160 likes | 287 Views
End-to-End VoMPLS Header Compression (draft-ash-e2e-vompls-hdr-compress-00.txt) End-to-End VoIP Header Compression Using cRTP (draft-ash-e2e-crtp-hdr-compress-00.txt). Jerry Ash AT&T gash@att.com. Jim Hand AT&T jameshand@att.com. Bur Goode AT&T bgoode@att.com.
E N D
End-to-End VoMPLS Header Compression(draft-ash-e2e-vompls-hdr-compress-00.txt)End-to-End VoIP Header Compression Using cRTP (draft-ash-e2e-crtp-hdr-compress-00.txt) Jerry Ash AT&T gash@att.com Jim Hand AT&T jameshand@att.com Bur Goode AT&T bgoode@att.com
Outline(draft-ash-e2e-vompls-hdr-compress-00.txt)(draft-ash-e2e-crtp-hdr-compress-00.txt)Outline(draft-ash-e2e-vompls-hdr-compress-00.txt)(draft-ash-e2e-crtp-hdr-compress-00.txt) • motivation/requirements for VoIP over MPLS (VoMPLS) • brief review of proposals • ‘you read the drafts’ • issues • PWE3 WG charter extension • protocol extensions for cRTP, RSVP-TE, RFC2547 VPNs • resynchronization, performance, & applicability of cRTP/'simple' mechanisms • scalability of E2E VoMPLS applied CE-CE • LDP application as the underlying LSP signaling mechanism
Motivation/Requirements forEnd-to-End VoMPLS Header Compression • motivation • carriers evolving to converged MPLS/IP backbone with VoIP services • enterprise VPN services with VoIP • legacy voice migration to VoIP • requirements • need mechanism for end-to-end VoIP over MPLS CE --> CE header compression • e.g., from CE1 --> PE1 --> P --> PE2 --> CE2 • prior work in MPLS WG by Swallow/Berger on ‘simple’ mechanism (I-D expired) • for efficient voice transport • support encapsulation of voice/RTP/UDP/IP/MPLS • header (uncompressed) = typically 40-48 bytes • voice = typically < 30 bytes • ~60% overhead, 10s of gigabits-per-second for headers alone • support voice encoding (G.729, G.723.1, etc.) • compress/decompress header using cRTP (RFC 2508) and/or ‘simple’ algorithm • ~90% header compression with cRTP • ~50% header compression with ‘simple’ • operate in RFC2547 VPN context • scalable to very large number of CE --> CE flows
Brief Review of ProposalsEnd-to-End VoMPLS Header Compression(draft-ash-e2e-vompls-hdr-compress-00.txt) • steps • use RSVP to establish LSPs between CE1-CE2 • use cRTP (or simple HC) to compress header at CE1, decompress at CE2 • CE1 requests SCIDs from CE2 • CE1 appends CE2 label to compressed header • header compression context routed from CE1 --> PE1 --> P --> PE2 --> CE2 • route compressed packets with MPLS labels CE1 --> CE2 • packet decompressed at CE2 using cRTP (or simple HC) algorithm • advantages • minimizes PE requirements (has to route HC context) • disadvantages • CE’s need to run RSVP, possible scalability issue
Brief Review of ProposalsEnd-to-End VoIP Header Compression Using cRTP (draft-ash-e2e-crtp-hdr-compress-00.txt) • steps • use RSVP to establish LSPs between PE1-PE2 • use cRTP to compress header at CE1, decompress at CE2 • CE1 requests SCIDs from CE2 • header compression context routed from CE1 --> PE1 --> P --> PE2 --> CE2 • PE1 & PE2 create ‘SCID routing tables’ & perform ‘SCID switching’ for compressed packets (SCID --> MPLS labels’) • route compressed packets with MPLS labels PE1 --> PE2 • packet decompressed at CE2 using cRTP algorithm • advantages • minimizes CE requirements (has to route HC context) • disadvantages • additional PE requirements (need to create ‘SCID routing tables’)
Issue 1 - PWE3 WG Charter Extension • VoMPLS header compression not within current charter of the PWE3 WG (or any WG) • why PWE3? • seems the best fit • header compression historically a transport item (PWE3 in transport area) • header compression more a function of application than "base MPLS technology” • current charter mentions work on header compression with RTP: “Specify methods with RTP(and possibly using header compression where needed) to ensure in-order final PDU delivery, perform clock recovery inband emulation and maintenance functions across the PSN, where required.” • what extension needed? • PWE3 not chartered to do work related to signaling PE-PE tunnel (PW signaling/maintenance in scope) • proposals in I-Ds require extensions to RSVP-TE to create tunnels • charter extension needed
Issue 2 - Protocol Extensionsfor cRTP, RSVP-TE, RFC2547 VPNs • extensions proposed to [cRTP] and [cRTP-ENHANCE] • new packet type field to identify FULL_HEADER, CONTEXT_STATE, etc. packets • new objects defined for [RSVP-TE] • extensions proposed to RFC2547 VPNs • create 'SCID routing tables' to allow routing based on the session context ID (SCID) • extensions need coordination with other WGs (MPLS, CCAMP, AVT, ROHC, PPVPN, etc.)
Issue 3 - Resynchronization, Performance, & Applicability of cRTP/’simple' Mechanisms • E2E VoMPLS using cRTP header compression might not perform well with frequent resynchronizations • applicability & performance need to be addressed • 'simple' header compression avoids need for resynchronization • cRTP header compression achieves greater efficiency than ‘simple’ (90% vs. 50% header compression)
Issue 4 - Scalability of E2E VoMPLSApplied CE-CE • E2E VoMPLS applied CE-CE would require a large number of LSPs to be created • concern for CE/PE/P ability to do necessary processing & architecture scalability • processing & label binding at every MPLS node on path between each GW-GW pair • processing every time resource reservation is modified (e.g., to adjust to varying number of calls on a GW-GW pair) • core router load from thousands of LSPs, setup commands, refresh messages
Issue 5 - LDP Application as Underlying LSP Signaling Mechanism • desirable to signal VoMPLS tunnels with LDP • many RFC2547 VPN implementations use LDP as underlying LSP signaling mechanism • may require substantial extensions to LDP • 2 I-D's propose ways for LDP to signal 'VC' (outer) labels for PWs • http://ietf.org/internet-drafts/draft-ietf-pwe3-control-protocol-01.txt • http://www.ietf.org/internet-drafts/draft-rosen-ppvpn-l2-signaling-02.txt • Rosen's I-D suggests ways to do auto-discovery • this together with LDP capability to distribute inner labels might support CE-CE VoMPLS LSPs (within the context of RFC 2547) • other LDP issues • no bandwidth associated with LSPs. • QoS mechanisms limited
Issue 4/5 - LDP vs. RSVP-TE for MPLS LSPs • no solution perfect • LDP • lack of TE, no bandwidth associated with LSPs • QoS mechanisms limited • perhaps too many bytes for an ATM packet • RSVP-TE • lack of scalability • however • LDP • used in many RFC 2547 VPNs • scalable • RSVP-TE • TE allows bandwidth assignment to LSPs • QoS mechanisms
Next Steps • propose Charter extension to PWE3 to include VoMPLS • propose I-D’s be accepted as PWE3 WG documents
VoIP/VoMPLS Control Plane • call control • establishes, modifies, and releases telephone calls • e.g., SIP or H.323 • in distributed VoIP architecture, Call Agents control gateways using MGCP or MEGACO/H.248 • arranges for media gateway to obtain address of terminating GW • determines, through negotiation if necessary, what processing functions the media GW must apply to the media stream • e.g., codec choice, echo cancellation, etc. • connection/bearer control • establishes, modifies, & releases logical connections between gateways • provide fairness in sharing of LSPs • TE provides low-delay, low jitter routes for VoIP • QoS guarantees for existing connections should be unaffected by new voice calls • new calls should be rejected at network edge if insufficient resources are available • network should be resilient to mass calling events
VoIP/VoMPLS Control Plane • interaction between call/connection control needed • BW reservation must occur before called party alerting • new sessions routed to use network efficiently • architecture to accommodate multimedia as well as VoIP