Overcoming OBI in RFoG Networks. Michael McWilliams ANGA Cologne, Germany June 9, 2016

Transcription

1 Overcoming OBI in RFoG Networks Michael McWilliams ANGA Cologne, Germany June 9, 2016

2 Agenda Optical Beat Interference (OBI) Causes Analysis Identification Mitigation The answer 2

3 OBI Causes OBI Occurs when two optical transmitters of the same or very similar wavelengths are seen at the same receiver. When two polarized optical signals are detected in the same optical receiver, then the photodiode current is equal to: I(t) = R (E 1 (t) + E 2 (t)) 2 (R is receiver responsivity) Expressed as power: = R [ P 1 (t) + P 2 (t) + ] Where α is a difference angle between the two polarized signals and ω 2 ω 1 is a difference frequency between the two optical signals 3

4 OBI Causes OBI induced with two optical transmitters Performance not affected at >0.2 nm separation between wavelengths 0.35nm Increased noise floor at separations <0.08 nm 4

5 OBI Causes OBI Network drivers It is a statistical probability with multiple variables Number of receivers combined at input to CMTS Port Number of active upstreams Data usage Number of customers Temperature variation between R ONUs 5

6 OBI Analysis Temperature variation Upstream lasers used in R ONU devices are un cooled DFBs Optical transmitters wavelength will vary over temperature 0.1 nm per Degree C or 1 nm for a 10 C temperature change Cannot assume all will be at the same temperature; placement of device will impact its operating temperature 6

7 OBI Analysis ARRIS has run a statistical analysis on incidences of OBI against certain variables To make the maths manageable various assumptions have been made Optical wavelengths of transmitters Follow Gaussian distribution For no OBI, separation between wavelengths assumed to be > 4 GHz (0.33 nm) Number of transmitters on Single MAC domain of DOCSIS 3.0 with n bonded upstreams Multiple MAC domains each with single upstream Upstream traffic is assumed to be at maximum capacity of link n upstream transmitters assumed to ON at all times Uniform usage amongst all users 7

8 OBI Analysis OBI Eight Receivers combined Typically 256 customer (8 x 32) 1 to 6 transmitters on For no temperature difference incidence is very low and remains almost constant; Red 0 C As temperature variations between R ONUs increase so does OBI Blue 10 C 8

9 OBI Analysis What does this mean? When OBI is present, data throughput to the CMTS port is zero Worst case scenario For lower traffic loads the % probability of OBI can be reduced by the usage factor; For example if upstream usage is at 50% OBI probability is ½ calculated value OBI is dynamically changing in a network R ONUs upstream wavelength change with temperature 9

10 OBI free: Identifying OBI is a short term burst type phenomenon Best indication of its presence is by examination of the CMTS messages When OBI is present the number of Codeword Errors will rise, in a normally operating system these are typically well below 1% Increase with usage Copyright , ARRIS Enterprises, LLC. All rights reserved. 10

11 OBI Identifying Due to its burst nature it will only be present when two transmitters are active 5-42 MHz Span Its presence will corrupt data on all four active carriers 10dB / div Challenge Which two R ONUs are causing it? 11

12 OBI Field Experience Upstream performance measured using Pathtrak TM Showing bursts of high Codeword Errors and the subsequent impact on system MER 12

13 OBI Deployments with Standard R ONUs For RFoG systems with single RF upstream, OBI cannot be a problem; The CMTS system manages the upstream usage and ensures only a single R ONU is active at any one time OBI generated data loss: some percentage may be acceptable For Data transfer, system will resend any lost packets, unlikely to be noticed by customer For active voice calls, low levels of OBI may result in intermittent loss of upstream. If too high, may result in dropped calls Received information from active deployments indicates VPN connections established through a cable modem drop out when Codeword Error percentages rise above 5% 13

14 OBI Mitigation with Standard R ONUs What Variables can you control? Upstream carrier count more increases OBI Expand bandwidth modulation to highest available: 128 QAM with 6.4 MHz BW and retain single RF upstream Limit the upstream carriers, there will be OBI present but may be acceptable on a lightly loaded network Maximize number of optical receivers connected to a single CMTS port Reduce number of R ONUs per receiver Spread connected customers across all active receivers if possible All of these options impose limitations on upstream bandwidth and limit service to customers. Not a good business model! 14

15 OBI DOCSIS 3.1 Techniques listed provide some level of OBI mitigation but they do not eliminate it DOCSIS 3.1 introduces OFDM upstream modulation with multiple 25/50KHz upstream carriers. Individual cable modems can be allocated a small subset of these carriers. 32 cable modems could be active simultaneously in the return Probability of OBI increases with the number of upstream channels. Expand on our earlier OBI Model to increase upstream carrier count Upstream Channels Probability of OBI 2% 10% 39% 89% 100% To support DOCSIS 3.1 on RFoG Systems requires a true OBI free solution 15

16 OBI Free For true OBI free operation the upstream optical frequencies must be managed to eliminate the chance of two transmitters operating at the same wavelength being seen at a common receiver Two principle technologies to solve this challenge: Eliminate OBI at the detection stage. Active optical splitter using multiple receivers, one per active transmitter Manage and control the upstream wavelengths from the R ONUs For more information on either option please visit the ARRIS booth E21, Hall 10.2 on the upper floor 16

17 Q&A Thank you for your time and attention Any Questions?

18 Michael McWilliams Senior Product Manager