Superhub Virgin Business Superhub now has Modem Mode - Page 9

sebyoung · 02-04-2015, 16:59

Assuming then that L2TP might be the only way to go, can anyone suggest what else needs completing on the attached?

qasdfdsaq · 02-04-2015, 17:11

Server address and gateway need to be completed. This is the information VM have never previously given out publicly, hence it being the crucial piece of information everyone else is also missing.

Ignitionnet · 02-04-2015, 17:16

Quote:

Originally Posted by qasdfdsaq

PPPoE makes no sense to me, L2TP does. PPPoE makes sense on ADSL where it's (encapsulated) ethernet to the BRAS. VM doesn't give you an ethernet link to the VPN "server", the ethernet is terminated at the local CMTS, the remainder of the path being over VM's standard IP core.

---------- Post added at 16:53 ---------- Previous post was at 16:49 ----------

I suspect you're wrong and the Superhub creates a direct L2TP tunnel to a remote LAC that VM are using for the static IP endpoint, and the CMTS has nothing to do with it.

This is because a) if it were PPPoE to the CMTS, then there would be no reason for the Superhub to be issued a public routablle DHCP IP on it's cable interface and b) There are both PPP(non-oE) and L2TP up/down messages in the Superhub's local event log.

The LAC is the start point of the L2TP tunnel, not the end. It isn't the static IP end point, that'd be the LNS, and the LAC in this case is the CMTS.

Part of the reason business services were late getting DOCSIS 3 was because VPDN support, as in being a LAC, was either dodgy or non-existent in the early software.

qasdfdsaq · 02-04-2015, 17:44

Quote:

Originally Posted by Ignitionnet

The LAC is the start point of the L2TP tunnel, not the end. It isn't the static IP end point, that'd be the LNS, and the LAC in this case is the CMTS.

Part of the reason business services were late getting DOCSIS 3 was because VPDN support, as in being a LAC, was either dodgy or non-existent in the early software.

Sorry, got my terminology the wrong way round (just like iSCSI, why can't they just call the blooming things "client" and "server")

Point still stands though: Superhub doesn't need a routable IP address for PPPoE but it does need one for L2TP. Superhub also has L2TP session entries in it's local log, which it would know nothing about if the CMTS was responsible for L2TP. Unless VM have sneakily shoehorned some proprietary remote L2TP tunnel state reporting mechanism into PPPoE...

ccarmock · 02-04-2015, 23:19

I hear what you say about PPPoE not needing a public routable IP address, but it would also work if there was one, and I wonder if it's purely since the VM infrastructure is setup top provide 1 public routable IP address to each device they just implemented that way.

I must admit I thought the tunnel was directly between the Superhub and the equipment within the VM network that provided the fixed IP end point. As qasdfdsaq says the superhub log contains entries relating to L2TP.

However I thought the reason VMB still can't offer fixed IP addresses on the 152 Mb/s service was their equipment has issues with PPP at higher speeds, which implies that it is PPP(oE) from the Superhub.

So a few differing views on how this works....

---------- Post added at 23:19 ---------- Previous post was at 23:15 ----------

here are some recent entries from my VMB superhub network log....

02/04/2015 20:15:48 2436694091 L2tp session up
02/04/2015 20:15:48 2436694086 PPP session up
02/04/2015 20:15:42 2436694090 Starting L2tp session
02/04/2015 20:15:06 2436694078 TOD established

qasdfdsaq · 03-04-2015, 11:50

Quote:

Originally Posted by ccarmock

I hear what you say about PPPoE not needing a public routable IP address, but it would also work if there was one, and I wonder if it's purely since the VM infrastructure is setup top provide 1 public routable IP address to each device they just implemented that way.

That's possible, but it's a waste of IP addresses in a time when many companies are struggling to get enough of them.

Quote:

However I thought the reason VMB still can't offer fixed IP addresses on the 152 Mb/s service was their equipment has issues with PPP at higher speeds, which implies that it is PPP(oE) from the Superhub.

Well, PPPoL2TP is not PPPoE but it's still PPP.

Quote:

So a few differing views on how this works....

Indeed. Ignition is usually right about these things but he's so far not said anything to explain the clear discrepancies between his suggestion of PPPoE to the CMTS and your logs showing PPPoL2TP to some unknown location.

Ignitionnet · 03-04-2015, 13:57

Quote:

Originally Posted by qasdfdsaq

Indeed. Ignition is usually right about these things but he's so far not said anything to explain the clear discrepancies between his suggestion of PPPoE to the CMTS and your logs showing PPPoL2TP to some unknown location.

Not this time.

Even though it's a crazy architecture they wanted to use the same one for both xDSL and business cable, so every business modem uses an L2TP tunnel to a PPP service module, these also terminate L2TP from BT, and from there get encapsulated again and terminate on the same BAMs/BRAS as the Virgin National service.

No-one could give a reason why it's done that way beyond a comment about the CMTS

. It's completely contrary to the 'standard' methods which you can find plenty of documentation on on vendor websites.

Name: oxfd-pppsm-1.network.virginmedia.net
Address: 194.145.149.115

Name: oxfd-bam-1.network.virginmedia.net
Address: 194.145.148.252

Seems to be a PPPSM in each regional hubsite along with some of the core sites, and BAMs scattered around.

I reckon if people do traceroutes and take note of which site their core router is at, and from there do an nslookup for a matching PPPSM to that core they might well be able to bring up an L2TP session.

ccarmock · 03-04-2015, 15:04

Do you think this non standard way they have gone bout this is causing the problems with the 152 MB/s fixed IP address service? Any news on that being resolved that you know?

Looks like my tunnel goes to Bradford. I am fed from the NMAL site though

1 <1 ms <1 ms <1 ms gw [10.20.10.1]
2 21 ms 17 ms 17 ms brad-bam-1.network.virginmedia.net [194.145.148.
188]
3 15 ms 15 ms 15 ms brad-core-2b-xe-1111-0.network.virginmedia.net [
213.105.159.42]
4 * * * Request timed out.
5 17 ms 17 ms 16 ms leed-bb-1b-ae1-0.network.virginmedia.net [62.254
.42.122]
6 25 ms 24 ms 22 ms tcl5-ic-3-ae0-0.network.virginmedia.net [62.253.
175.98]
7 24 ms 23 ms 22 ms 233-14-250-212.static.virginm.net [212.250.14.23
3]
8 25 ms 23 ms 25 ms 216.239.47.221
9 24 ms 26 ms 24 ms 216.239.47.229
10 24 ms 23 ms 23 ms google-public-dns-a.google.com [8.8.8.8]

Trace complete.

Ignitionnet · 03-04-2015, 15:22

Your first tunnel probably goes here:

Name: croy-pppsm-1.network.virginmedia.net
Address: 194.145.148.253

Then the connection will either be round robined to a BAM or statically mapped to Bradford.

Odd decision mapping from Croydon to Bradford when there's a BAM in Brentford but it is what it is.

drsox · 03-04-2015, 15:53

May as well throw my one in here then too:

Code:

  1     *        *        *     Request timed out.
  2    10 ms     8 ms    12 ms  lutn-bam-1.network.virginmedia.net [62.253.189.2
26]
  3     9 ms    10 ms    10 ms  lutn-core-2a-xe-113-0.network.virginmedia.net [6
2.252.71.126]
  4     *        *        *     Request timed out.
  5     *        *        *     Request timed out.
  6     *        *        *     Request timed out.
  7    16 ms     *       16 ms  brhm-bb-1c-ae0-0.network.virginmedia.net [62.254
.42.110]
  8    18 ms    17 ms    17 ms  tcl5-ic-2-ae0-0.network.virginmedia.net [212.250
.15.210]
  9    63 ms    64 ms   140 ms  te8-4.core1.thn.as20860.net [195.66.224.207]

qasdfdsaq · 03-04-2015, 16:07

Quote:

Originally Posted by Ignitionnet

Even though it's a crazy architecture they wanted to use the same one for both xDSL and business cable, so every business modem uses an L2TP tunnel to a PPP service module, these also terminate L2TP from BT, and from there get encapsulated again and terminate on the same BAMs/BRAS as the Virgin National service.

So what you're saying is, the business fixed-IP service is dependent on now obsolete kit used to run the now non-existent national ADSL service and the partly-installed-but-later-abandoned FTTC-LLU project? Lovely!

Quote:

No-one could give a reason why it's done that way beyond a comment about the CMTS

. It's completely contrary to the 'standard' methods which you can find plenty of documentation on on vendor websites.

Vendor schmendor.

I've seen enterprise deployments that rely on nothing but MAC-address based port security and who genuinely believed it was actually effective, based on vendor documentation...

Ignitionnet · 03-04-2015, 16:29

Quote:

Originally Posted by qasdfdsaq

So what you're saying is, the business fixed-IP service is dependent on now obsolete kit used to run the now non-existent national ADSL service and the partly-installed-but-later-abandoned FTTC-LLU project? Lovely!

Yep.

Sephiroth · 03-04-2015, 21:21

Quote:

Originally Posted by qasdfdsaq

So what you're saying is, the business fixed-IP service is dependent on now obsolete kit used to run the now non-existent national ADSL service and the partly-installed-but-later-abandoned FTTC-LLU project? Lovely!
<SNIP>...

Reminds me of Patrick McGoohan's famous quote from Ice Station Zebra.

https://www.youtube.com/watch?v=-ht_Z_Dn3Uc

drsox · 08-04-2015, 14:48

So I finally got time to go up to this business and play about...
Modem only mode gave my laptop:

Code:

   IPv4 Address. . . . . . . . . . . : 80.235.149.51(Preferred)
   Subnet Mask . . . . . . . . . . . : 255.255.255.240
   Lease Obtained. . . . . . . . . . : 05 April 2015 20:46:52
   Lease Expires . . . . . . . . . . : 12 April 2015 20:46:52
   Default Gateway . . . . . . . . . : 80.235.149.49
   DHCP Server . . . . . . . . . . . : 62.253.131.192
   DNS Servers . . . . . . . . . . . : 194.168.4.123
                                       194.168.8.123

The IP address isn't the static IP address the business is supposed to have..
I had no internet access and couldn't ping the default gateway - but did have ARP to it:

Code:

80.235.149.49         e0-2f-6d-6c-d8-da     dynamic

Possibly I got dumped into the walled garden "autoregister" mode network?

PPPoE produced absolutely no response.
One thing I forgot to do was to try spoofing the MAC of the cable modem on my machine or ethernet router

I now wonder if that is all that needs to be done? Or go through the Autoregister stuff, if that is even still a thing?

sebyoung · 08-04-2015, 16:24

I had the same result when I connected the SH to our DrayTek Vigor 2860. The IP address was not our static IP (I believe it was 80.193.19.151 with a default gateway of 80.193.19.145). I did try spoofing the MAC address of the SH but it didn't make a difference.