Re: Carriers

[Jay Hanke]

On Wed, Jan 22, 2020 at 9:12 AM Neader, Brent
 wrote:
>
> 5nines is in that building, they list available carriers on their spec sheet, 
> it is a pretty respectable selection.  Unsure of how the MMR or cross 
> connects work in building though.
>

It's a semi free for all. You'll want to get into the suite where most
of your connections are located.

Re: MAP-E

[Jay Hanke]

> I can't think of a public presentation off the top of my head that
> explains how each major transition technology works, and the pros and
> cons of each. There must be one, but it's hard to cover the major
> options in an hour.

Actually your post is better than a presentation. I was quite
surprised at the adoption rate of DS-Lite. There must be some pretty
decent B4 implementations with that many operators deployed.

Even though the spreadsheet is small sample size, there isn't much
DS-lite deployment in the US.

So from 10k feet, MAP-E is basically the same thing as DS-Lite except
for the location of the NAT?

Re: MAP-E

[Jay Hanke]

Is there a summary presentation someplace laying out the options that
are active in the wild with some deployment stats?

On Fri, Aug 2, 2019 at 10:34 AM JORDI PALET MARTINEZ via NANOG
 wrote:
>
> I understand that, but the inconvenient is the fix allocation of ports per 
> client, and not all the clients use the same number of ports. Every option 
> has good and bad things.
>
>
>
> MAP is less efficient in terms of maximizing the “use” of the existing IPv4 
> addresses.
>
>
>
> https://datatracker.ietf.org/doc/draft-lmhp-v6ops-transition-comparison/
>
>
>
>
>
> Regards,
>
> Jordi
>
> @jordipalet
>
>
>
>
>
>
>
> El 2/8/19 17:25, "NANOG en nombre de Baldur Norddahl" 
>  escribió:
>
>
>
> Hi Jordi
>
>
>
> My alternative to MAP-E is plain old NAT 444 dual stack. I am trying to avoid 
> the expense and operative nightmare of having to run a redundant NAT server 
> setup with thousands of users. MAP is the only alternative that avoids a 
> provider run NAT server.
>
>
>
> Regards,
>
>
>
> Baldur
>
>
>
>
>
> On Fri, Aug 2, 2019 at 3:38 PM JORDI PALET MARTINEZ via NANOG 
>  wrote:
>
> Ask the vendor to support RFC8585.
>
>
>
> Also, you can do it with OpenWRT.
>
>
>
> I think 464XLAT is a better option and both of them are supported by OpenWRT.
>
>
>
> You can also use OpenSource (Jool) for the NAT64.
>
>
>
> Regards,
>
> Jordi
>
> @jordipalet
>
>
>
>
>
>
>
> El 2/8/19 14:20, "NANOG en nombre de Baldur Norddahl" 
>  escribió:
>
>
>
> Hello
>
>
>
> Are there any known public deployments of MAP-E? What about CPE routers with 
> support?
>
>
>
> The pricing on IPv4 is now at USD 20/address so I am thinking we are forced 
> to go the CGN route going forward. Of all the options, MAP-E appears to be 
> the most elegant. Just add/remove some more headers on a packet and route it 
> as normal. No need to invest in anything as our core routers can already do 
> that. No worries about scale.
>
>
>
> BUT - our current CPE has zero support. We are too small that they will make 
> this feature just for us, so I need to convince them there is going to be a 
> demand. Alternatively I need to find a different CPE vendor that has MAP-E 
> support, but are there any?
>
>
>
> What is holding MAP-E back?  In my view MAP-E could be the end game for IPv4. 
> Customers get full IPv6 and enough of IPv4 to be somewhat compatible. The ISP 
> networks are not forced to do a lot of processing such as CGN otherwise 
> requires.
>
>
>
> I read some posts from Japan where users are reporting a deployment of MAP-E. 
> Anyone know about that?
>
>
>
> Regards,
>
>
>
> Baldur
>
>
>
>
> **
> IPv4 is over
> Are you ready for the new Internet ?
> http://www.theipv6company.com
> The IPv6 Company
>
> This electronic message contains information which may be privileged or 
> confidential. The information is intended to be for the exclusive use of the 
> individual(s) named above and further non-explicilty authorized disclosure, 
> copying, distribution or use of the contents of this information, even if 
> partially, including attached files, is strictly prohibited and will be 
> considered a criminal offense. If you are not the intended recipient be aware 
> that any disclosure, copying, distribution or use of the contents of this 
> information, even if partially, including attached files, is strictly 
> prohibited, will be considered a criminal offense, so you must reply to the 
> original sender to inform about this communication and delete it.
>
>
> **
> IPv4 is over
> Are you ready for the new Internet ?
> http://www.theipv6company.com
> The IPv6 Company
>
> This electronic message contains information which may be privileged or 
> confidential. The information is intended to be for the exclusive use of the 
> individual(s) named above and further non-explicilty authorized disclosure, 
> copying, distribution or use of the contents of this information, even if 
> partially, including attached files, is strictly prohibited and will be 
> considered a criminal offense. If you are not the intended recipient be aware 
> that any disclosure, copying, distribution or use of the contents of this 
> information, even if partially, including attached files, is strictly 
> prohibited, will be considered a criminal offense, so you must reply to the 
> original sender to inform about this communication and delete it.
>

Re: provider email maintenance standard

[Jay Hanke]

> https://github.com/jda/maintnote-std/blob/master/standard.md
>
> NTT / AS 2914’s NOC follows this process to keep customers and partners 
> informed about maintenances.

Is there commercial or open source software that already has this implemented?

Re: Non-profit IX vs. neutral for-profit IX

[Jay Hanke]

MICE is technically a cooperative not a non-profit. The fees cover the
costs and just the costs and the members are owners.

Also MICE does not provide any transport. All transport to remote locations
is provided by the network hosting the remote switch.

Jay

On Fri, Dec 21, 2018, 9:16 AM Mike Hammett  Someone's typically paying the difference in a non-profit IX. Someone's
> donating piles of cash, free dark fiber, free colo, etc. You're either
> paying your own way, or you have a port subsidized by someone else. There's
> not necessarily anything wrong with that, but you have to make sure you
> count that when you talk about "cost".
>
> They're also over twice the size, and in half the number of buildings (per
> PeeringDB, anyway). They've also been around over twice as long. Scale
> helps with cost.
>
>
>
> -
> Mike Hammett
> Intelligent Computing Solutions 
> 
> 
> 
> 
> Midwest Internet Exchange 
> 
> 
> 
> The Brothers WISP 
> 
> 
> --
> *From: *"Darin Steffl" 
> *To: *"Mike Hammett" 
> *Cc: *"Mehmet Akcin" , "NANOG Mailing List" <
> nanog@nanog.org>
> *Sent: *Friday, December 21, 2018 8:34:32 AM
> *Subject: *Re: Non-profit IX vs. neutral for-profit IX
>
> http://micemn.net/services.html
>
> MICE in Minneapolis is a great IX that we are on and their port fees are
> very reasonable. They used to be completely free up until this year. Even
> so, their fees are virtually nothing which encourages more operators to
> connect to it versus For-Profit IX's where sometimes the fees are almost as
> much as transit.
>
> For example Midwest-IX is $9,300 per year for a 10G port but MICE is only
> $250 per year. That's a HUGE difference and MICE also has way more peers
> and traffic overall due to how easy and cheap it is to join.
>
> On Fri, Dec 21, 2018 at 8:27 AM Mike Hammett  wrote:
>
>> Not all transit is cheap and not all transit is good quality, regardless
>> of what it costs. ;-)
>>
>> At our IX, we regularly see clients whose total network usage goes up
>> once they're on the IX.
>>
>>
>>
>> -
>> Mike Hammett
>> Intelligent Computing Solutions 
>> 
>> 
>> 
>> 
>> Midwest Internet Exchange 
>> 
>> 
>> 
>> The Brothers WISP 
>> 
>> 
>> --
>> *From: *"Mehmet Akcin" 
>> *To: *"Clayton Zekelman" 
>> *Cc: *"Mike Hammett" , "NANOG Mailing List" <
>> nanog@nanog.org>, "Tim Raphael" 
>> *Sent: *Friday, December 21, 2018 8:19:43 AM
>> *Subject: *Re: Non-profit IX vs. neutral for-profit IX
>>
>> Torix and Six are great examples.
>>
>> If you want to be for profit, make sure to publish port pricing and keep
>> it fair. Transit is cheap and good quality
>>
>> On Fri, Dec 21, 2018 at 08:14 Clayton Zekelman  wrote:
>>
>>>
>>> TorIX is a great example of a not for profit IX that is very successful.
>>>
>>> https://www.torix.ca/
>>>
>>> A very dedicated team of people provide an incredible level of service.
>>>
>>> Thave a very transparent process.  Their pricing is listed up front on
>>> their website:
>>>
>>> https://www.torix.ca/peering/#pricing
>>>
>>>
>>>
>>> At 09:03 AM 21/12/2018, Mike Hammett wrote:
>>>
>>> As far as neutral, I meant separate from the datacenters in which
>>> they're housed. People in NA seem to think there are only two kinds of
>>> IXes, Equinix, DRT, Coresite types and NWAX, SIX, MICE types.
>>>
>>>
>>>
>>> -
>>> Mike Hammett
>>> Intelligent Computing Solutions 
>>>
>>> Midwest Internet Exchange 
>>>
>>> The Brothers WISP 
>>>
>>> --
>>> *From: *"Tim Raphael" 
>>> *To: *"NANOG Mailing List" 
>>> *Sent: *Thursday, December 20, 2018 8:39:42 PM
>>> *Subject: *Re: Non-profit IX vs. neutral for-profit IX
>>>
>>> The other point to consider is that a NFP can justify more locations and
>>> offer services (such as extended reach) that don’t have the same profit
>>> margins or ROI as for-profits.
>>> This often leads to greater value to those with sm

Re: Moving fibre trunks: interruptions?

[Jay Hanke]

I'd expect at least a couple of hours of outage while the cable is reconnected.

When doing the move on the live cable (assuming 1 cable). There will
be a splicing crew at each end of the move. They will then break a
tube or ribbon at a time and splice into the new cable.

Splicing unused portions of the cable and then moving patches is also
done. In my experience, it's much more common to resplice on the
existing strands.

A large cable will take quite a while to resplice, likely more than
just overnight depending on the size of the cable.

Jay

On Fri, Sep 1, 2017 at 1:32 PM, Jean-Francois Mezei
 wrote:
>
> A large highway interchange is being rebuilt in Montréal (Turcot) and
> this requires that the CN mainline tracks out of downtown be moved a few
> hundred metres to the north for a couple of kilometres until it rejoins
> the existing alignment.
>
> Part of the contract involves the cost of moving the fibre trunks along
> with the tracks. (old alignment will become commercial properties).
>
>
> So they have new cable that goes through the new alignment and joins the
> old one at both ends.  So they'll have hundreds of strands to splice.
>
> When doing that type of work, how much downtime can be expected for each
> strand?
>
> Would they typically use patch panels in central offices to move a
> customer to a spare strand while they splice their assigned strand to
> use the new cable segment (and then move traffic back to that assigned
> strand?). Or would they switch customers around to new strands and
> update their documentation on which customer is on which strand?
>
> Or do they do nothing at patch panels in COs and just take whatever time
> it is needed to have crews at both ends of the work site splice each
> strand at same time (I assume about 5 minutes outage for each strand?)
>
>
> Would they normally involve the customer advising them of upcoming
> outage? Would the folks working trackside be limited to overnight hours
> to make outages less significant, or do they work around the clock ?
>

Re: Virtual or Remote Peering

[Jay Hanke]

I think you are talking about different applications of remote peering.

If you connect to a remote IX via transport the routing decision is
more along the lines is this packet destined to me. Having a router
sitting in the "remote" colo is of little value. It would not help to
keep the traffic local as there are only two paths. The router would
just forward between the ports on either side. A common application of
this is a "backup" IX to pick up content in the event of a failure at
the primary IX. A wave service is just a very long cross connect in
this regard.

If you provide services across the IX and start bouncing things
through remote ports (that could stay local). That is a different
animal.

Jay


On Thu, Aug 17, 2017 at 6:41 AM, Mike Hammett  wrote:
> A company you have a contractual arrangement with vs. random operators of 
> which neither you nor the end party have any relationship with. Which one's 
> unreliable, again?
>
> From a technical perspective:
> router located with IX > wave to IX > switched PtP\PtMP to IX > remote 
> peering service > transit
>
> Fiscally, it's almost the other way around, with where transit goes being 
> variable based on locations and volumes.
>
>
>
>
> -
> Mike Hammett
> Intelligent Computing Solutions
> http://www.ics-il.com
>
> Midwest-IX
> http://www.midwest-ix.com
>
> - Original Message -
>
> From: "Måns Nilsson" 
> To: "Mike Hammett" 
> Cc: nanog@nanog.org
> Sent: Thursday, August 17, 2017 12:42:21 AM
> Subject: Re: Virtual or Remote Peering
>
> Subject: Re: Virtual or Remote Peering Date: Wed, Aug 16, 2017 at 08:02:47AM 
> -0500 Quoting Mike Hammett (na...@ics-il.net):
>
 How well does this service work? I understand it usually involves 
 point-to-multipoint Switched Ethernet with VLANs and resold IX ports. 
 Sounds like a service for ISP that would like to peer, but have relatively 
 small volumes for peering purposes or lopsided volumes.
>
>>> Its like buying regular ip-transit, but worse.
>
>> That seems to be a rather lopsided opinion.
>
> You get connections to other operators over an unreliable path that you
> have no control over, and the opportunities to keep traffic local are
> limited. Adding to that, it is all your fault since your provider does
> not do L3 and can claim a very passive rôle in the process.
>
> Like transit, but worse.
>
> --
> Måns Nilsson primary/secondary/besserwisser/machina
> MN-1334-RIPE SA0XLR +46 705 989668
> YOW!! The land of the rising SONY!!
>

Re: Peering BOF/Peering social @NANOG69?

[Jay Hanke]

The peering social at previous NANOG meetings has been excellent and
very useful. As you mentioned, the peering personals are perhaps not
as valuable. It would be great to see the social portion come back in
some form.

Jay

On Mon, Feb 6, 2017 at 4:06 PM, Dave Temkin  wrote:
> The Peering Personals has been shelved while we try to figure out a better 
> option.
>
> There was no peering content submitted to the Program Committee that 
> justified a separate track, and so they chose to include the content in the 
> general session throughout the program.
>
> Regards,
>
> -Dave
>
> On Feb 6, 2017, 8:12 AM -0500, Matthew Petach , wrote:
>> I'm squinting at the Guidebook for NANOG69,
>> and I don't seem to see any peering BOF or
>> peering social this time around. Am I being
>> blind again, and it's on the agenda somewhere
>> but I'm just overlooking it?
>> Pointers in the right direction would be appreciated.
>>
>> Thanks! :)
>>
>> Matt

Re: Optical Wave Providers

[Jay Hanke]

There are lots of national carriers in the US. A much smaller number
of those carriers actually own the fiber cables. There are a handful
(Zayo, Level3, CenturyLink, Windstream, Earthlink, Verizon) that have
very large national, or semi-national foot prints.

The carriers frequently trade and lease strands of fiber from each
other to create a national network. Be careful on the commodity routes
diversity wise. There are a lot of places with 20+ carriers in the
same cable (or duct) each claiming to own the route.

Jay

On Wed, Aug 31, 2016 at 6:08 PM,   wrote:
> I have been looking at optical wave carriers for some long haul 1G/10G
> across the US. All to major cities and well known POP's.
> I am finding that there are not a lot of carriers who are offering wave
> services, usually just ethernet/MPLS.
> Particularly across the North west.
> Can someone shed some light on who some of the bigger carriers are and any
> challenges you have encountered with services like this?
> Who actually owns the fiber across the US?
>
> Thanks
>
> Tim

Mellanox 100GE switches

[Jay Hanke]

Does anyone have experiences they can share regarding the Mellanox
100GE Ethernet switches?

Thanks!

Jay

Re: Peering and Network Cost

[Jay Hanke]

Getting networks to connect to an ix is Uber expensive in relation to the
overall costs. Specifically before critical mass is reached. Getting the
first X gig of traffic is a hard problem that takes money to fix.
On Apr 19, 2015 7:51 AM, "Mike Hammett"  wrote:

> There is a revenue floor where it doesn't matter how much or how little
> service is provided, simply having a customer period requires a certain
> amount of revenue.
>
> Route servers, IXP Manager, AS112, route collectors, DNS, etc. all cost
> money.
>
> Maintenance costs money. The organization itself costs money. Upgrades
> cost money. Racks cost money. Power costs money.
>
> I'm sure I've left some things out.
>
>
>
>
> -
> Mike Hammett
> Intelligent Computing Solutions
> http://www.ics-il.com
>
>
>
> - Original Message -
>
> From: "Baldur Norddahl" 
> To: nanog@nanog.org
> Sent: Sunday, April 19, 2015 4:23:53 AM
> Subject: Re: Peering and Network Cost
>
> So why is IX peering so expensive?
>
> Again if I look at my local IX (dix.dk) they have about 40 networks
> connected. Each network pays minimum 5800 USD a year. That gives them a
> budget of 24+ USD a year.
>
> But the only service is running an old layer 2 switch.
>
> Why do these guys deserve to be paid that much for so little?
>
> Recently we had a competitor show up in the form of Netnod. However the
> pricing is almost exactly the same, although Netnod tries to deliver
> slightly more service.
>
> Seems to me that this an unsound market. The 40 dix particants should
> donate 1000 USD once and get a new layer 2 switch. Why does that not
> happen?
>
> Does not look like it is a local phenomenon either. IX'es all over are way
> more expensive than they should be.
>
> Regards
>
> Baldur
>
>

Re: Vancouver IXP - VanTX - BCNet

[Jay Hanke]

On Fri, Aug 23, 2013 at 5:59 AM, Niels Bakker wrote:

>
>> Indeed.  I think that ISPs who understand their business model well
>> enough to understand the effect the IXP will have on their
>> average-per-bit-delivery-cost is essential.  I think it's also essential
>> that they have some basic familiarity with the different ways IXPs can
>> fail, or fail to thrive, so that they can avoid mistakes others have made
>> in the past.  Over-spending, particularly on switches, is a huge killer of
>> IXPs.  Under-provisioning of circuits to the IXP is another big mistake.
>>  Failure to encourage local content and hosting is another.
>>
>
> Can you cite a few examples of an IXP going under because of overspending
> on switch hardware?  You call this a "huge killer" so there must be dozens
> you can choose from.anke
>

Having recently been through the startup process at an independent
non-profit IXP, I can see spending too much on hardware being a problem
particularly on supporting the ongoing hardware/software maintenance on the
switch. Something else to think about, if an exchange is given an endowment
from from outside entity it might be harder to build the commitment levels
of participants/volunteers because it's too easy to solve the problems with
money as opposed to member contributions.

We went through 3 switch upgrades in 2 years and IMHO it built a lot of
community.

Each IXP will have a set of faithful founders and the key is growing that
group beyond the initial group before the founders lose interest and move
on to bigger things. Particularly before connecting to the IXP makes
business sense to a company that won't connect just because it's cool.

When the exchange gets over the hump were companies are saving real money
it is much easier to keep the ball rolling and the exchange growing. Many
exchange points never make it to that level.

Jay

St Louis Internet Exchange

[Jay Hanke]

After a bit of googling, I found some references to an Internet
Exchange in St. Louis, MO called the St. Louis Regional Exchange.

Is this project still active?

Thanks,

Jay

Re: Cheap Juniper Gear for Lab

[Jay Hanke]

> We have 3 J2320s in the lab, all running 9.3R3.8. That's the last
> *real* JunOS (no session/flow tracking) for these boxes.
>

+1 on that. We have a number of 2300s in our lab for the same purpose
running 8.x code.

We also use Junosphere extensively, but nothing beats real hardware.
j2300s are cheap.

Jay

BGP MD5 at IXP

[Jay Hanke]

How critical is BGP MD5 at Internet Exchange Points? Would lack of
support for MD5 authentication on route servers prevent some peers
from multilaterally connecting? Do most exchange operators support it?

Thanks!

Jay

Re: Cogent & HE

[Jay Hanke]

On Wed, Jun 8, 2011 at 3:19 PM, Paul Stewart  wrote:
> Or peer with HE and buy transit from Cogent (or someone on Cogent's friendly
> list) - this is where I think their strategy is going to go after a while
> with a lot of folks (if they have the option - that's the key).  HE will
> peer with anyone I believe - Cogent has much more stringent "tier1" rules on
> peering.

How divided is the table? I see about 98 routes transiting Cogent ASN
via a HE connection. Customer has only has HE as v6 upstream. An
previous post listed about a 1300 prefix difference. That's pretty
significant unless it's due to aggregation or something. I'd also be
interested to see the size of the other major carriers v6 tables so I
can patch a whole until the other upstream is ready.

Jay

Small IXP [was Alaska IXP?]

[Jay Hanke]

[snip]

> Does anybody have some numbers they're able to share?  In the "two small
ISPs
> in the boonies" scenario, *is* there enough cross traffic to make an
> interconnect worth it? (I'd expect that gaming/IM/email across town to a
friend
> on The Other ISP would dominate here?) Or are both competitors too busy
> carrying customer traffic to the same sites elsewhere (google, youtube,
amazon,
> etc)?  Phrased differently, how big/small a cross-connect is worth the
effort?
> 

Or at the cogent website ($4/meg) do the cost justify peering anymore?

Obviously some of this always depends on the loop costs.

Going to try to write something up that would be useful for smaller ISPs.

The BGP barrier IMHO is quite high in most cases, not all the small ISPs
carry their routes out to the edge in the same manner as the larger SPs.

- Jared

In our efforts, BGP hasn't come up as often as the Cogent (low cost) issue.
I think there are two aspects, one is the opportunity. If you need to build
or bury it gets pretty tough to keep costs below $4/meg. The second is
traffic volume, if you can set up a peering connection for $200 per month
for a full GE you need to stuff 50 Mb/s over the link to break even. That
may be tough unless you have an anchor institution like a college or a
content network. Rural wholesale (delivered to ISP) is going at $50-60 per
Mb in large parts of the US. That brings the breakeven to about 4 Mb which
is much easier for the small guys.

I think the dominate application driving cross connects right now is might
be business VPN between the small ISPs either at L2 or L3. 

Also, keep in mind though the cheap Internet is only at a limited number of
metro area and you still need to pay to transport that Internet back to your
network.

jay

RE: Alaska IXP?

[Jay Hanke]

On 3/4/10 8:57 AM, "Jay Hanke"  wrote:

>>
>> We've seen the same issues in Minnesota. Locally referred to as the
"Chicago
>>. Problem". Adding on to point 3, there is also a lack of neutral
facilities
>> with a sufficient amount of traffic to justify the next carrier
connecting.
>> In rural areas many times the two ISPs that provide services are enemies
at
>> the business level. A couple of us have started to talk about starting an
>> exchange point. With transit being so cheap it is sometimes difficult to
>> justify paying for the x-connects for a small piece of the routing table.
>>
>> Have you considered starting your own exchange point with some of the
local
>> players? Just having the connectivity in place may help with DR
situations
>> in addition to all of the benefits of an exchange point.
>
>Any interest by other anchor tenants in the area, such as the higher
>education facilities? In Madison, we have MadIX[1], an exchange point
hosted
>by the University of Wisconsin-Madison, with a presence in one of the
>neutral carrier hotels in Madison.
>
>That eliminates the carrier to carrier issues you run into in the smaller
>cities, also helps with the "Chicago Problem" which we are very familiar
>with here as well.
>
>[1] http://kb.wisc.edu/ns/page.php?id=6636
>
>Andrew

>From the looks of the link it looks like there is a bit of traction at the
MadIX. One of the other interested carriers has talked to the University of
MN and they showed some interest in participating. The trick is getting the
first couple of participants to get to critical mass. Is the MadIX using a
route server or is it strictly layer2?

Thanks,

Jay

RE: Alaska IXP?

[Jay Hanke]

On Mar 4, 2010, at 8:13 AM, Sean Donelan wrote:

> On Wed, 3 Mar 2010, Antonio Querubin wrote:
>> On Wed, 3 Mar 2010, Sean Donelan wrote:
>> 
>>> Are there any common locations in Alaska where multiple local ISPs
exchange traffic, either transit or peering?  Or is Seattle the closest
exchange point for Alaska ISPs?
>> 
>> peeringdb.com lists only SIX (in Seattle) and PAIX Seattle.
> 
> Thanks and also thanks to the other folks that replied privately.  That
matches basically what I had found, but I wanted to check.
> 
> Transit is also ok, I'm doing the usual minimum connections/maximum
communications in case of (earthquake, volcano, tsunomi, etc) math.  Is
> there someplace in Anchorage that buying transit or peering from one or a
few ISPs is significant enough, or is it going back to Seattle anyway and
> the local ISPs already have done the math.
>
>What I've seen is that in smaller markets (in my previous life), eg:
Michigan, even when the providers are all in the same facility they
>
>1) Lacked understanding of traffic-patterns to understand peering savings
>2) Lacked ability to interconnect (eg: no switch on-site, no bgp/routing
capability)
>3) CLEC or other colo provider prohibited #2
>
>This meant traffic would regularly be diverted to Chicago or similar for
exchange between local ISPs.

>The one time I was able to pull off a local facility cross-connect, it was
difficult to get it at a speed greater than 10megs (this was 1999 or so).

>With the dropping metro-ethernet/ftth type equipment that can do 1G for
"cheap", perhaps a short fiber build for x-connect would help faciltiate
things >these days.  (i should model that and post the results).
>- Jared


We've seen the same issues in Minnesota. Locally referred to as the "Chicago
Problem". Adding on to point 3, there is also a lack of neutral facilities
with a sufficient amount of traffic to justify the next carrier connecting.
In rural areas many times the two ISPs that provide services are enemies at
the business level. A couple of us have started to talk about starting an
exchange point. With transit being so cheap it is sometimes difficult to
justify paying for the x-connects for a small piece of the routing table.

Have you considered starting your own exchange point with some of the local
players? Just having the connectivity in place may help with DR situations
in addition to all of the benefits of an exchange point.

I would also be very interested in seeing any modeling on the subject. There
was a document a couple of years ago that was pretty good talking about when
to peer but if memory serves it was more focused on the larger carriers.

Jay