I have stumbled upon this site [1] which seems to offer /27 IPv4 leasing.
They also claim "All of our IPv4 address space can be used on any network in any location."
I thought that the smallest prefix size one could get routed globally is /24?
So how does this work?
I have stumbled upon this site [1] which seems to offer /27 IPv4 leasing.
They also claim "All of our IPv4 address space can be used on any network
in any location."
I thought that the smallest prefix size one could get routed globally is
/24?
With the cooperation of your local ISP, it's possible to get clever about
this.
If your ISP sets its filter to allow it, you can send packets from the /27
directly without having to transit the GRE tunnel. So, half the path has no
latency hit at all.
The tunnel ingress which takes the /24 off the Internet and sends the /27
to you does not have to be a single node in a single location. GRE and IPIP
both support stateless multipoint tunnels where they can receive packets
from multiple sources. The /24 can be anycasted from multiple nodes around
the world allowing near-optimal routing from most origins.
IPv4 /24 is commonly the minimal chunk advertised to (and accepted by) neighbors. If I run a global (or regional) network, I may advertise this /24 -- or rather an aggregate covering it -- over my diverse interconnection with neighbors, your /27 being part of the chunk and routed to you internally (if you're va customer)-- no need for encapsulation efforts. Similar scenario may be multi-upstream, subject to acceptance of "punching holes in aggregates"... Am I missing something? What's the trigger for doing tunneling here?
With "IP address leasing" you aren't connected to the network which holds
the address registration.
For leasing less than a /24, they need a plan other than "advertise to your
peers with BGP" because even if your peer accepts a /27, most of their
peers will not.
Yes, we do this for several clients. We route them a smaller than 24 block over a tunnel.
Which bring up an interesting question. Will there be a time where the smallest block size recognized will be something smaller than a /24? /25, /26 ? Most modern routers have the horsepower to deal with larger route tables.
I know of dozens, if not hundreds of small ISPs that can’t participate in BGP because they don’t have big enough blocks. Many others who do are not utilizing their /24 so it just kinda sits there. They have to have their provider assigned IP space be advertised. Does not help them getting on to an IX though.
I know I know IPV6 is the answer not going to accepting smaller blocks.
Not much of an ISP if they can't get a /24. We're talking about a one-time
market purchase under $5000 and the ARIN justification for that small a
block almost writes itself.
What's the business model, if you have less than 120 customers? Selling
value-add services on top of moving the packets? Or just be in a country
where cost-of-everything is so cheap that you can make a profit on 120
customers at $20/mo?
And hundreds? Is that "in the US", or "worldwide"?
No disrespect, but anyone who can't afford to spend $5000 on resources
critical to their activity is not in the Internet business or any other
kind of business and should probably stop lying to themselves about that.
I can tell you that when we started (and there were IP's still available)
we first leased from another company to get our feet when and run tests
before we requested our own resources.
Most of the ones I know personally are doing CGN and have no real need for IP addresses. I know of Wireless ISPs with 2000 customers and only about 50 IPv4 addresses in use for nat and the occasional Public IP customer.