Proxy types explained: datacenter, ISP, residential, and mobile

A proxy forwards a request so the target sees the proxy’s address instead of the client’s. Proxy types differ mainly in where the address comes from, how identifiable it is, and what consent exists from the person whose connection is used. Those differences matter more than the marketing labels.

The four types

Type	Address source	Detectability	Typical cost	Consent question
Datacenter	Cloud/hosting providers	High; ranges are known	Low	None; it is the provider’s address
ISP	ISP-assigned, hosted	Medium	Medium–high	Depends on the provider
Residential	Consumer devices	Lower	High	Does the device owner consent?
Mobile	Mobile carrier networks	Lower	Very high	Does the device owner consent?

“Residential” and “mobile” proxies route traffic through other people’s devices. A provider that cannot explain how it obtains consent is a risk. Using someone else’s connection to evade a block is not made lawful by the label.

Datacenter proxies

Datacenter proxies are the simplest and cheapest. They are useful for:

controlled egress from a private network;
caching and architecture;
testing a service from a known IP;
distributing load across your own infrastructure.

They are easy for targets to identify and block. If a site blocks datacenter ranges, that is often a deliberate policy. Do not treat a block as an invitation to switch to residential.

ISP proxies

ISP proxies use addresses registered to an ISP but hosted in a datacenter. They can look less like a cloud range than a pure datacenter proxy, but they are still a commercial service. The ethical questions are the same: what are you using it for, and does the target permit it?

Residential proxies

Residential proxies exit through consumer devices. They are marketed for ad verification, brand protection, and market research. The technical appeal is that the address looks like a home connection; the ethical issue is that the traffic uses someone else’s device and bandwidth.

Questions to ask a provider:

How are exit nodes recruited, and do users opt in?
Is there a clear acceptable-use policy?
Can a user see and revoke participation?
What data does the provider log?
Is the provider’s sourcing lawful in the relevant jurisdictions?

If the answers are vague, do not use the service.

Mobile proxies

Mobile proxies exit through carrier networks. They are the most expensive and the most sensitive because mobile devices are personal. The same consent and acceptable-use questions apply, with a higher bar.

What proxies do not change

Terms of service. The site’s rules apply to the request, not the IP.
robots.txt. It is addressed to crawlers, not addresses.
Copyright and database rights. The data is still the data.
Privacy law. Personal data is still personal data.
Access controls. A login, paywall, or CAPTCHA is still a boundary.

A proxy changes the network path, not the authorisation.

Configuration notes

Use http:// and https:// proxy URLs for HTTP proxies; use socks5:// for SOCKS5.
Keep credentials in environment variables or a secret manager.
Set a proxy connection timeout separate from the target timeout.
Log the proxy used per request for debugging, but do not log credentials.
Remove failing proxies from the pool instead of retrying them forever.
Test the exit IP with a service you control, not by querying a random “what is my IP” page on every request.

When to use a proxy

Your service needs a controlled egress.
You have a contractual right to access the data and need a stable IP.
You are testing geo-specific behaviour that the site permits.
You are caching or load-balancing at the edge.

When not to use one

To bypass a block, ban, paywall, login, or CAPTCHA.
To hide a crawler’s identity while ignoring robots.txt.
To collect personal data at scale without a lawful basis.
Because a vendor’s marketing says “unblock anything”.

The proxies and IP rotation guide covers the operational and ethical side. The terms of service guide covers the contractual side.