Proxy Pricing Models
The unit a provider charges in: a gigabyte, an address per month, a port or a day. Two providers can look equally priced and differ several times over on the same job.
There are four common units and they answer different questions. Per gigabyte suits work whose size follows the pages you pull. Per address per month suits work that needs the same address to persist. Per day suits mobile, where the cost is the modem rather than the traffic. Per port is the same idea as per address, wearing different words.
Comparing quotes across units is where most of the confusion lives. A cheap per-gigabyte rate is expensive for a job that holds one address for a month and pulls almost nothing. A cheap monthly address is expensive for a job that needs a thousand different exits for one afternoon.
What counts as traffic is the second question, and it is asked far less often than it should be. Metering normally covers both directions and includes protocol overhead: request headers, response headers, TLS handshakes and every retry. A failed request that returned a 403 still moved bytes, and those bytes are billable.
This is why an estimate built from page counts is unreliable and an estimate built from measured page weight is not. Twenty pages fetched through the exit you intend to use will tell you more than any calculator.
What the meter actually sees
A proxy counts bytes crossing it, not documents delivered. That includes the request line and headers going out, the response headers coming back, TLS negotiation and anything a redirect chain added along the way.
Images, fonts, video and analytics scripts are the bulk of a modern page and they cross the same meter as the data you wanted. Blocking them is not an optimisation, it is the difference between two invoices.
On per-address products the meter is irrelevant: traffic through an ISP or datacenter address is not metered here, so the cost is fixed regardless of how heavy the pages turn out to be.
Estimating before committing
Measure the weight, then multiply. The measurement takes a minute and removes the guesswork:
From twenty pages to a monthly figure
# average bytes per page through the exit you will actually use
for i in $(seq 1 20); do
curl -o /dev/null -s -w "%{size_download}
" \
-x login_c_DE:password@proxy.sotaproxy.com:10000 https://target.example/product/$i
done | awk '{s+=$1} END {print s/NR " bytes average"}'
# pages per day multiplied by that average, divided by a gigabyte, times $1.00- Run the measurement with your real client settings. A curl fetch and a headless browser produce different numbers for the same page.
- Our residential is billed per gigabyte in both directions with headers included, and traffic does not expire while the package is active.
- ISP at $1.90 and datacenter at $1.15 per address per month are not metered, which makes them the cheaper answer for anything long-running on one address.
Where comparisons go wrong
Comparing rates instead of invoices
The rate is per unit. The invoice is the rate times your actual volume, and volume depends on page weight rather than page count.
Assuming failed requests are free
A refused request still carried headers in both directions. It is small, but on a large crawl it is not nothing.
Ignoring the minimum
Committed monthly volumes and expiring traffic are where cheap per-gigabyte prices are recovered. We have neither, which is a claim you can check on the pricing page.
Treating per-day as expensive by default
Mobile at $5.76 a day is dear for volume and reasonable for one profile that must not change address for a month.
Пов'язані терміни
Дивись на практиці
Готовий використовувати proxy pricing models?
SotaProxy надає доступ до ротуючих резидентських, мобільних, дата-центр та ISP проксі. Без мінімальних платежів.
Почати