Re: [AusNOG] HE Routing (not a joke)

[Benoit Page-Guitard]

Hi Bradley,

They seemed entirely uninterested in solving their lack of domestic GoF
routes when I challenged them on this topic when they tried to sell me IP
transit a few months back.

My guess is they built out their AU presence on the back of a few lucrative
L2 transport customers, and are basically just fishing for any poor suckers
that want to pick up some cheap international-only transit until they
further expand their AU footprint (and GoF routing).


Regards,

Benoit Page-Guitard
Network Engineer
Digital Pacific Group

Phone: 1300 MY HOST (694 678)
Email: benoit.p...@digitalpacific.com.au
Website: https://www.digitalpacific.com.au

On 15 May 2018 at 19:59, Bradley Amm <b...@bradleyamm.com> wrote:

> So I guess its unlikely that HE will fix the routing to Telstra, Optus,
> AAPT anytime soon.
>
>
>
> Sent from my Samsung Galaxy smartphone.
>
>  Original message 
> From: Nick Stallman <n...@agentpoint.com>
> Date: 16/5/18 7:32 am (GMT+08:00)
> To: Pieter Berkel <pieter.ber...@gmail.com>, McDonald Richards <
> mcdonald.richa...@gmail.com>
> Cc: ausnog@lists.ausnog.net
> Subject: Re: [AusNOG] HE Routing (not a joke)
>
> A beancounter must have realised they had lots of unused Australia -> US
> capacity and decided they wanted to make some profit from it.
>
> On 16/05/18 09:29, Pieter Berkel wrote:
>
> Fair -- but where does "bits we both end up paying for" fit into that
> spectrum?
>
>
> On 16 May 2018 at 09:25, McDonald Richards <mcdonald.richa...@gmail.com>
> wrote:
>
>> standard route policy: bits you pay me for > bits I get for free > bits I
>> pay for.
>>
>> On Wed, May 16, 2018 at 9:23 AM, Pieter Berkel <pieter.ber...@gmail.com>
>> wrote:
>>
>>> Matt and Gavin are right: this asymmetric routing seems to only impacts
>>> providers that buy transit from HE in SJC (TPG and Vocus being the two main
>>> ones I've noticed).  It would appear from the below post that HE policy is
>>> to preference transit over peering links (presuming that AS paths for both
>>> are the same length?):
>>>
>>> https://www.geekzone.co.nz/forums.asp?forumid=81=233
>>> 763_no=2#2006292
>>>
>>> (Although I'm not quite sure how the reverse might break routing, it
>>> might break their business model?).  I guess it's up to Vocus & TPG to
>>> figure out how fix if it becomes a problem for them.
>>>
>>> Edwin: there is some merit to having a routable /48 to experiment with,
>>> not sure if Internode supply that sort of IPv6 subnet to ADSL customers?
>>>
>>>
>>>
>>> On 16 May 2018 at 09:05, Dave Browning <d...@sentrian.com.au> wrote:
>>>
>>>> All good if on MegaIX SYD
>>>>
>>>> Tracing route to tserv1.syd1.he.net [216.218.142.50]
>>>> over a maximum of 30 hops:
>>>>
>>>>   1<1 ms<1 ms<1 ms  vl666.cr01.b1.bne.qld.au.sentr
>>>> ian.net.au [103.226.9.138]
>>>>   215 ms16 ms
>>>> <https://maps.google.com/?q=2%C2%A0+%C2%A0+15+ms%C2%A0+%C2%A0+16+ms=gmail=g>
>>>>   14 ms  vl3.cr01.s1.syd.nsw.au.sentrian.net.au [103.226.9.245]
>>>> <https://maps.google.com/?q=245%5D+%0D%0A%C2%A0+3%C2%A0+%C2%A0+13+ms%C2%A0+%C2%A0+16+ms=gmail=g>
>>>>   313 ms16 ms14 ms  as6939.sydney.megaport.com
>>>> [103.26.68.236]
>>>>   412 ms19 ms
>>>> <https://maps.google.com/?q=4%C2%A0+%C2%A0+12+ms%C2%A0+%C2%A0+19+ms=gmail=g>
>>>>   15 ms  tserv1.syd1.he.net [216.218.142.50]
>>>>
>>>> Dave Browning | Network Engineer
>>>> P 1300 791 678
>>>> Level 1, 12 Railway Tce, Milton QLD 4064
>>>> <https://maps.google.com/?q=12+Railway+Tce,+Milton+QLD+4064=gmail=g>
>>>> ___
>>>> AusNOG mailing list
>>>> AusNOG@lists.ausnog.net
>>>> http://lists.ausnog.net/mailman/listinfo/ausnog
>>>>
>>>
>>>
>>> ___
>>> AusNOG mailing list
>>> AusNOG@lists.ausnog.net
>>> http://lists.ausnog.net/mailman/listinfo/ausnog
>>>
>>>
>>
>
>
> ___
> AusNOG mailing 
> listAusNOG@lists.ausnog.nethttp://lists.ausnog.net/mailman/listinfo/ausnog
>
>
> --
> Nick Stallman
> Technical Director
> [image: Email] n...@agentpoint.com
> [image: Phone] 02 8039 6820 <0280396820>
> [image: Website] www.agentpoint.com.au
> [image: Agentpoint] <https://www.agentpoint.com.au/>
> [image: Netpoint] <https://netpoint.group/>
> Level 3, 100 Harris Street, Pyrmont NSW 2009 [image: Facebook]
> <https://www.facebook.com/agentpoint/> [image: Twitter]
> <https://twitter.com/agentpoint> [image: Instagram]
> <https://www.instagram.com/Agentpoint/> [image: Linkedin]
> <https://www.linkedin.com/company/agentpoint-pty-ltd>
>
> ___
> AusNOG mailing list
> AusNOG@lists.ausnog.net
> http://lists.ausnog.net/mailman/listinfo/ausnog
>
>
___
AusNOG mailing list
AusNOG@lists.ausnog.net
http://lists.ausnog.net/mailman/listinfo/ausnog

Re: [AusNOG] Vocus doody st power ?

[Benoit Page-Guitard]

Hi Ben,

Yeah. We lost our racks there between 07:13 and 07:15 AEDT (approx).

Power appears to have been pretty quickly restored.

A recent hazard notification suggests they were doing UPS maintenance on
site. Maybe they got unlucky?


Regards,

Benoit Page-Guitard
Network Engineer
Digital Pacific Group

Phone: 1300 MY HOST (694 678)
Email: benoit.p...@digitalpacific.com.au
Website: https://www.digitalpacific.com.au

On 12 February 2018 at 15:40, Ben Cornish <b...@overthewire.com.au> wrote:

> Anyone else seeing power loss in Sydney Vocus doody st ?
>
> Regards
> Ben Cornish
> ___
> AusNOG mailing list
> AusNOG@lists.ausnog.net
> http://lists.ausnog.net/mailman/listinfo/ausnog
>
___
AusNOG mailing list
AusNOG@lists.ausnog.net
http://lists.ausnog.net/mailman/listinfo/ausnog

Re: [AusNOG] web24 contact

[Benoit Page-Guitard]

Hi Anthony,

We'll get in touch with you shortly off-list to sort this out.


Regards,

Benoit Page-Guitard
Network Engineer
Hostopia Australia

Phone: 1300 MY HOST (694 678)
Email: benoit.p...@hostopia.com.au
Website: https://www.hostopia.com.au/

On 9 August 2018 at 23:55, Anthony Bortolotto 
wrote:

> Does anyone have a network contact for someone at web24 that they can send
> me off list? Appears to be some routing issues we would like to resolve,
> appears to be affecting AussieBB too but not others such Telstra.
>
>
>
> AJB
>
> ___
> AusNOG mailing list
> AusNOG@lists.ausnog.net
> http://lists.ausnog.net/mailman/listinfo/ausnog
>
>
___
AusNOG mailing list
AusNOG@lists.ausnog.net
http://lists.ausnog.net/mailman/listinfo/ausnog

Re: [AusNOG] Telstra Insanity?

[Benoit Page-Guitard]

Hi Skeeve,

As Brad said, that's just ECMP working its multipath magic.

It looks like your network has two connections into Telstra's network (one
each off chw-edge901 and ken-edge901), which are of unequal path lengths as
seen from the perspective of your local machine (with the last hop on your
network on the way to ken-edge901 not shown in the above output).

ECMP combined with a low number of UDP probes (and thus insufficient
randomnness to pluck all possible paths out of the woodwork) can lead to
extremely confusing ECMP traceroutes.


Try running an ICMP trace instead to get a solid idea of what (one) of your
paths is like:

  $ sudo traceroute -I 1.1.1.1


Regards,

Benoit Page-Guitard
Network Engineer
Hostopia Australia

Phone: 1300 MY HOST (694 678)
Email: benoit.p...@hostopia.com.au
Website: https://www.hostopia.com.au/

On 10 October 2018 at 19:54, Skeeve Stevens <
skeeve+aus...@eintellegonetworks.com> wrote:

> Hey all,
>
> Does anyone understand this... it hurts my brain.
>
> Skeeve-Faqtory-27:~ skeeve.stevens$ traceroute 1.1.1.1
> traceroute to 1.1.1.1 (1.1.1.1), 64 hops max, 52 byte packets
>  1  x.x.x.10 (x.x.x.10)  0.949 ms  0.615 ms  0.585 ms
>  2  x.x.x.1 (x.x.x.1)  1.169 ms  1.106 ms  1.022 ms
>
>  3  172.18.208.9 (172.18.208.9)  19.487 ms  19.802 ms  19.576 ms
>  4  172.18.66.197 (172.18.66.197)  20.170 ms
> 172.18.66.193 (172.18.66.193)  20.538 ms  19.848 ms
>  5  bundle-ether4.chw-edge901.sydney.telstra.net (203.50.12.108)  20.093
> ms  20.163 ms  21.795 ms
>  6  bundle-ether2.ken-edge901.sydney.telstra.net (203.50.11.102)  20.296
> ms  19.528 ms
> bundle-ether13.chw-core10.sydney.telstra.net (203.50.11.98)  21.252 ms
>  7  bundle-ether1.chw-edge902.sydney.telstra.net (203.50.11.101)  19.891
> ms
> bundle-ether13.ken-core10.sydney.telstra.net (203.50.11.94)  22.118
> ms  21.234 ms
>  8  bundle-ether2.ken-edge902.sydney.telstra.net (203.50.11.104)  19.428
> ms  20.432 ms
> bundle-ether1.ken-edge902.sydney.telstra.net (203.50.11.97)  20.719 ms
>  9  clo2241682.lnk.telstra.net (144.130.174.50)  20.385 ms  20.533 ms
> 20.532 ms
> 10  one.one.one.one (1.1.1.1)  20.224 ms  19.687 ms  19.703 ms
>
> It looks like it going to Chatswood, Kent St and back and forwards. It is
> odd.
>
> ...Skeeve
>
> *Skeeve Stevens - Founder & Chief Architect - *eintellego Networks Pty Ltd
> ske...@eintellegonetworks.com ; www.eintellegonetworks.com
>
> Phone: 1300 239 038; Cell +61 (0)414 753 383 ; skype://skeeve
>
> facebook.com/eintellegonetworks ;  <http://twitter.com/networkceoau>
> linkedin.com/in/skeeve
> Cumulus Linux / Open Networking - Cloud - Consulting - Juniper - Cisco - IPv4
> Brokering
>
> ___
> AusNOG mailing list
> AusNOG@lists.ausnog.net
> http://lists.ausnog.net/mailman/listinfo/ausnog
>
>
___
AusNOG mailing list
AusNOG@lists.ausnog.net
http://lists.ausnog.net/mailman/listinfo/ausnog

Re: [AusNOG] AWS With Megaport

[Benoit Page-Guitard]

As Steve indirectly pointed out above, an AWS VPC is not a real layer 2
domain.. so there's nothing really to "extend into".

The best way to conceptualise a VPC at a layer 2 level (at least from a
network engineer's perspective) is to imagine a full mesh of hosts,
connected by overlay tunnels, where every host has (or can have) full
knowledge of MAC-to-tunnel-endpoint mappings for every VPC member. It's
also unicast-only.

Basically, the only way for a MAC to be "added" to a VPC is for it to be
programmed onto each host.

It's not a completely accurate picture of how things actually work, but it
serves well enough to explain why you can't just extend your own layer 2
domain onto a VPC and expect broadcasts and L2 learning to "just work".

Hope that helps..


Benoit

On Tue, 12 Feb 2019 at 03:56, simon thomason  wrote:

> Friends don't let friends build large layer2 networks!
>
> That said; if you wanted to get a little creative you could run up a
> csr1000v and setup otv and lisp from AWS to your on prem infra. I have
> never set it up but would be interesting to see if the MP transport
> supports it.
>
> I guess the question is really, why would you want to extend your layer2
> domain into the cloud.
> You will not be able to run redundant connections.
> You have just increased your fail domain size and will be tromboning
> traffic.
>
>
>
> On Tue., 12 Feb. 2019, 6:17 pm Tim Raphael  wrote:
>
>> That’s still routed and not bridged over the DC. It’s all layer 3 over
>> DCs unless (as mentioned) you use an L2 tunnelling method onto a compute VM.
>>
>> - Tim
>>
>> On 12 Feb 2019, at 19:01, Andras Toth  wrote:
>>
>> I'm not sure if the Megaport Direct Connect is any different to a regular
>> AWS Direct Connect, but a friend of mine has successfully used a regular
>> AWS Direct Connect that they extended into their EC2 VPC and assigned IP
>> addresses to their instances from their own /28 range.
>>
>> The subnet has a Route Table that would say send everything to the DX VGW
>> to make this work.
>>
>> The only caveat was the .1 and .2 addresses cannot be used because that's
>> the AWS Default Gateway and DNS, etc.
>>
>> Andras
>>
>>
>> On Tue, Feb 12, 2019 at 1:30 PM Steve Tu  wrote:
>>
>>> Hi Joseph,
>>>
>>> Thank your for your enquiry.
>>>
>>> As Simon Thomason pointed out, AWS Direct Connect - Private Virtual
>>> Interface is mainly for connecting two private networks together by
>>> exchanging routes via BGP.
>>>
>>> In your scenario of extending your on-premises network into the native
>>> AWS VPC so they are on the same broadcast domain, this is not a supported
>>> scenario. Also on AWS VPC route table, the destination via next-hop/default
>>> gateway will need to be an AWS component i.e. Virtual Private Gateway or
>>> Internet Gateway etc. and not something outside of AWS.
>>>
>>> https://aws.amazon.com/vpc/faqs/
>>> Please also check the Topology section of the FAQ
>>>
>>> There are options such as VMware on AWS and creating a L2VPN over Direct
>>> Connect, that may suits your requirements.
>>>
>>> I hope this answers your question, and if you have any further
>>> questions, please do not hesitate to contact me,
>>>
>>> *SteveTu*
>>> *Solutions Architect APAC*
>>>
>>> Phone: +61 430 965 798
>>> [image: Connect on LinkedIn] 
>>>
>>>
>>>
>>> [image: Visit Megaport.com] 
>>>
>>> This message is intended for the addressee named and may contain
>>> confidential information. If you are not the intended recipient please
>>> delete it and notify the sender.
>>>
>>>
>>> On Tue, Feb 12, 2019 at 12:12 PM Gavin Tweedie <
>>> gavin.twee...@megaport.com> wrote:
>>>
 I'm not our cloud wizard so let me throw on a technical solutions guy
 or two on this thread to bring it to their attention!

 Gav

 On Tue, 12 Feb 2019 at 10:08, simon thomason  wrote:

> I would be interested to know how you do this also.
>
> My understanding is that over MP you need to peer with AWS in BGP.
> https://knowledgebase.megaport.com/cloud-connectivity/aws-cloud/
>
> Even looking at AWS doco it says you can not
> https://aws.amazon.com/directconnect/faqs/.
>
> Unless you want to run GRE or something over this but i am not certain
> why you would.
>
> On Tue, Feb 12, 2019 at 11:53 AM Nathan Brookfield <
> nathan.brookfi...@simtronic.com.au> wrote:
>
>> 100% possible :)
>>
>> Nathan Brookfield
>> Chief Executive Officer
>>
>> Simtronic Technologies Pty Ltd
>> http://www.simtronic.com.au
>>
>> On 12 Feb 2019, at 12:40, Joseph Goldman  wrote:
>>
>> Hi *
>>
>>  Just wondering if the following scenario is supported for EC2
>> instances with AWS.
>>
>>  Over megaport, I'd like to use a VXC (Or Direct Connect) - On that
>> interface on my router, I put x.x.x.1/24, then on my EC2 instances I'd 
>> want
>> to put x.x.x.2-254/24 

Re: [AusNOG] Digital Pacific - Mail Cluster - Support Contact

[Benoit Page-Guitard]

Hi David,

I'll have someone from our operations team reach out to you.

Cheers.

Benoit Page-Guitard
Network Engineer
Hostopia Australia

On Thu, 26 Sep 2019 at 23:00, David Moyle  wrote:

> Hi all,
>
>
>
> Non-urgent issue – Is anyone able from Digital Pacific Tech Support team
> able to contact me off-list.
>
>
>
> Regarding Mail Cluster (mailcluster.com.au) rejecting emails for one of
> our business clients due to a bizarre header error?
>
>
>
> Cheers all!
>
> David
>
>
> ___
> AusNOG mailing list
> AusNOG@lists.ausnog.net
> http://lists.ausnog.net/mailman/listinfo/ausnog
>
___
AusNOG mailing list
AusNOG@lists.ausnog.net
http://lists.ausnog.net/mailman/listinfo/ausnog

Re: [AusNOG] Intermittent problem accessing Goggle services.

[Benoit Page-Guitard]

Hi Paul,

Definitely been seeing the same thing on our end (AS55803).

As others have noted, we also coincidentally saw Google flap on MegaIX NSW
last night, along with a few others (e.g. AS15133, AS46489, etc). I'm
deactivating our peering on that exchange now to see if it is also the
source of our own troubles.

This smells more like a MegaIX NSW VPLS issue than it does a Google issue.
It definitely wouldn't be the first time (I can recall a couple of
incidences of sporadic loss to/from some peers in the last 1-2 years?).

Regards,

Benoit Page-Guitard
Network Engineer
Hostopia Australia


On Fri, 17 Jan 2020 at 00:58, Paul Holmanskikh  wrote:

> Hi,
>
> Is anyone experiencing problems with Google today?  Some IPs can't even
> reach 8.8.8.8. but majority are working fine. I can't find any problems
> with our network.
>
> ---
> NEXON - I.T. FOR THE DYNAMIC BUSINESS
> Paul Holmanskikh
> Senior Network Engineer
>
> Disclaimer: The contents of this email represent my own views and not
> necessarily the views of my employer
>
> ___
> AusNOG mailing list
> AusNOG@lists.ausnog.net
> http://lists.ausnog.net/mailman/listinfo/ausnog
>
___
AusNOG mailing list
AusNOG@lists.ausnog.net
http://lists.ausnog.net/mailman/listinfo/ausnog