connection to ix

Mike Hammett nanog at ics-il.net
Thu Jan 18 13:36:36 UTC 2018


I doubt you're going to find IX-specific services among Caribbean islands. I also doubt that there is 10G wave-level of capacity needs to each IX, given that AMS-IX Caribbean has a peak of 19 Gb/s. 

Finding someone that does Carrier Ethernet or VPLS services to do aggregation would be the best hope, but even that is limited given the limited carrier selection in the Caribbean. 




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

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

----- Original Message -----

From: "Olivier Benghozi" <olivier.benghozi at wifirst.fr> 
To: "NANOG list" <nanog at nanog.org> 
Sent: Thursday, January 18, 2018 7:12:10 AM 
Subject: Re: connection to ix 

Sure, plenty of people use such service. 

If you have much traffic, you'll rely on one dedicated port for each IX, and maybe one full 10G wave or more toward each one (so the remote service is only transport, you have your own port at each IX). 

If not, you'll use ethernet over MPLS service for remote peering, provided by guys like IX-Reach, who will provide you shared bandwidth over their own connection at each IX, delivered on one single port (or more) toward you. 
It's shared bandwidth, so you must trust the provider network, in case of DDoS on its ports your traffic will be at risk, you don't see if the remote port is down. 
But the price can be quite more interesting, you don't have to deal with the remote physical interco, and so on. 

Now, you have to find someone providing remote peering to the IXs you want, to check the price... 

> Le 18 janv. 2018 à 13:24, Rudolph Daniel <rudi.daniel at gmail.com> a écrit : 
> 
> How do i remotely connect to an ix? 
> In the en. speaking caribbean we have probably six to eight small national 
> ixps. I want to be present at all of them, & its been suggested i can make 
> some cost savings by remote connection using mpls. Is this a viable 
> option? Is there a disadvantage in doing this? Thx. 
> 
> Rudi 





More information about the NANOG mailing list