOR S < C XOR S

Access Routers

Host

B sends an RA immediately

A sends an RA after 20 ms

C sends an RA after 40 ms

RS

Page 5: 11 draft-pentland-dna-protocol3-00.txt Brett Pentland

5

Link Identification

Routers listen for all prefixes on a link In general, all prefixes included in RAs

• New option for learned prefixes One noted as LinkID prefix

• MUST be included in RAs

Hosts SHOULD include a landmark in RSs If the landmark is on the link, routers may send an abbreviated

RA containing only the landmark If the landmark is not on the link or there’s no landmark routers

send a Complete RA

Unsolicited RAs MAY carry a subset of prefixes but MUST include the LinkID prefix

Page 6: 11 draft-pentland-dna-protocol3-00.txt Brett Pentland

6

Link Identification

B AA B DC

ARs

APs

Host

A,B A,B A,B

D C

RA(C) RA(D)

RS(A?)

RA(A) RA(A,B)RA(C) RA(C,D)

RA(A-Yes)RA(A-Yes)

RA(A-No,C,D) RA(A-No,C,D)

RS(A?)

C,D

Page 7: 11 draft-pentland-dna-protocol3-00.txt Brett Pentland

7

Open Issues

Identification based on non-prefix information No supporting emails on mailing list

Explicit link identifiers Some other discussion suggests there is interest

TSLLAO Needed for unicast RS/RA w/o NS/NA Where does this belong?

Delivery of MLD packets during DNA? presentation to follow

Page 8: 11 draft-pentland-dna-protocol3-00.txt Brett Pentland

8

Open issues (cont)

Packet delivery to/from node during DNA i.e. between the RS and RA What should hosts do with queued packets?

Do we need separate Y/N flags for the Landmark Option? Probably not.

Page 9: 11 draft-pentland-dna-protocol3-00.txt Brett Pentland

9

Possible LPO format

Use a flag to mark the first prefix as the LinkID

Use a further flag to mark it as “not a prefix”

Need host processing rules

0 1 2 3 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Length |L|P| Reserved | Prefix Len 1 | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | ... | Prefix Len N | Padding | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | | + Prefix 1 + | | + + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | | + Prefix 2 + | | + + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ ~ ... +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | | + + | | + Prefix N + | | + + | | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

Page 10: 11 draft-pentland-dna-protocol3-00.txt Brett Pentland

10

Where to now?

Is this close enough to the mark to accept as a WG draft?