The Great Big NBN Sticky Thread

Discussion in 'Networking, Telephony & Internet' started by Akh-Horus, Dec 21, 2016.

  1. The Sentinel

    The Sentinel Member

    Joined:
    Jan 30, 2002
    Messages:
    2,778
    Says who?

    Have not received any emails about my recently connected Telstra NBN plan going up by $10.
     
  2. caspian

    caspian Member

    Joined:
    Mar 11, 2002
    Messages:
    10,016
    Location:
    Melbourne
    you can connect to one and one only node. there is never any choice in the matter, there is no diversity in the copper network.

    if you want to know where the node is, ping me with your location details.

    I think a large part of the confusion there arises from the incorrect use of the DELT number. without too much boring detail, it's very much not the length of the line. it's the equivalent estimated length of the line, something which is of little use to anyone other than an internal operator who has very good visibility of the physical construction details of the line.

    that's why it's called an estimate - crystal-balling a very variable real world from a database can't take into account any real world factors. the database estimate is a "perfect" result, reality will always be less than that (assuming the database is accurate!).

    the database estimate also can't take into account parts of the line not considered to be part of the network per se, and that includes your lead-in or drop cable, and all of the wiring inside your house.
     
  3. Doc-of-FC

    Doc-of-FC Member

    Joined:
    Aug 30, 2001
    Messages:
    3,294
    Location:
    Canberra
    I've been tasked with investigating a scenario where a Cisco 800 series device does not come back into service when a DSL retrain event (of a particular nature) occurs on a NBN FTTN TC-2 access tail.

    Scenario:
    NBN NEBS TC-2 requires the CPE device to set the L2 COS tag to a value of 4 as per section 4.8 (Priority Encoding) found in the following document: https://www.nbnco.com.au/content/da...catalogue-nebs-product-tech-spec_20170724.pdf

    To accomplish this on a Cisco router, you configure a outbound QOS policy via MQC to set the COS to a value of 4.
    This policy map is subsequently configured against the Ethernet interface as an outbound service-policy.

    When the policy-map is applied to the parent interface (Ethernet0) the following behaviour can be observed when the erratic behaviour occurs:
    Code:
    .Jun  6 10:50:58.992: %LINEPROTO-5-UPDOWN: Line protocol on Interface Ethernet0, changed state to down
    .Jun  6 10:51:00.280: VDSL 0: SM_LINE_DOWN boolean event
    .Jun  6 10:51:02.280: VDSL 0: SM_LINE_TRAIN boolean event
    .Jun  6 10:51:02.280:     vdsl_daemon_sm VDSL 0: during state training, got event 17(line_training)
    .Jun  6 10:51:02.280: @@@ vdsl_daemon_sm VDSL 0: training -> training
    .Jun  6 10:51:27.324: VDSL 0: SM_LINE_SHOWTIME boolean event
    .Jun  6 10:51:27.324: VDSL 0: line state : showtime !
    .Jun  6 10:51:28.324:     vdsl_daemon_sm VDSL 0: during state training, got event 18(showtime)
    .Jun  6 10:51:28.324: @@@ vdsl_daemon_sm VDSL 0: training -> ready
    .Jun  6 10:51:28.324: %CONTROLLER-5-UPDOWN: Controller VDSL 0, changed state to up
    .Jun  6 10:51:28.324:     vdsl_daemon_sm VDSL 0: during state ready, got event 20(conn_mode_chk)
    .Jun  6 10:51:28.324: @@@ vdsl_daemon_sm VDSL 0: ready -> mode_pending
    .Jun  6 10:51:28.324:     vdsl_daemon_sm VDSL 0: idle during state mode_pending
    .Jun  6 10:51:28.324: @@@ vdsl_daemon_sm VDSL 0: mode_pending -> ready
    .Jun  6 10:51:28.324:     vdsl_daemon_sm VDSL 0: during state ready, got event 22(if_state_chk)
    .Jun  6 10:51:28.324: @@@ vdsl_daemon_sm VDSL 0: ready -> ready
    .Jun  6 10:51:28.324:     vdsl_daemon_sm VDSL 0: during state ready, got event 21(if_wakeup)
    .Jun  6 10:51:28.324: @@@ vdsl_daemon_sm VDSL 0: ready -> running
    .Jun  6 10:51:28.324: %HQF-4-NOT_ENOUGH_BW: interface Ethernet0: available bandwidth may not meet bandwidth guarantees on direction OUT
    .Jun  6 10:51:28.332: CBWFQ: Removing service policy on Ethernet0
    .Jun  6 10:51:28.332:     vdsl_daemon_sm VDSL 0: during state running, got event 23(linkup)
    .Jun  6 10:51:28.332: @@@ vdsl_daemon_sm VDSL 0: running -> running
    .Jun  6 10:51:37.996: %LINK-3-UPDOWN: Interface Ethernet0, changed state to up
    .Jun  6 10:51:38.996: %LINEPROTO-5-UPDOWN: Line protocol on Interface Ethernet0, changed state to up
    
    %HQF-4-NOT_ENOUGH_BW: interface Ethernet0: available bandwidth may not meet bandwidth guarantees on direction OUT
    CBWFQ: Removing service policy on Ethernet0


    When the policy-map is applied to the sub-interface (Ethernet0.1) the following behaviour can be observed when the erratic behaviour occurs:
    Code:
    .Jun  7 10:02:21.913: %LINEPROTO-5-UPDOWN: Line protocol on Interface Ethernet0, changed state to down
    .Jun  7 10:02:22.285: VDSL 0: SM_LINE_DOWN boolean event
    .Jun  7 10:02:24.285: VDSL 0: SM_LINE_TRAIN boolean event
    .Jun  7 10:02:24.285:     vdsl_daemon_sm VDSL 0: during state training, got event 17(line_training)
    .Jun  7 10:02:24.285: @@@ vdsl_daemon_sm VDSL 0: training -> training
    .Jun  7 10:02:48.916: VDSL 0: SM_LINE_SHOWTIME boolean event
    .Jun  7 10:02:48.916: VDSL 0: line state : showtime !
    .Jun  7 10:02:49.916:     vdsl_daemon_sm VDSL 0: during state training, got event 18(showtime)
    .Jun  7 10:02:49.916: @@@ vdsl_daemon_sm VDSL 0: training -> ready
    .Jun  7 10:02:49.916: %CONTROLLER-5-UPDOWN: Controller VDSL 0, changed state to up
    .Jun  7 10:02:49.916:     vdsl_daemon_sm VDSL 0: during state ready, got event 20(conn_mode_chk)
    .Jun  7 10:02:49.916: @@@ vdsl_daemon_sm VDSL 0: ready -> mode_pending
    .Jun  7 10:02:49.916:     vdsl_daemon_sm VDSL 0: idle during state mode_pending
    .Jun  7 10:02:49.916: @@@ vdsl_daemon_sm VDSL 0: mode_pending -> ready
    .Jun  7 10:02:49.916:     vdsl_daemon_sm VDSL 0: during state ready, got event 22(if_state_chk)
    .Jun  7 10:02:49.916: @@@ vdsl_daemon_sm VDSL 0: ready -> ready
    .Jun  7 10:02:49.916:     vdsl_daemon_sm VDSL 0: during state ready, got event 21(if_wakeup)
    .Jun  7 10:02:49.916: @@@ vdsl_daemon_sm VDSL 0: ready -> running
    .Jun  7 10:02:49.916: %QOS-4-INVALIDBW: interface Ethernet0: Not enough bandwidth to configure service-policy
    
    .Jun  7 10:02:49.924:     vdsl_daemon_sm VDSL 0: during state running, got event 23(linkup)
    .Jun  7 10:02:49.924: @@@ vdsl_daemon_sm VDSL 0: running -> running
    .Jun  7 10:03:00.912: %LINK-3-UPDOWN: Interface Ethernet0, changed state to up
    .Jun  7 10:03:01.912: %LINEPROTO-5-UPDOWN: Line protocol on Interface Ethernet0, changed state to up
    
    %QOS-4-INVALIDBW: interface Ethernet0: Not enough bandwidth to configure service-policy


    A reapplication of the service policy on either the parent or sub-interface is required to restore service functionality, otherwise a CPE power cycle is required to re-instate services.

    A TAC case will shortly be under way as this is not expected behaviour.
    This device is a Cisco C867VAE2 series router running IOS 15.8(3), I have been made aware that this matter may also be impacting other devices within the Cisco 800 series.

    I have not yet been asked to create a TCL script to re-apply the service policy on the interface based on LINEPROTO-5-UPDOWN, if this occurs I'll provide the script for public consumption.
     
    Last edited: Jun 7, 2019
    chip likes this.
  4. caspian

    caspian Member

    Joined:
    Mar 11, 2002
    Messages:
    10,016
    Location:
    Melbourne
    Wuffles - ran some checks for you. your node is actually just under 1km away including your internal wiring. your node is outside no.41 in your street (weird street numbering) but the main problem is the pillar is at the western end of your street where it intersects with B****lle Rd, so the total wiring run is a lot longer than you might think.

    your actual line speeds are running at around 41/9Mbps which is quite reasonable for that distance and coexistence. I don't know where Future Broadband are getting those numbers, because when I run an SQ on your address it comes back with a very realistic 34/15Mbps.
     
    banshee and callan like this.
  5. banshee

    banshee Member

    Joined:
    Sep 8, 2001
    Messages:
    1,738
    Location:
    Springwood, West of Hell (Sydney, actually)
    Anyone have any ideas what I can try to see why I am getting, on a 50/20 plan with FTTC, 47.5 down, (full speed as expected), but under 10 up? It varies on tests between ~8 &10, and for most things that is fast enough, but if it should be 20 up, is there anything that I can look at that might be the cause?

    All gigabit ethernet connection, not using wireless. Internal transfers between devices with SSDs can run at ~80MBps. QoS on router is disabled.
     
    Last edited: Jun 9, 2019
  6. caspian

    caspian Member

    Joined:
    Mar 11, 2002
    Messages:
    10,016
    Location:
    Melbourne
    get your ISP to check the sync speed of the modem. it's possible the upstream sync speed is lower than expected.
     
  7. banshee

    banshee Member

    Joined:
    Sep 8, 2001
    Messages:
    1,738
    Location:
    Springwood, West of Hell (Sydney, actually)
    I'll give them a call tomorrow.
     
  8. r8response

    r8response Member

    Joined:
    Jun 9, 2004
    Messages:
    6,424
    Location:
    Sunshine Coast
    caspian, Have any areas in Australia within an FTTN footprint had the legacy network shut off yet (vectoring turned on?) ?
     
  9. chip

    chip Member

    Joined:
    Dec 24, 2001
    Messages:
    3,638
    Location:
    Pooraka Maccas drivethrough
    does this mean there's no end-user accessible information from the FTTC NTD about line stats?
     
  10. caspian

    caspian Member

    Joined:
    Mar 11, 2002
    Messages:
    10,016
    Location:
    Melbourne
    no, the FTTC is not user accessible in any way other than the ability to connect your CPE to it. there's no GUI or any other form of reporting you can get to.

    every line in the network always runs vectoring. I think you mean end of coexistence, which is disconnection of all interfering services (not copper disconnection, non-interfering services will remain), which in turn allows relaxation of the DSLAM settings that allow legacy and NBN data services to coexist in the same copper network.

    but yes, heaps have reached end of coexistence and are now exclusively NBN serviced from a data perspective.
     
  11. Wuffles

    Wuffles Member

    Joined:
    Apr 20, 2004
    Messages:
    124
    Location:
    Canberra
    Thanks Caspian, you really do a great service to the people in this thread. Much appreciated mate. I was thinking of getting a cabler to run cat6 from the lead-in but given those distances I'm not sure it'll be worth it.
     
  12. caspian

    caspian Member

    Joined:
    Mar 11, 2002
    Messages:
    10,016
    Location:
    Melbourne
    yeah, TBH I don't think it is worth your spending the money. wait until coexistence ends and you'll get something like another 8-15MBps or so.
     
  13. Sphinx2000

    Sphinx2000 Member

    Joined:
    Sep 16, 2001
    Messages:
    7,427
    Location:
    Brisbane
    The Fastest NBN Plans, According To The ACCC
    https://www.kotaku.com.au/2019/06/heres-the-fastest-nbn-plans-according-to-the-accc/

    The ACCC recently released its fifth report on real-world NBN speeds, and the results weren’t too shabby. Download speeds generally increased when compared to last quarter’s report, including during peak usage hours of 7pm - 11pm. The report’s sample is large enough to provide speed data about nine major ISPs: Aussie Broadband, Dodo, Exetel, iiNet, iPrimus, MyRepublic, Optus, Telstra, and TPG.

    TPG came out on top this time around, Telstra came 3rd last..

    [​IMG]
     
  14. caspian

    caspian Member

    Joined:
    Mar 11, 2002
    Messages:
    10,016
    Location:
    Melbourne
    very close grouping of results.... that's the result of the bundled plans where NBN sold a predetermined adequate-for-purpose chunk of CVC capacity along with the service AVC. nicely completes the job of blowing away the myth that CVC congestion was ever anything other than the ISPs deliberately unprovisioning capacity, once the ability to do so was taken away from them.

    they still need to exclude the services sold as knowingly unable to reach peak speed though. if a line can "only" reach 90Mbps and the end user makes a fully informed opt-in decision to purchase a 100Mbps service, that's not a problem and it should not be reflected as such in the results.
     
  15. mr camouflage

    mr camouflage Member

    Joined:
    May 25, 2012
    Messages:
    770
    Have been in the new house almost a year, can get NBN, but still on ADSL until my conrtact expires, and I can shop around.

    Put the old address in the NBN website. They still cant get NBN at the old house. Sept 2019 they reckon. The node box has been in for over 2 years now and no one in the area can get NBN yet.
     
  16. trevor68

    trevor68 Member

    Joined:
    Jun 28, 2001
    Messages:
    4,262
    Location:
    Canberra
    I'm with The Signal Co, but getting spammed hard with NBN stuff since a node went live a few k's away.
    Hard to work out if it's worthwhile changing, given I have bad memories of max 2Mb down and 20-50 disconnects daily when I was on ADSL.
     
  17. cvidler

    cvidler Member

    Joined:
    Jun 29, 2001
    Messages:
    12,006
    Location:
    Canberra
    They say they've got a new range of plans coming. I've got no choice yet, but will be keen to see what they offer compared to FTTN (Fail to the Nation).

    I've been very happy with their service, hope they can bump speeds.
     
  18. chip

    chip Member

    Joined:
    Dec 24, 2001
    Messages:
    3,638
    Location:
    Pooraka Maccas drivethrough
    Bugger. The access the ISPs have, does it let them do things like read SNMP data?
     
  19. caspian

    caspian Member

    Joined:
    Mar 11, 2002
    Messages:
    10,016
    Location:
    Melbourne
    that sounds rather unusual. what location? just street and suburb will do.

    they get a similar level of access to other NBN technologies. definitely no direct access to the device, they have limited access to an OSS solution that can run some predefined tests. device status, speeds, CPE interface status etc.
     
  20. archie

    archie Member

    Joined:
    Mar 26, 2002
    Messages:
    3,635
    Location:
    western sydney
    is it possible to have two concurrent NBN connections?

    I'm currently on proper Fibre with Optus ($900/month) at 50/50, contract just finished. its been awesome.
    However thinking of ways to save bit of money, hence 2x 100/40 NBN connections with a load balancing router, shouldn't feel too much of a drop & if one ISP goes down the other is fine.

    We are getting HFC at our industrial complex.
    spoke to the contractor laying the cables from the street to my factory. he has no issues running a second cable but cannot connect it (which I understand).
    however down the track can I actually make use of this?
    is it possible/feasible?
     

Share This Page

Advertisement: