Insight into EPON GPON Sept 1 st 2007

  • Slides: 34
Download presentation
Insight into EPON & GPON Sept. 1 st, 2007 2021/12/17 Fiberhome Proprietary and Confidential

Insight into EPON & GPON Sept. 1 st, 2007 2021/12/17 Fiberhome Proprietary and Confidential

Outline Ø EPON vs. GPON Ø CTC EPON IOP Ø Summary 2021/12/17 Fiberhome Proprietary

Outline Ø EPON vs. GPON Ø CTC EPON IOP Ø Summary 2021/12/17 Fiberhome Proprietary and Confidential 2

EPON Vs. GPON • Protocol & Framing • Qo. S & TDM Support •

EPON Vs. GPON • Protocol & Framing • Qo. S & TDM Support • System Costs • Upgrade Path • Interoperability & service migration • Split ratios, maximum reach, & traffic management • Users Forecast 2021/12/17 Fiberhome Proprietary and Confidential 3

PON Framing ITU-TGPON 125 sec ATM 125 sec GEM ATM 125 sec ATM GEM

PON Framing ITU-TGPON 125 sec ATM 125 sec GEM ATM 125 sec ATM GEM ATM “GPON Lite” GEM IEEE EPON GPON is evolving GEM to look like EPON! GEM OAM & MPCP No Fixed Frame 2021/12/17 Fiberhome Proprietary and Confidential 4

GPON Frame Details Downstream 125 us PCBd n PSync 4 Bytes PCBd n+1 Payload

GPON Frame Details Downstream 125 us PCBd n PSync 4 Bytes PCBd n+1 Payload n Ident 4 Bytes Rate 1. 244 G 2. 488 G 125 us PLOAMd 13 Bytes BIP 1 Byte PLend 4 Bytes Coverage of this BIP (Includes Payload n-1) Payload n+1 Plend 4 Bytes Frame Size 19440 38880 US BW Map N*8 Bytes Coverage of next BIP (Includes Payload n) Upstream ONT 1 Guard Time 2021/12/17 125 us Frame n+1 Gap PLOu PLOAMu 13 bytes Alloc #a SStart ONT 2 DBRu 1 Payload 1 Alloc #a Gap PLOAMu 13 bytes Alloc #b Rate 1. 244 G 2. 488 G ONT n Frame Size 19440 38880 Gap DBRu 1 Payload 1 Alloc #b Fiberhome Proprietary and Confidential 5

Protocol ≠ Qo. S DBA (Out-of-Scope) OAM (In-Scope) Framing (In-Scope) Physical Layer (In-Scope) •

Protocol ≠ Qo. S DBA (Out-of-Scope) OAM (In-Scope) Framing (In-Scope) Physical Layer (In-Scope) • Neither the EPON nor the GPON specification defines the Qo. S mechanism (DBA algorithm); it is out-of-scope, meaning it is up to the system/chip vendor. • EPON and GPON have identical service requirements. High-performance, Qo. S-capable systems can be built with either protocol. Evaluate PON systems on performance and price, not protocol. 2021/12/17 Fiberhome Proprietary and Confidential 6

End-To-End Service Architecture PSTN Core Network Home Network Soft Switch OLT Triple-Play FTTH ONT

End-To-End Service Architecture PSTN Core Network Home Network Soft Switch OLT Triple-Play FTTH ONT Video / IP STB PON System: A L 2/L 3/L 4 Ethernet Switch • Connects the Core & Home Networks • Multi-service • Strict enforcement of service contracts • Designed to reduce end-to-end cost Vo. D Server 2021/12/17 Fiberhome Proprietary and Confidential 7

TDM Services over x. PON • Telco-grade Qo. S is required • Jitter, wander,

TDM Services over x. PON • Telco-grade Qo. S is required • Jitter, wander, delay, Stratum-traceability Apartment Building Cell Site • Must be compatible with triple-play networks • EPON & GPON: identical service-layer requirements for TDM. Copper Network Mgmt Channel Bank Video Network E 1 TDM Network Gb. E ONT n x E 1 Data Network Ethernet Switch ONT Business Optical Line Terminal E 1 ONT Central Office Gb. E PSTN ONT ONT 2021/12/17 Triple-Play Residential Customers Fiberhome Proprietary and Confidential 8

GPON Has More Complex Chips and Buffers • GPON uses GEM to Segment and

GPON Has More Complex Chips and Buffers • GPON uses GEM to Segment and Reassemble Ethernet frames • Each connection (Port-ID) requires a separate SAR buffer • An additional 1 MB external buffer memory is required GPON OLT • 100’s to 1000’s of SAR buffers • Frame from ONU must wait until all bytes are received upstream from ONU before it can be processed GPON ONU • Segmentation buffers for every Port ID 2021/12/17 Fiberhome Proprietary and Confidential 9

EPON Uses Less Expensive Optics – Proven GPON 1 G/2 G/10 G EPON Downstream

EPON Uses Less Expensive Optics – Proven GPON 1 G/2 G/10 G EPON Downstream data rate (Mbps) 1244 or 2488 1000, 2500, 10000 Upstream data rate (Mbps) 155, 622, 1244 1000 Payload encapsulation Laser on/off AGC CDR (Clock Data Recovery) GPON Encapsulation Method (GEM) 13 ns * 44 ns * Ethernet framing 512 ns 400 ns * Short laser on/off times in GPON require high-speed laser drivers * Short AGC intervals in GPON require optical power leveling – Additional protocol to negotiate power level – Digital interface to transceiver to set the values * Relaxed optical specification parameters in EPON less expensive devices 2021/12/17 Fiberhome Proprietary and Confidential 10

Two Very Different Choices Support for Advanced Services 802. 3 ah EPON Continuity of

Two Very Different Choices Support for Advanced Services 802. 3 ah EPON Continuity of Services & Network Management. 100% Ethernet Seamless Migration 2. 5 Gb/s EPON 1. 25 Gb/s EPON X ATM BPON 622 Mb/s 2021/12/17 10 Gb/s EPON 100% Ethernet Seamless Migration New Protocol Forklift Upgrade? 1. 25 Gb/s 2. 5 Gb/s GPON X ? No roadmap beyond 2. 5 G for GPON. Disruption of Services & Network Management. ITU-T GPON 2. 5 Gb/s Speed Fiberhome Proprietary and Confidential 10 Gb/s 11

1. 25 & 2. 5 Gb/s EPON: Line Rates & Framing 1. 25 Gb/s

1. 25 & 2. 5 Gb/s EPON: Line Rates & Framing 1. 25 Gb/s Downstream Line Rate: Data Rate: Line Encoding: MPCP Timing: 1. 25 Gb/s (. 8 ns per bit) 1 Gb/s (1 ns per bit) 8 B/10 B Time Quanta (16 ns units) 64 bits of Preamble 1518 Byte Packet PRE 2. 5 Gb/s Downstream Line Rate: Data Rate: Line Encoding: MPCP Timing: 2. 5 Gb/s (. 4 ns per bit) 2 Gb/s (. 5 ns per bit) 8 B/10 B Time Quanta (16 ns units) 64 bits of Preamble 1518 Byte 2021/12/17 P 1518 Byte Fiberhome Proprietary and Confidential P 1518 Byte P 12

Backward & Forward Compatibility 1. 25 G 1. 25 G 2. 5 G 1.

Backward & Forward Compatibility 1. 25 G 1. 25 G 2. 5 G 1. 25 G 2. 5 G 2021/12/17 Fiberhome Proprietary and Confidential 1. 25 G 13

Progression from 1 G to 10 G EPON 2021/12/17 Fiberhome Proprietary and Confidential 14

Progression from 1 G to 10 G EPON 2021/12/17 Fiberhome Proprietary and Confidential 14

Interoperability & Service Migration “Full” Specification (ITU-T GPON) “Open” Specification (IEEE EPON) Management Layer

Interoperability & Service Migration “Full” Specification (ITU-T GPON) “Open” Specification (IEEE EPON) Management Layer (In-Scope) Management Layer (Out-of-Scope) Why are these different? Services Layer (In-Scope) Services Layer (Out-of-Scope) System Layer (Out-of-Scope) Allows Telcos & OEMs to differentiate products System Layer (Out-of-Scope) “Upper” PON Layer (Out-of-Scope) DBA Algorithm, etc. “Upper” PON Layer (Out-of-Scope) “Lower” PON Layer (In-Scope) Different Objectives Different Scopes 2021/12/17 Fiberhome Proprietary and Confidential 15

Scope of the IEEE 802. 3 Standard IEEE 802. 3 covers only the Physical

Scope of the IEEE 802. 3 Standard IEEE 802. 3 covers only the Physical Layer & part of the Data Link Layer 2021/12/17 Fiberhome Proprietary and Confidential 16

Interoperability: Two World Views World View 1: PON equipment that complies with a complete

Interoperability: Two World Views World View 1: PON equipment that complies with a complete specification, such as ITU-T GPON, is mandatory. • Aspiration: A “complete” specification leads to interoperable equipment from multiple suppliers, leading in turn to lower cost. World View 2: PON equipment that allows transparent re-use of existing IP-based services is mandatory. • Aspiration: Interoperability at the service and management layers with other access systems (e. g. , DSL). 2021/12/17 Fiberhome Proprietary and Confidential 17

Split-Ratio Myths Logical vs physical split-ratio limits • Logical – One BPON OLT can

Split-Ratio Myths Logical vs physical split-ratio limits • Logical – One BPON OLT can address 253 BPON ONUs; – One GPON OLT can address 4 k GPON ONUs; – One EPON OLT can address 32 k EPON ONUs. • Physical – In real deployments, all technologies are limited to 1 x 32 or 1 x 64, depending on reach, condition of the fiber plant, service mix, optical performance. – There is no practical split-ratio limit for any of the PON protocols; all have ample “ONU address space. ” Myth: “EPON is only a 1 x 16 solution, while GPON supports 1 x 128” • Statements like this combine willful mis-reading of the EPON spec, which specifies a minimum split-ratio of 1 x 16, not a maximum split-ratio, with some very simplistic BW utilization calculations. Myth: “GPON has twice the split-ratio” because it’s downstream is twice as fast as EPON’s • 2. 5 G EPON is here and 10 G EPON is coming soon this issue will disappear. • Latency requirements, bandwidth guarantees, and fairness requirements are more important than raw bandwidth. • Stated another way, if solution A has more raw bandwidth than solution B, but cannot distribute that bandwidth with enough precision and accuracy to meet the SLAs, then solution A, and its higher bandwidth, are useless. 2021/12/17 Fiberhome Proprietary and Confidential 18

Traffic Management & Maximum Reach Too much emphasis is placed on the PON protocol

Traffic Management & Maximum Reach Too much emphasis is placed on the PON protocol (EPON vs GPON), and too little attention is paid to the traffic-management and service-level issues. • The magic of EPON is not so much that it uses Ethernet framing (although that does lead to the lowest costs), but rather that Fiberhome has built in the traffic management functions that are really needed to make the services work properly. • High-performance (or low-performance) systems in principle could be built with either protocol, hence manufacturers and carriers should place highest priority on feature set and performance, not details of the framing. Comments on maximum reach • Optics performance, split-ratio, and fiber-plant particulars determine the reach, not the PON protocol. Again there is FUD that confuses minimum requirements in the EPON standard with what is actually achievable (and legal) in real systems. • Basically, you can dial up very long-reach PONs using any of the protocols, provided you are willing to choose the right optics, reduce the split ratio, etc. • None of the framing definitions contain any long-distance magic; it’s all about optics and physics. 2021/12/17 Fiberhome Proprietary and Confidential 19

EPON in Asia, 2007 • Japan: 300 k+ lines/month. • NTT, KDDI, Tepco, K-Opticom,

EPON in Asia, 2007 • Japan: 300 k+ lines/month. • NTT, KDDI, Tepco, K-Opticom, Chubu Electric, Energia, Kintetsu, & many others. • Korea: Now in mass deployment, KT and others, 1 M+ new subscribers in 2007. • China: 50+ EPON deployments currently underway, 400 k+ new subscribers in 2007. • Taiwan, Malaysia, Singapore, Hong Kong, Thailand, Indonesia, Australia, etc. • 6 EPON deployments currently underway (including 3 PTTs). • Cost is key. • Since IEEE 802. 3 ah approval in 2004, EPON equipment costs have decreased by 60+% and optics costs have decreased by 80+%. 2021/12/17 Fiberhome Proprietary and Confidential 20

2004– 2009 PON Subscribers ~ 5 M EPON end CY 2006 Source: Infonetics Metro

2004– 2009 PON Subscribers ~ 5 M EPON end CY 2006 Source: Infonetics Metro Ethernet Equipment, April 2006 With permission: Copyright © 2006 by Infonetics Research, Inc 2021/12/17 Fiberhome Proprietary and Confidential 21

2005– 2011 FTTH Subscribers Source: HEAVY READING | VOL. 4, NO. 9, JUNE 2006

2005– 2011 FTTH Subscribers Source: HEAVY READING | VOL. 4, NO. 9, JUNE 2006 | FTTH WORLDWIDE MARKET & TECHNOLOGY FORECAST 2021/12/17 Fiberhome Proprietary and Confidential 22

Connecting two Ethernet networks Ethernet-over-GEM-over-SDH or Ethernet? ITU-T GPON or IEEE EPON? 2021/12/17 Fiberhome

Connecting two Ethernet networks Ethernet-over-GEM-over-SDH or Ethernet? ITU-T GPON or IEEE EPON? 2021/12/17 Fiberhome Proprietary and Confidential 23

Lessons from History • Ethernet has won every time it has competed with “higher

Lessons from History • Ethernet has won every time it has competed with “higher speed” and “higher efficiency” technologies – Ethernet vs. Token Ring – Ethernet vs. FDDI – Ethernet vs. ATM – Ethernet vs. SONET – Ethernet vs. ATM in the DSLAM – Ethernet vs. Multi-service in the Metro • Ethernet is cheap, simple, easy to install & manage Prediction Ethernet all the way will win a large fraction of the market. 2021/12/17 Fiberhome Proprietary and Confidential 24

Outline Ø EPON vs. GPON Ø CTC EPON IOP Ø Summary 2021/12/17 Fiberhome Proprietary

Outline Ø EPON vs. GPON Ø CTC EPON IOP Ø Summary 2021/12/17 Fiberhome Proprietary and Confidential 25

CTC EPON IOP Key Features 2021/12/17 Fiberhome Proprietary and Confidential 26

CTC EPON IOP Key Features 2021/12/17 Fiberhome Proprietary and Confidential 26

CTC EPON System Evaluation Test 2021/12/17 Fiberhome Proprietary and Confidential 27

CTC EPON System Evaluation Test 2021/12/17 Fiberhome Proprietary and Confidential 27

CTC EPON System Evaluation Test Achievements 2021/12/17 Fiberhome Proprietary and Confidential 28

CTC EPON System Evaluation Test Achievements 2021/12/17 Fiberhome Proprietary and Confidential 28

CTC View of EPON is mature and suitable for mass deployment in CTC –Simple,

CTC View of EPON is mature and suitable for mass deployment in CTC –Simple, easy to develop –Sufficient chip and system vendors –Large-scale, all-around, chip-level and system-level IOP –Mass deployment in east Asia –Stable operation in the field trial of CTC for one and a half years –Decreasing cost 2021/12/17 Fiberhome Proprietary and Confidential 29

CTC View of EPON (Cont. ) After improved by CTC spec, EPON has no

CTC View of EPON (Cont. ) After improved by CTC spec, EPON has no distinctive and essential difference in technical capability compared with GPON -Transport capability -DBA & Qo. S -Operation & Management -Security -Multicast -Fiber protection -Multi-play support 2021/12/17 Fiberhome Proprietary and Confidential 30

NTT View of EPON “Two types of giga-bit PON systems have been standardized: G-PON

NTT View of EPON “Two types of giga-bit PON systems have been standardized: G-PON by ITU-T and GE-PON by IEEE. Now the question is which one is more promising? . . . In Japan, we have seen a drastic price reduction of media converters which could be realized by sharing the technology and products of the LAN market. For services, high quality IP Telephone and IP video are becoming critical basic FTTH services. And for the core network, in NTT we have a full IP backbone network for the FLET’s service. Switches and routers in the network employ Ethernet interfaces. Given these factors, we decided to develop GE-PON as the next-generation FTTH system. ” ----Hiromichi Shinohara, Director of NTT Access Labs (IEEE Communications Magazine, September 2005) 2021/12/17 Fiberhome Proprietary and Confidential 31

Outline Ø EPON vs. GPON Ø CTC EPON IOP Ø Summary 2021/12/17 Fiberhome Proprietary

Outline Ø EPON vs. GPON Ø CTC EPON IOP Ø Summary 2021/12/17 Fiberhome Proprietary and Confidential 32

Summary 1. EPON is more mature & cost-effective than GPON. 2. Both GPON &

Summary 1. EPON is more mature & cost-effective than GPON. 2. Both GPON & EPON will coexist in a long time. 3. Fiberhome is a FTTH leader in China. 4. With our effort, Fiberhome FTTH system will be deployed worldwide soon. 2021/12/17 Fiberhome Proprietary and Confidential 33

Thank you for attention! 2021/12/17 Fiberhome Proprietary and Confidential 34

Thank you for attention! 2021/12/17 Fiberhome Proprietary and Confidential 34