Gbps → TB / Day Egress

Sustained 1 Gbps equates to approximately 10.8 TB transferred per 24-hour period. This tool converts between network throughput (Gbps) and daily transfer volume (TB) for cloud egress cost modeling and CDN capacity planning.

Bidirectional Egress Calculator

Convert between sustained Gbps throughput and 24-hour transfer volume in both directions.

Cloud Egress Cost Reference

Estimated monthly cloud egress costs at various throughput levels (pricing approximated).

Throughput (Gbps) TB/Month Est. AWS Cost
0.132.4$2,754
1324$26,798
51,620$133,488
103,240$263,088
4012,960$1,049,088
10032,400$2,620,488

Gbps → TB/Day: The Egress Equation

The conversion is straightforward: 1 Gbps sustained for 24 hours = (1 × 10⁹ bits/s × 86,400 s) ÷ (8 bits/byte × 10¹² bytes/TB) ≈ 10.8 TB/day. This assumes 100% link utilization — in practice, bursty traffic patterns, TCP slow-start, and inter-packet gaps reduce effective throughput by 10–30%.

For cloud egress planning, this is the single most important conversion to internalize. A 10 Gbps Direct Connect link can push up to 3,240 TB/month — at AWS standard egress pricing (~$0.09/GB after the first 10 TB), that's over $263,000/month in data transfer charges alone. Enterprises often discover this only after their first cloud bill shock. Budget egress capacity and costs with this tool before architecting cross-region or hybrid-cloud data pipelines.

Decimal and binary again: 10.8 TB is not 9.83 TiB

1 Gbps sustained for 24 hours is (109 bits/s × 86,400 s) ÷ (8 bits/byte × 1012 bytes/TB) = 10.8 TB/day. In binary units the same traffic is 9.83 TiB/day, because a tebibyte is 1.0995 × 1012 bytes. Cloud billing uses decimal units, storage dashboards usually use binary ones, and a 9% discrepancy between a transfer log and an invoice is almost always this rather than a rounding error.

Worked example: the 50 Mbps line that costs $1,458 a month

Small sustained rates are where this conversion earns its keep, because a continuous trickle becomes a large monthly total. Take a 50 Mbps link running flat out:

50 × 106 bits/s × 2,592,000 s/month ÷ 8 ÷ 1012 = 16.2 TB/month
At AWS standard egress of roughly $0.09/GB that is 16,200 GB × $0.09 ≈ $1,458/month.

Fifty megabits per second is a video stream, a log shipper, a cross-region replication feed — nothing that registers as a capacity problem. The billing unit is the thing to watch, and it accumulates at 10.8 TB per Gbps per day regardless of how unremarkable the instantaneous rate looks.

Asymmetry is the architectural fact

Ingress is free on essentially every major cloud. Egress is metered. That single asymmetry should drive more architecture than it usually does: put data where it will be consumed, replicate toward readers rather than pulling toward a central store, and treat any regularly scheduled cross-boundary pull as a design smell. A nightly job fetching 500 GB from a central region costs about $450/month purely in transfer — the same data pushed to the edge once, or read in place, does not.

Peak-based thinking does not transfer either. Enterprise circuits are often billed on 95th-percentile utilisation, which rewards flattening peaks. Cloud egress is billed on total volume, so flattening a peak changes nothing and reducing the total is the only lever. Teams arriving from network engineering habitually optimise the wrong variable for their first several months on cloud.

What CDN offload actually changes

A CDN does not reduce egress, it relocates it. With a 95% cache hit ratio, origin egress falls twentyfold — a service pushing 100 TB/month to viewers drops to 5 TB/month from origin, about $450/month. But the CDN's own delivery charges are a separate line item, usually on the same order of magnitude. The saving comes from hitting cache and from CDN egress rates sitting below direct object-storage egress — not from making bytes disappear. Planning an offload migration around "95% of traffic goes away" overstates the benefit by the difference between the two rate cards.