ERP OperationsFree calculatorPDF, Excel, Word export

ERP WAN Bandwidth Sizing Calculator for Plants and Remote Sites

Plants ask carriers for bandwidth by gut feel and then blame the ERP when screens lag. Most slow sessions are latency or packet loss, but undersized circuits do cause real queueing. This calculator builds the number from users, reports, voice and replication.

Your numbers

Concurrent ERP users at site

users

Peak concurrent sessions, not total accounts.

Interactive traffic per user

kbps

Thin client or web ERP is often 50 to 200 kbps. Measure your own.

Report and file downloads per user

MB per hour

Burst factor on interactive traffic

2 x

Concurrent voice or video calls

calls

Backup or replication traffic in business hours

Mbps

Protocol and VPN overhead

20 %

Target peak link utilization

70 %

Your results

Total peak demand with overhead

41.2352

Recommended circuit size

58.9074

Round up to the next carrier tier in Mbps.

Interactive ERP traffic

Report and file traffic

5.3333

MB per hour converted to Mbps.

Peak interactive traffic

28.6667

Voice and video

0.696

About 87 kbps per G.711 call with headers. Video needs far more.

Planning estimate. Measure real flows with NetFlow or your firewall reports before ordering a circuit, and add latency requirements separately.

Free, optional. Reopen it on any device.

Save results:

Recommended circuit (Mbps) vs concurrent users

Sweeps Concurrent ERP users at site from 5 to 400, holding your other inputs fixed.

Optional network readiness check

Tell us your sites and ERP and we will flag the links most likely to hurt user experience. No obligation.

Optional. The tool and every export work without this. No spam.

Was this tool useful?

How to use it

Enter the peak number of concurrent users at the site. For interactive traffic per user, start with 100 to 200 kbps for RDP or web ERP, then replace it with a measured figure from your firewall or NetFlow. Add the average data users pull each hour such as reports, attachments and drawings. Include concurrent calls and any backup replication that runs in the working day, then set overhead and a target peak utilization.

Use the busiest 15 minutes, not the daily average
Count both directions and size for the busier one
Keep drawings, CAD and software updates out of the average

The method

Interactive demand is users times kbps. Report demand converts MB per hour into Mbps using 8 bits per byte over 3,600 seconds. Their sum is multiplied by a burst factor, since people click at the same moment after a shift change. Voice and replication are added, then protocol and VPN overhead. Dividing by the utilization target leaves headroom, because queueing delay rises quickly beyond about 70 percent load.

Worked example

Sixty users at 150 kbps is 9 Mbps. Reports at 40 MB per hour per user add about 5.3 Mbps. With a 2x burst the interactive peak is 28.7 Mbps. Eight calls add 0.7 Mbps and replication 5 Mbps. With 20 percent overhead the total is about 41 Mbps, and at 70 percent utilization the recommended circuit is around 59 Mbps, so a 100 Mbps tier is the sensible purchase.

Common mistakes

Sizing from total employee count, forgetting that shop floor terminals and Wi-Fi scanners also generate traffic, and running full backups across the same circuit as production. Also check latency: a clean 20 ms path to a cloud ERP matters more than raw megabits. Use QoS to protect voice and ERP traffic, and schedule large transfers outside shift hours.

Frequently Asked Questions

How much bandwidth does one ERP user need?

Typically 50 to 200 kbps for interactive sessions over RDP or a web client, plus bursts for reports and attachments. Heavy use of drawings or BI dashboards can add several megabits per user, so measure your own traffic.

Is latency or bandwidth more important for ERP?

For chatty client-server ERP and RDP, latency and packet loss usually matter more than raw bandwidth. Aim for low and stable round trip time. Extra bandwidth does not fix a poor path.

Why use a 70 percent utilization target?

Queueing delay and packet drops rise sharply as a link approaches saturation, and bursts will exceed the average. Designing for a peak of about 70 percent leaves room for bursts and growth.

Should I include backup traffic?

Include anything that runs during working hours. If backups or replication run overnight on a schedule, exclude them from the daytime number but check they finish inside the window.

If you are weighing a cloud move or an on-prem private AI layer for the ERP, Netray can help you map which traffic must leave the plant and which can stay local.

Talk to an Expert Explore More Free Tools