← All Tools

NAS File Transfer Speed Estimator

This NAS file transfer speed estimator calculates how long a file transfer will actually take based on your connection type, NAS drive speed, and file size. Covers LAN, 10GbE, and WAN (remote access) scenarios with Australian NBN upload context.

Upgrading to 10GbE doesn't guarantee 10GbE speeds. Your actual file transfer speed is limited by the slowest link in the chain, your network, your computer's drive, your NAS storage, or the type of files you're moving.

This estimator models the bottleneck. Select your network, drives, NAS storage, and workload type to see a realistic speed range and find out what's actually limiting your transfers.

Your Setup

Jumbo Frames (9000 MTU)
SSD Read Cache
Enter GB to calculate time-to-transfer
Estimated Read Speed
- MB/s
megabytes per second
Real-world speeds depend on hardware, protocol, network conditions, and workload.
Primary bottleneck: -

Time to Transfer

100 GB
1 TB

Times shown as fastest-slowest based on estimated speed range.

Component Breakdown

Network
Client drive
NAS storage
Estimated result

Not sure which connection type you have?

Get the free NBN Remote Access Logic Tree, one page covering every major AU ISP, whether they use CGNAT, and the fastest fix for each situation.

WAN / Internet Transfer Speed

Estimate how long a transfer would take over your internet connection. Select your ISP plan to pre-fill speeds.

Connection preset:

Defaults to Australian NBN tiers, if you're outside Australia, select your plan above or enter speeds manually.

Upload: 100 GB
Download: 100 GB
Upload: 1 TB
Download: 1 TB

Times assume 100% line utilisation. Real-world transfers typically achieve 80-90% of advertised speeds due to overhead and contention.

AU NBN Speed Tiers: Real-World NAS Transfer Rates

NBN speed tiers are advertised as download/upload Mbps. Upload speed is the binding constraint for remote NAS access and off-site backup , most residential plans cap upload at 20-50 Mbps regardless of download tier.

NBN Plan Download Upload Real-world DL Real-world UL
NBN 2525 Mbps / 3.1 MB/s5 Mbps / 0.6 MB/s~2.5 MB/s~0.5 MB/s
NBN 5050 Mbps / 6.3 MB/s20 Mbps / 2.5 MB/s~5 MB/s~2 MB/s
NBN 100 (most common)100 Mbps / 12.5 MB/s20 Mbps / 2.5 MB/s~10 MB/s~2 MB/s
NBN 250250 Mbps / 31 MB/s25 Mbps / 3.1 MB/s~25 MB/s~2.5 MB/s
NBN 10001000 Mbps / 125 MB/s50 Mbps / 6.3 MB/s~90 MB/s~5 MB/s

Key insight: Moving from NBN 100 to NBN 250 or NBN 1000 dramatically improves download but upload stays at 20-50 Mbps on most FTTC/FTTN connections. For remote NAS access and off-site backup, upload is what matters, not download tier.

LAN Protocol Overhead: SMB vs NFS on 1 Gbps

1 Gbps Ethernet has a theoretical ceiling of 125 MB/s. Protocol choice and file size determine how close you get.

Protocol Common use Large file reads Small file reads
SMB 3.xWindows, macOS Finder80-112 MB/s5-30 MB/s
NFS v4Linux clients, VMs, TrueNAS90-118 MB/s10-40 MB/s
SMB Multichannel (2× 1GbE)Windows + Synology dual-port160-220 MB/s10-50 MB/s
AFP (legacy macOS)Old macOS clients (deprecated)50-90 MB/s3-15 MB/s

For most AU home users on 1 Gbps LAN, SMB to a 4-bay Synology or QNAP NAS delivers 80-110 MB/s for large sequential files. The network becomes the ceiling before HDD array throughput in most 4-bay RAID 5 setups, so a 10GbE upgrade only helps if your drives exceed ~115 MB/s aggregate.

Example: Synology DS925+ Over SMB on 1 Gbps LAN

A DS925+ (4-bay, AMD Ryzen R1600) with 4× Seagate IronWolf 6 TB drives in RAID 5:

CGNAT affects most AU residential users. Most Australian ISPs place residential customers behind CGNAT (Carrier-Grade NAT), which blocks standard port-forwarding for remote NAS access. Synology QuickConnect bypasses CGNAT via relay servers, but relay-routed traffic is typically capped at 5-10 MB/s. For faster remote access, you need a static IP (ask your ISP, usually $10-20/month extra) or a WireGuard VPN tunnel on the NAS.

How We Calculate

The estimator uses a bottleneck model with four components:

Your estimated speed is the minimum of all four components. The "primary limiter" shows which component is the bottleneck.

All throughput ranges are based on published benchmarks, manufacturer specifications, and typical real-world observations. They are shown as ranges rather than single numbers because actual performance varies with hardware, firmware, protocol overhead, and workload patterns.

Last reviewed: 20 March 2026

Frequently Asked Questions

Why do you show a range instead of an exact speed? +
Real-world transfer speeds depend on many variables, network overhead, protocol efficiency (SMB vs NFS), CPU load, caching, file sizes, and even antivirus scanning. A range is more honest than a single number. The primary limiter label tells you where to focus upgrades.
Why isn't my 10GbE network giving me 10GbE speeds? +
10GbE provides a theoretical maximum of 1,250 MB/s, but your actual speed is limited by the slowest component. If your NAS uses spinning HDDs in RAID 5, the drives may only deliver 300-450 MB/s, making them the bottleneck, not the network. The estimator identifies this for you.
Does SSD cache help with transfer speeds? +
SSD cache primarily improves random read performance and small file access. For large sequential transfers (e.g., copying video files), SSD cache provides minimal benefit because the data is read sequentially from the HDD array. The estimator adjusts its range based on your workload type.
Why are small files so much slower to transfer? +
Each file transfer involves metadata operations, opening, writing, closing, verifying. For large files, this overhead is negligible compared to the data transfer. For thousands of small files, the overhead dominates. A folder of 100,000 × 10KB files transfers much slower than a single 1GB file, even though the total data is the same.
Should I upgrade my network or my NAS drives first? +
Use the estimator to find your primary limiter. If it says "Network," upgrade your network first. If it says "NAS storage," faster or more drives will help more. In many home setups with HDDs, the NAS storage is the bottleneck well before 10GbE, meaning a 2.5GbE upgrade may be sufficient.
Does RAID type affect transfer speed? +
Yes. RAID types that stripe data across drives (RAID 0, 5, 6, 10) can read faster than a single drive because multiple drives contribute data simultaneously. RAID 1 (mirror) can also read faster than a single drive. Generally, more drives in a striped array means higher aggregate read throughput, up to the network or controller limit.
How does Unraid affect transfer speed? +
Reads: Unraid does not stripe data, each disk is formatted independently (XFS or BTRFS). For a single sequential read, throughput is capped at one data drive's sequential read speed (typically 100-160 MB/s for a 7200 RPM HDD). The parity disk plays no role in reads. Multiple concurrent users or streams can read from different disks simultaneously, but a single transfer will never exceed one drive's speed. On 1 GbE (~115 MB/s), the network is usually the bottleneck before the drives, which is why many Unraid users see no benefit from upgrading to 2.5 GbE for sequential reads.

Writes: Every write to the array must update the parity disk, and Unraid supports two write modes with different speed and power tradeoffs. Read/Modify/Write (the default) reads the existing data block and existing parity, calculates the new parity, and then writes both, only the target data drive and the parity drive spin up. This is energy-efficient but slow: typically 20-40 MB/s. Turbo Write (Reconstruct Write) reads all data drives simultaneously to recalculate parity from scratch, then writes. All drives spin up, power consumption rises significantly, but speeds reach 40-120 MB/s. Neither mode achieves single-drive sequential write speed, you will always be slower than a drive's raw write throughput when writing to an Unraid parity array.
What about write speeds to NAS? +
This estimator focuses on read transfer speeds (copying from NAS to client), which is the most common scenario. For RAID 5/6, write speeds are typically 10-30% slower than reads due to parity calculation overhead. For RAID 1 and RAID 10, write speeds are similar to single-drive speeds. For Unraid, write speeds depend on write mode: Read/Modify/Write (the default) achieves 20-40 MB/s; Turbo Write (Reconstruct Write) achieves 40-120 MB/s. Neither Unraid write mode approaches single-drive sequential write speed, parity recalculation is always the limiting factor. The Unraid Write Speed panel above your results shows the estimated write range for your selected mode.
How accurate are these estimates? +
The ranges are based on published benchmarks and typical observed performance. Most users should see speeds within the estimated range. However, specific hardware combinations, firmware versions, protocol settings, and network congestion can push speeds outside the range.
Why is my remote NAS access so slow in Australia? +
Two constraints are specific to Australian residential internet. First, NBN upload speeds are capped at 20-50 Mbps regardless of download tier: NBN 100, 250, and 1000 all have 20-50 Mbps upload limits on FTTC/FTTN connections. Remote NAS access is upload-limited (your NAS uploads to you remotely), so a faster download plan doesn't help. Second, most AU residential ISPs use CGNAT (Carrier-Grade NAT), which blocks standard port forwarding. Synology QuickConnect bypasses CGNAT via relay servers, but relay-routed traffic is typically capped at 5-10 MB/s. For faster remote access, you need a static IP (ask your ISP, usually $10-20/month extra) or a WireGuard VPN tunnel on the NAS.