Dedicated ports and dedicated IPs are not the same thing

A customer once bought a 25-port plan, tested the exit addresses, found that they changed between requests and that some of them had clearly been used elsewhere, and asked for a refund on the grounds that we had sold him something we did not deliver. He had bought twenty-five dedicated ports. He had read that as twenty-five dedicated IPs — addresses nobody else on earth would ever touch.

He was not being unreasonable, and neither were we. We were describing two different layers of the same connection using the same overloaded word. That misunderstanding is the single most expensive one in proxies, and it is entirely avoidable once you can see the two layers separately.

The two layers

Every proxied request has a front door and a back door.

The port is the front door. It is your private entry point into the network: a host and a number that you connect to, with your credentials on it. When a plan gives you twenty-five ports, you have twenty-five separate doors that are yours. Nobody else connects through them.

The IP is the back door. It is the address the destination website actually sees when your request arrives. It is the thing that shows up in a server log, gets rate-limited, gets recognised, or gets challenged.

Everything else falls out of that split. A dedicated port is a statement about who can send traffic. A dedicated IP is a statement about who can be seen sending it. They are related, but they are not the same claim, and a vendor can honestly make the first without making the second.

What a dedicated port actually buys you

On Budget Unlimited you buy ports — 5, 15, 25, 50 or 100 of them. What you own is a fixed number of simultaneous ways into the network, each independently configured. In the Configuration tab you set a country per port and a rotation interval per port, anywhere from one minute to twenty-four hours. Port 1 can sit on Germany changing every hour while port 2 runs the United States changing every minute.

That is real, and it is genuinely yours. The bandwidth through those doors is not metered, the configuration is not shared, and nobody else is competing with you for that entry point.

What it does not buy you is ownership of the addresses on the other side. Each port draws its exit from a large rotating pool. The address changes on your schedule, and an address that served you at ten past two may serve someone else at half past four. That is not a flaw in the product; it is the mechanism by which a rotating pool works at all. Breadth and exclusivity are opposites. You cannot have a very wide pool of addresses and have every address in it reserved for you.

Premium Unlimited sits in the same family with a different dial: it hands you a fresh exit on every new connection, and you buy a speed tier — 200 through 1000 Mbps — rather than a port count. Same principle. Your access is yours; the addresses circulate.

What a dedicated IP actually buys you

Static ISP is the other end of the scale, and it is the product people are usually imagining when they say “dedicated”.

Each line is a static US residential address, allocated to you and single-tenant for the whole term. It does not rotate. Nothing else routes through it while it is yours. Bandwidth is unmetered, it bills at a flat monthly rate per IP, and orders start at five IPs.

You are buying the back door this time. The destination sees the same address on Monday that it saw on Friday, and no other customer’s traffic has shaped that address in between. That is a much stronger claim than a dedicated port, it costs more, and it comes with a real trade-off: a fixed set of US lines is fixed and US-only. It will not cover a job that needs forty countries.

Concurrency lives on ports. Identity lives on IPs.

That sentence is the whole article, and it is the one worth keeping.

If your constraint is how much you can do at once — parallel workers, simultaneous sweeps, throughput — you are shopping for ports or a speed tier. Adding dedicated IPs will not help; five static lines run just as parallel as one does. We wrote about sizing that dimension properly in how many proxies do you actually need, and about how it lands on the bill in per-GB, per-port, or unlimited.

If your constraint is being recognised as the same visitor — logins, allowlists, dashboards, anything with an account attached — you are shopping for a stable exit address, and no number of ports will get you there. That is the layer a sticky session touches too, which is covered in rotating or sticky.

Most disappointment in this industry is someone buying hard on one axis to solve a problem that lives on the other.

The thirty-second test

Do not take anyone’s word for which layer you bought, ours included. Connect through the same port twice and read what comes back. Substitute your own credentials, endpoint and port from the dashboard:

curl -x "http://user123:pass456@<your-endpoint>:<your-port>" https://api.ipify.org; echo
curl -x "http://user123:pass456@<your-endpoint>:<your-port>" https://api.ipify.org; echo

Two different addresses means you are on a rotating pool: the port is yours, the exit is borrowed. That is expected on Premium Unlimited, which takes a fresh exit per connection.

Two identical addresses means the exit is being held. On Static ISP that is permanent — run it again next week and you will get the same answer. On Budget Unlimited it means you are inside your configured rotation window, so wait out the interval you set on that port and run it a third time; it should move.

Thirty seconds of terminal beats an afternoon of assuming. If the result is not what your plan should be doing, that is a concrete thing to send us, along the same lines as the checks in why a proxy connection gets reset.

The question to ask any vendor

“Dedicated” is not a defined term in this industry. Nobody is policing it, and different providers attach it to different layers in good faith. So do not ask whether something is dedicated. Ask this instead:

“Can the exit IP I get on this plan be served to another customer at any point?”

There are only two honest answers. Yes, it comes from a shared rotating pool — which is what a port-based or per-GB plan means, and which is fine for most work. Or no, it is allocated to you alone for the term — which is what a static, single-tenant address means, and which costs accordingly.

If you get a third answer, or a paragraph instead of one of those two, treat that as data.

Which one for which job

Rotating pools for scale and coverage: sweeps, discovery, price and availability checks, anything spread across many countries where each request stands on its own. Ports on Budget Unlimited, a speed tier on Premium Unlimited, or per-GB on Platinum when you need state, city or ASN precision.

Static addresses for identity: account work, services with IP allowlists, dashboards and panels that treat a new address as a new login to challenge, and anything that must be recognised as the same visitor tomorrow. That is Static ISP.

Plenty of teams run both on one account and pick per job. The credential pattern is identical either way, and HTTP and SOCKS5 sit on the same port, so moving a job from one to the other is a configuration change, not a migration.

If you are unsure which layer your problem actually lives on, describe the job — how many things at once, and whether the destination needs to recognise you — and we will tell you which of the two you are shopping for, including when the answer is the cheaper plan. More detail on both sides in ISP or residential and at docs.proxyomega.com.