Backup Window and Throughput Calculator
A backup that does not finish before the first shift is a backup problem and an ERP performance problem. This calculator estimates how long fulls and incrementals take and how fast the link must be to fit your window.
Your numbers
Protocol overhead, disk speed and contention. 60 to 80 is common.
Source-side reduction before data crosses the link.
Your results
Throughput depends on source disk, backup software and network conditions. Validate with a timed test job.
Full backup hours vs link speed (Mbps)
Sweeps Backup link speed from 100 to 10000, holding your other inputs fixed.
Optional backup health review
Share your backup job summary and we will highlight windows that are close to overrunning. No obligation.
Optional. The tool and every export work without this. No spam.
How to use it
Enter the size of data in scope for a full backup and your daily change rate, which you can read from the backup software reports. Choose the link speed that the backup traffic uses, then set a realistic efficiency. If the software does source-side dedupe or compression, enter the observed ratio. Finally enter the time window available. You get the full duration, the incremental duration, how much of the window the incremental uses, and the link speed required for a full in-window.
- Use observed dedupe ratios, not vendor claims
- Plan fulls for weekends and incrementals for nights
- Test restore speed separately, it is usually slower
The method
Effective throughput in MB per second is link speed in Mbps times efficiency divided by 8. Data volume in MB is terabytes times 1,048,576, divided by the reduction ratio because less data crosses the wire. Duration in hours is data divided by throughput divided by 3,600. The required link speed formula inverts this: reduced data in megabits divided by window seconds, then divided by efficiency.
Worked example
Twelve terabytes with a 2x reduction sends about 6 TB. A 1 Gbps link at 70 percent efficiency moves roughly 87.5 MB per second, so the full takes about 20 hours, which will not fit an 8 hour window. A 3 percent change rate is 0.36 TB, which takes about 0.6 hours, comfortably inside the window. To do the full overnight you would need around 2.4 Gbps, so use weekend fulls or a faster path.
Common mistakes
Assuming the link is the bottleneck when the source disk or the backup proxy is, running backups while month-end jobs are active, and measuring only backup speed while ignoring restore time. Also check that application-consistent snapshots for the ERP database are taken before the transfer starts, and confirm that offsite copies finish before the next cycle.
Frequently Asked Questions
How long does it take to back up 1 TB over 1 Gbps?
At about 70 percent efficiency a 1 Gbps link moves roughly 87 MB per second, so 1 TB takes about 3.3 hours without any dedupe. With 2x reduction it is under 2 hours. Source disk speed can lower this.
What is a typical daily change rate?
Transactional ERP databases often change 2 to 5 percent a day, file shares 1 to 3 percent, and engineering or CAD vaults can spike when releases happen. Read the actual figure from your backup reports.
Why is my backup slower than the link speed suggests?
Protocol overhead, many small files, slow source disks, encryption, antivirus scans and competing traffic all reduce throughput. 60 to 80 percent of line rate is a common real-world result.
Does restore time matter as much as backup time?
Yes, restore speed defines your recovery time objective and is often slower than backup because of random writes and verification. Test a full restore at least once a year and record the time.
If backup logs and ERP job schedules are scattered across tools, Netray can build a private assistant that answers when and why a job overran, on your own infrastructure.
Related Tools
ERP WAN Bandwidth Sizing Calculator
Estimate the circuit size a plant or branch needs for ERP sessions, report downloads, voice and replication, with burst, overhead and utilization targets.
Industry TemplatesRansomware Downtime Cost Calculator for Manufacturers
Turn restore effort, detection time and plant revenue into a recovery time and a cost estimate, and see how faster recovery changes the loss.
ERP Migration & SelectionOn-Prem Server Refresh vs Cloud TCO Calculator
Put hardware, support, space, admin time and hosting fees on one page to compare a server refresh with a cloud hosting move over your planning horizon.
ERP OperationsPlant IT Change Request (CAB) Form Template
A change request form that covers scope, risk, rollback, production impact and approvals for IT changes that touch ERP, MES or shop floor systems.
More ERP Operations tools
See allKeep exploring
Free tools
- Infor LN Batch Job Window CalculatorEstimate batch duration from records, throughput and parallel processes, and find the processes needed to fit a nightly window.
- Oracle EBS Maintenance Window and Backup Time CalculatorWork out how long the outage needs to be, including a backup, the work itself, validation and a contingency buffer, and how long a rollback restore would take.
- Legacy ERP Data Cutover Window CalculatorEstimates final load duration, window utilisation and total rehearsal time from record volume, throughput and validation overhead.
ERP fixes and how-tos
- How to Back Up the JobBOSS DatabaseBack up and restore the JobBOSS SQL Server database: scheduled full and log backups, verification, SQL Express options and pre-upgrade snapshots.
- How to Back Up and Restore a QAD Progress Database with probkup and prorestStep by step guide to online backup of a QAD Progress database with probkup, incremental backups, verification, and restore with prorest, plus the traps to avoid.
- How to Back Up Fishbowl Inventory (and Restore It)How to back up and restore the Fishbowl Inventory database: built-in backup, scheduling, mysqldump, offsite copies, and restore testing.
Guides
- How to Set Up Database Backup Automation in SyteLineAutomate SyteLine database backups with SQL Server maintenance plans, transaction log management, and disaster recovery procedures for Infor CloudSuite Industrial.
- SyteLine Database Backup and Recovery StrategyPlan database backup and disaster recovery for Infor SyteLine SQL Server environments. Configure full, differential, and transaction log backup strategies.
- LLM Batching and Throughput Tuning: A Field GuideTune LLM inference batching and throughput: max-num-seqs, latency-throughput tradeoffs, load testing methodology, and scaling patterns that hold up.
Troubleshooting
- How to tune LLM inference batch size for the right throughput vs latency tradeoffBatch size trades throughput for latency, and the right value is the knee of that curve for your SLO, not the largest number that fits in memory. Here is how to find it.
- Why vLLM throughput is far below expected, and how to tune itLow vLLM throughput usually traces to max-num-seqs, chunked prefill, dtype, or CPU-bound tokenization. Diagnose and tune each in a step-by-step pass.