80 likes | 107 Views
IETF-57 PWE3 Workgroup 16-July-03 draft-vainshtein-cesopsn-06.txt: Summary of Discussion. Sasha Vainshtein, Tim Frost, Prayson Pate. CESoPSN-06 Status (1). The CESoPSN draft : First presented at the 52th IETF meeting Evolved: To cover additional functionality
E N D
IETF-57PWE3 Workgroup16-July-03draft-vainshtein-cesopsn-06.txt:Summary of Discussion Sasha Vainshtein, Tim Frost, Prayson Pate
CESoPSN-06 Status (1) • The CESoPSN draft: • First presented at the 52th IETF meeting • Evolved: • To cover additional functionality • To remove unnecessary restrictions • Fully aligned with the following basic documents: • The WG Charter • The PWE3 Requirements draft - all the relevant requirements covered, non-goals not pursued • The PWE3 Architecture draft (payload types, partitioning between NSP and PW IWF, the principle of minimum intervention) • The TDM PW requirements draft (all the mandatory requirements covered) • RFC 2736 (explicitly referenced in the WG Charter) • The SONET Emulation (CEP) draft
CESoPSN-06 Status (2) • The -06 revision of the CESoPSN draft: • A few updates from the -05 version: • Support of trunk-specific Nx64 kbit/s with CAS has been defined • The CW has been fully aligned with CEP • Request to adopt as a WG item as the basis for further WG work posted 06/10/03 • 4 successful interoperability tests have been publicly announced • Prove simplicity and robustness of the proposed solution • Multiple silicon implementations are now available, including a high-density solution • Danny has asked for the WG opinion 06/12/03 • A lively discussion has been maintained on the WG mailing list • A few TECHNICAL issues have been raised by both the supporters and the opponents • Suggested resolutions to these issues are presented here
Unstructured Services • No conceptual objections raised to the CESoPSN for emulation of UNSTRUCTURED services • A new proposal for unstructured service (draft-stein-pwe3-stpp-00): • Conceptually identical to unstructured mode in CESoPSN-06 • See also draft-vainshtein-pwe3-ucesopsn-00.txt (presented in Atlanta and incorporated into CESoPSN-05/06) • The following minor differences found: • No option to use RTP • Different control word format • Different default payload length/latency for E1 • No showstoppers!
Specific Technical Issues - 1 • Issue: Non-ECMP-compliant CW • The CESoPSN CW is fully aligned with that of CEP • It is technically possible to use a standard CW: • Sufficient Flags space (4 flags needed) • Sufficient Sequence Number space • Use Fragmentation bits for splitting basic structures of trunk-specific Nx64 kbit/s with CAS (as per draft-ietf-pwe3-fragmentation-02) • Alignment with CEP has certain advantages • Suggested resolution: follow the WG input 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 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ |0|0|0|0|R|D|N|P|B|E| Reserved | Sequence Number | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ An ECMP-compliant CESoPSN CW
Specific Technical Issues - 2 • Issue: Usage of RTP header • Claims made on the PWE3 list: • RTP does not provide any advantages for clock recovery with emulated TDM • G.823 / G.824-compliant clock recovery is impossible with TDM emulated over PSN, because IP and MPLS do not provide clock at their physical layer • Contention: RTP can facilitate G.823/G.824-compliant clock recovery when a common clock source is available at both PEs terminating the TDM PW • The common clock can be provided using BITS, GPS etc. (same as with CEP-EPAR) • The timestamps are generated in accordance with this clock (differential mode) • Suggested resolution: • Preserve the OPTION to use RTP header • Follow the WG input regarding the default mode
Specific Technical Issues - 3 • Issue: Default packet latency for unstructured E1 • The draft proposes 1 ms • An alternative proposal: 0.5 ms • Suggested resolution: follow the WG input • Issue: Frame alignment for unstructured services • The draft does NOT require or imply that • Suggested resolution: editorial (explicitly state in the text) • Issue: T1-SF Nx64 kbit/s w/CAS • Basic structure for trunk-specific Nx64 kbit/s with CAS defined as the trunk multiframe • De-facto basic structure used with T1-SF: a pair of consecutive multiframes • Suggested resolution: editorial (update the corresponding portions of the text)