Re: Curiosity about AS3356 L3/CenturyLink network resiliency (in general)

[Mike Hammett]

To circle back to the original post... Level 3 does have multiple routes out of 
Tampa. They just apparently don't use them all for their transit service. Why 
not? 




- 
Mike Hammett 
Intelligent Computing Solutions 
http://www.ics-il.com 

Midwest-IX 
http://www.midwest-ix.com 

- Original Message -

From: "valdis kletnieks"  
To: "Baldur Norddahl"  
Cc: nanog@nanog.org 
Sent: Sunday, May 20, 2018 5:43:42 PM 
Subject: Re: Curiosity about AS3356 L3/CenturyLink network resiliency (in 
general) 

On Sun, 20 May 2018 09:16:25 +0200, Baldur Norddahl said: 

> He is complaining about AS3356 in specific and claiming they COULD 
> reroute around it but choose not to. This leads me to assume there are 
> alternatives. Two places, Miami and Texas, are mentioned and that a 
> double fault, one in Miami and another in Texas would bring down the 
> network. I am from Europe, but am I to believe that Miami and Texas (or 
> anywhere between those two) are served by only two fiber conduits? 

There's a difference between "route around it by flipping some BGP magic" and 
"route around it by digging a ditch to a third city". 

The fact that other places have other conduits doesn't change the fact that a 
given city may only have two physical conduits handy. Often, there are other 
*possible* paths that could be built out, but other providers have looked at 
the cost of digging a ditch from the city, out a third path, to their closest 
POP, and decided it's not economically feasible. You can only route across the 
fiber that's actually there and lit up. 

You're from Europe? OK, consider this setup: Andorra. Two providers, one of 
who backhaul that path all the way to Madrid, and the other that backhauls to 
Marseilles. Sure, there's other cities along the way, but there's no fiber path 
from where you are to there. For instance, the fiber path may run from Madrid 
to Zaragoza, where it splits 3 ways to Pamplona, Andorra, and Barcelona - but 
if Barcelona and Pamplona don't provide alternate paths out to the net, you're 
still going to Madrid. Meanwhile, other companies may provide service to lots 
of smaller places along the border on the Spain side, and other companies 
provide service to lots of places on the French side, but not into Andorra 
itself. 

You don't like that, consider any one of the many European cities that are in a 
deep river valley, so the only realistic ways to the outside world are 
"upstream" and "downstream". 

> The question was if downtime on a transit provider of many hours is 
> unacceptable. I am offering my experience that this happens to all of 
> them. Some of them can have problems that last days not hours. Do not 
> ever assume that a so called "tier 1" network is good as your only transit. 

The gotcha here is the very high danger than with only two paths out of the 
city, your second and third choices are fate-sharing with that Tier 1. If 
you're 
in Andorra, and you have 8 providers that share a path through a tunnel to 
Toulouse, 
and another 6 that share a bridge to Barcelona, you still have a problem. 

(That, and anybody who buys transit only from one Tier 1 is going to have 
a really hard time getting routes to the *rest* of the internet...) 

Re: Curiosity about AS3356 L3/CenturyLink network resiliency (in general)

[valdis . kletnieks]

On Sun, 20 May 2018 09:16:25 +0200, Baldur Norddahl said:

> He is complaining about AS3356 in specific and claiming they COULD
> reroute around it but choose not to. This leads me to assume there are
> alternatives. Two places, Miami and Texas, are mentioned and that a
> double fault, one in Miami and another in Texas would bring down the
> network. I am from Europe, but am I to believe that Miami and Texas (or
> anywhere between those two) are served by only two fiber conduits?

There's a difference between "route around it by flipping some BGP magic" and
"route around it by digging a ditch to a third city".

The fact that other places have other conduits doesn't change the fact that a
given city may only have two physical conduits handy.  Often, there are other
*possible* paths that could be built out, but other providers have looked at
the cost of digging a ditch from the city, out a third path, to their closest
POP, and decided it's not economically feasible.  You can only route across the
fiber that's actually there and lit up.

You're from Europe?  OK, consider this setup:  Andorra.  Two providers, one of
who backhaul that path all the way to Madrid, and the other that backhauls to
Marseilles. Sure, there's other cities along the way, but there's no fiber path
from where you are to there.  For instance, the fiber path may run from Madrid
to Zaragoza, where it splits 3 ways to Pamplona, Andorra, and Barcelona - but
if Barcelona and Pamplona don't provide alternate paths out to the net, you're
still going to Madrid. Meanwhile, other companies may provide service to lots
of smaller places along the border on the Spain side, and other companies
provide service to lots of places on the French side, but not into Andorra
itself.

You don't like that, consider any one of the many European cities that are in a
deep river valley, so the only realistic ways to the outside world are
"upstream" and "downstream".

> The question was if downtime on a transit provider of many hours is
> unacceptable. I am offering my experience that this happens to all of
> them. Some of them can have problems that last days not hours. Do not
> ever assume that a so called "tier 1" network is good as your only transit.

The gotcha here is the very high danger than with only two paths out of the
city, your second and third choices are fate-sharing with that Tier 1.  If 
you're
in Andorra, and you have 8 providers that share a path through a tunnel to 
Toulouse,
and another 6 that share a bridge to Barcelona, you still have a problem.

(That, and anybody who buys transit only from one Tier 1 is going to have
a really hard time getting routes to the *rest* of the internet...)



pgpPJE3sF9gLw.pgp
Description: PGP signature

Re: Curiosity about AS3356 L3/CenturyLink network resiliency (in general)

[Christopher Morrow]

On Sun, May 20, 2018 at 12:33 PM Rubens Kuhl  wrote:
>
> CenturyLink bought Level 3, which bought Global Crossing, which bought
> Impsat; this makes every market unique, for the good and bad of it.
>
> What I have as a customer feeling is that Global Crossing was the most
> quality-minded of the 4, while the other 3 is/were more "take what we give
> you and shut up".
that might be a thing related to the time when GC was around individually
though, right?
they could have been considered 'boutique' network provider at the time...
The L3/GC merger was ~10 yrs ago? much has changed in the carrier space
since...
being bigger dpesn't often make companies higher touch :)

Re: Curiosity about AS3356 L3/CenturyLink network resiliency (in general)

[Rubens Kuhl]

CenturyLink bought Level 3, which bought Global Crossing, which bought
Impsat; this makes every market unique, for the good and bad of it.

What I have as a customer feeling is that Global Crossing was the most
quality-minded of the 4, while the other 3 is/were more "take what we give
you and shut up".


Rubens


On Wed, May 16, 2018 at 1:59 PM, David Hubbard <
dhubb...@dino.hostasaurus.com> wrote:

> I’m curious if anyone who’s used 3356 for transit has found shortcomings
> in how their peering and redundancy is configured, or what a normal
> expectation to have is.  The Tampa Bay market has been completely down for
> 3356 IP services twice so far this year, each for what I’d consider an
> unacceptable period of time (many hours).  I’m learning that the entire
> market is served by just two fiber routes, through cities hundreds of miles
> away in either direction.  So, basically two fiber cuts, potentially 1000+
> miles apart, takes the entire region down.  The most recent occurrence was
> a week or so ago when a Miami-area cut and an Orange, Texas cut (1287
> driving miles apart) took IP services down for hours.  It did not take
> point to point circuits to out of market locations down, so that suggests
> they even have the ability to be more redundant and simply choose not to.
>
> I feel like it’s not unreasonable to expect more redundancy, or a much
> smaller attack surface given a disgruntled lineman who knows the routes
> could take an entire region down with a planned cut four states apart.
> Maybe other regions are better designed?  Or are my expectations
> unreasonable?  I carry three peers in that market, so it hasn’t been
> outage-causing, but I use 3356 in other markets too, and have plans for
> more, but it makes me wonder if I just haven't had the pleasure of similar
> outages elsewhere yet and I should factor that expectation into the
> design.  It creates a problem for me in one location where I can only get
> them and Cogent, since Cogent can't be relied on for IPv6 service, which I
> need.
>
> Thanks
>
>
>

Re: Curiosity about AS3356 L3/CenturyLink network resiliency (in general)

[joel jaeggli]

On 5/17/18 6:24 AM, Mike Hammett wrote:
> I often question why\how people build networks the way they do. There's some 
> industry hard-on with having a few ginormous routers instead of many smaller 
> ones. I've learned that when building Internet Exchanges, the number of 
> networks that don't have BGP edge routers in major markets where they have a 
> presence is quite a bit larger than one would expect. I heard a podcast once 
> (I forget if it was Packet Pushers or Network Collective) postulating that 
> the reason why everything runs back to a few big ass routers is that someone 
> decided to spend a crap-load of money on big ass routers for bragging rights, 
> so now they have to run everything they can through them to A) "prove" their 
> purchase wasn't foolish and B) because they now can't afford to buy anything 
> else. 
There  seems to be a bit of overstatement with respect to how large
these are...

alcatel/nokia 7750 (L3's newer PE platform) is large but not outlandish
and they've been deployed for a couple years. it's relatively similar in
capacity  or a to to the devices that many of us interconnect with them
using.  Most of their customers probably though not always need less fib
then they need on a PE router.

There is a longer time-scale overhang from the choice to design of MPLS
core networks 15-20 years ago where PE routers have more to do fib wise
then do cores (which may well be larger and simpler, since most of what
they do is label switching), that drives the selection of what hardware
goes in the edge in ways than an IP only carrier might make different
choices (e.g. this big fib/queue/buffer router might have been a large
l3 switch).
> There's no reason why Tampa doesn't have a direct L3 adjacency to Miami, 
> Atlanta, Houston, and Charlotte over diverse infrastructure to all four. 
> Obviously there's room to add\drop from that list, but it gets the point 
> across. 
the number of paths available into and out a market seems somewhat
orthogonal to the number of routers.
>
>
> - 
> Mike Hammett 
> Intelligent Computing Solutions 
> http://www.ics-il.com 
>
> Midwest-IX 
> http://www.midwest-ix.com 
>
> - Original Message -
>
> From: "David Hubbard"  
> To: nanog@nanog.org 
> Sent: Wednesday, May 16, 2018 11:59:42 AM 
> Subject: Curiosity about AS3356 L3/CenturyLink network resiliency (in 
> general) 
>
> I’m curious if anyone who’s used 3356 for transit has found shortcomings in 
> how their peering and redundancy is configured, or what a normal expectation 
> to have is. The Tampa Bay market has been completely down for 3356 IP 
> services twice so far this year, each for what I’d consider an unacceptable 
> period of time (many hours). I’m learning that the entire market is served by 
> just two fiber routes, through cities hundreds of miles away in either 
> direction. So, basically two fiber cuts, potentially 1000+ miles apart, takes 
> the entire region down. The most recent occurrence was a week or so ago when 
> a Miami-area cut and an Orange, Texas cut (1287 driving miles apart) took IP 
> services down for hours. It did not take point to point circuits to out of 
> market locations down, so that suggests they even have the ability to be more 
> redundant and simply choose not to. 
>
> I feel like it’s not unreasonable to expect more redundancy, or a much 
> smaller attack surface given a disgruntled lineman who knows the routes could 
> take an entire region down with a planned cut four states apart. Maybe other 
> regions are better designed? Or are my expectations unreasonable? I carry 
> three peers in that market, so it hasn’t been outage-causing, but I use 3356 
> in other markets too, and have plans for more, but it makes me wonder if I 
> just haven't had the pleasure of similar outages elsewhere yet and I should 
> factor that expectation into the design. It creates a problem for me in one 
> location where I can only get them and Cogent, since Cogent can't be relied 
> on for IPv6 service, which I need. 
>
> Thanks 
>
>
>
>

Re: Curiosity about AS3356 L3/CenturyLink network resiliency (in general)

[Mark Tinka]

On 20/May/18 09:16, Baldur Norddahl wrote:

>
> The question was if downtime on a transit provider of many hours is
> unacceptable. I am offering my experience that this happens to all of
> them. Some of them can have problems that last days not hours. Do not
> ever assume that a so called "tier 1" network is good as your only
> transit.

And that is where the sage advice is...

Just because they are "large", "global", "transit-free",
"international", "Tier this or Tier that", don't think they are beyond
fault. And more importantly, don't allow your customers to assume they
are beyond fault, just because you aren't them.

Take control of your situation, especially if you can.

Mark.

Re: Curiosity about AS3356 L3/CenturyLink network resiliency (in general)

[Mark Tinka]

On 19/May/18 22:28, Baldur Norddahl wrote:

> What happened to do not trust anyone? Create your own resiliency by being
> multihomed to as many transits you can afford.
>
> You need the ability to shutdown a transit that is having trouble. It
> happens to all of them.

Agreed.

Mark.

Re: Curiosity about AS3356 L3/CenturyLink network resiliency (in general)

[Baldur Norddahl]

Den 20/05/2018 kl. 05.43 skrev valdis.kletni...@vt.edu:

On Sat, 19 May 2018 22:28:07 +0200, Baldur Norddahl said:

What happened to do not trust anyone? Create your own resiliency by being
multihomed to as many transits you can afford.

Re-read what David Hubbard said:


unacceptable period of time (many hours).  I’m learning that the entire
market is served by just two fiber routes, through cities hundreds of miles
away in either direction.  So, basically two fiber cuts, potentially 1000+
miles apart, takes the entire region down.

If in fact there's only two fiber conduit approaches to the area,  he's
basically stuck no matter how many companies sell him bandwidth in those two
conduits. He can contract with 8 companies to have 4 paths through each
conduit, and 2 cable cuts *still* leave him dead in the water.


He is complaining about AS3356 in specific and claiming they COULD 
reroute around it but choose not to. This leads me to assume there are 
alternatives. Two places, Miami and Texas, are mentioned and that a 
double fault, one in Miami and another in Texas would bring down the 
network. I am from Europe, but am I to believe that Miami and Texas (or 
anywhere between those two) are served by only two fiber conduits? This 
would have several big states only connected two ways.


The question was if downtime on a transit provider of many hours is 
unacceptable. I am offering my experience that this happens to all of 
them. Some of them can have problems that last days not hours. Do not 
ever assume that a so called "tier 1" network is good as your only transit.


Also a total cut of from the world is the good kind of trouble they can 
have. That would just lead them to lose a large part of the global 
routing table. Your router will automatically choose one of your other 
transits. The bad kind of trouble is when they have packet loss to some 
few (but important) destinations and your customer thinks it is you that 
is having issues. And basically all you can do about it is to "shutdown" 
the session and wait until they fixed the issue.


I am offering the view that one might consider that kind of downtime 
unacceptable, but it is just a matter of fact that they all have it. The 
two options to avoid it is to buy from a smaller local ISP instead - one 
that has multiple transits. Or to have multiple transits yourself and be 
prepared to deal with it.


Regards

Baldur