I did a search online regarding how a smartphone gets an ip address over a cellular network, but I couldn’t find information in the first page of Google search, so I thought I might ask a question in the networking forum (maybe it should belong in “Uncategorized?”).
As the subject mentions, how does a smartphone get an IP address from T-Mobile, AT&T, Verizon, or any other cellular companies once the smartphone is subscribed to the cell phone provider? Does a smartphone get its own dynamic public IP address or does it get an IP address from a Carrier-Grade NAT (CG-NAT)?
I saw the page titled “Does a phone have an IP address,” but that isn’t the question I asked on Google. I am playing with Cisco Packet Tracer and I am thinking about adding cell phone networks into my global networking project. I’m using BGP for ISP-to-ISP networking. This is the challenging big project I am working on.
Thanks. It seems like I should implement private IP addressing for smartphones in Packet Tracer. I’m not sure what block of private IP addresses carriers use though. Maybe 10.x.x.x/8?
(A few minutes later…)
Oh, I should have done a search! Yep. Carriers do use CGNAT and I saw it in the picture below the CGNAT section of the article:
So I did answer my question. However, I might leave this thread as is for anyone who might have the same question as I did.
(A few minutes later while reading the article…) Aha… So the 100.64.0.x/24 is used by ISPs for CGNAT. Interesting. Can multiple ISPs that use CGNAT use the same 100.64.0.x/24 as an address space for customers? I am not sure if 100.64.0.x/24 is publicly-routable…
You should check out EVE-ng for building virtual networks. I use it for all my labbing and it supports lots more devices than what PT offers. Also, since you are running actual code you won’t have the limitations of PT.
The max number of nodes is 63. If I am right about this, each switch, router, PC, and server is a node. I’m not sure if EVE-ng supports clusters that groups the nodes into a single cluster that Packet Tracer does.
What I have in my Packet Tracer project is I have 3 multi-layer switches, a router, a single computer, a server that provides DHCP services, and between 6 to 7 bridges that function as media converters for converting RJ45 to SFP for long-distance communication between different ISPs and to residential and business networks. The router uses NAT for a server and computer. Add all of that and multiply by 9 cities and I can easily go over 63 nodes, not to mention that a business network would have as little as 2 routers, 2 distributed/core switches, and 3 access switches and as large as two routers, two core switches, 3 distribution switches, and 5 access switches. Plus, computers and servers would connect to access switches.
Purchasing EVE-ng PRO will set me back $117 for a 1-year license with 1024 nodes. No, this won’t do it for me. And besides, Packet Tracer does not limit me to 63 nodes. Sure, 1024 nodes is great, but unlike Packet Tracer, I do not see myself spending $120 every year.
That’s one of the limitations of PT. I do pay for my EVE-ng license, but I consider it the cost of doing business being in IT. Another free option to consider is GNS3. You still have an option to run actual code and many other vendors are supported as well.
I don’t think I can make use of 6to4 given the limitations of Packet Tracer. Well, better stick with IPv4 in Packet Tracer until I have the money for Cisco equipment or maybe plan on spending $115/year (€99) so I can make use of 1024 nodes max. I might have questions regarding how carriers use of IPv4 NAT pools. This could be interesting. Might be beyond my knowledge until I gain experience in network engineering. I’m not sure if CCNP or CCIE covers that.