File Transfer
High-Speed File Transfer
Introducing INNORIX's high-speed file transfer, which connects high-speed networks to actual file transfer performance.
File Transfer
Make Your Files Move as Fast as Your Network.
Building a high-speed network does not automatically make file transfers run at the same speed. Even with 10 Gbps or 40 Gbps of bandwidth, actual transfer throughput varies depending on latency, RTT, packet loss, file profile, storage I/O, concurrent traffic, and the transfer method.
- Available Network
- Bandwidth, RTT, Loss — Storage, File Profile
- INNORIX High-Speed Transfer
- Actual File Throughput
INNORIX High-Speed File Transfer uses the available capacity of your existing network for actual file transfer through Parallel Transfer, Dynamic Concurrency, and Long-Distance Optimization, rather than relying on the characteristics of a single protocol.
File Transfer
Turn Network Speed into Actual File Throughput.
The most important number in high-speed file transfer is not simply the maximum bandwidth shown on the network interface. Even on the same 10 Gbps link, actual file transfer performance can differ significantly between a short path inside a data center and a WAN connection across continents because of RTT, packet loss, storage I/O, and competing traffic.
Performance therefore needs to be evaluated across the entire data path, from reading the file at the source, through the network, to writing it at the target, rather than by link speed alone. INNORIX manages usable transfer capacity by considering latency, file profile, source and target storage, and concurrent traffic together with available bandwidth.
Transfer Condition, Impact on Actual Throughput
- Source Storage
- Transfer Processing
- Network
- Target Storage
- Actual File Throughput
File Transfer
Use the Capacity of Your Existing Network for High-Speed Transfer.
There are multiple approaches to high-speed file transfer, including UDP-based acceleration, Parallel TCP, High-Speed TCP, and WAN Optimization. No single approach can be simplified as universally superior in every environment. The right transfer architecture depends on the actual network architecture, security policy, operating environment, and required throughput.
Rather than assuming UDP-based data transfer, INNORIX applies Parallel Transfer and Transfer Optimization on top of existing TCP-based network and security environments. Instead of requiring a separate transfer-only network, it focuses on turning the network capacity already available between connected sources and targets into actual file throughput.
Existing Enterprise Network
- Firewall, VPN, WAN, Cloud
- TCP Connectivity
- INNORIX Transfer
File Transfer
Expand Usable Bandwidth with Parallel Transfer and Concurrency.
On high-speed, long-distance networks, a single connection may not be able to use all available bandwidth effectively. INNORIX processes data through multiple parallel sessions and adjusts concurrency according to source storage, target storage, and network conditions.
The goal is not simply to increase the number of sessions. If source read performance or target write performance reaches its limit first, or if other workloads share the same network, excessive parallelism can reduce efficiency. The objective is to turn the capacity actually available along the current data path into file throughput.
Environment, Transfer Operation
- Source
- Session 1
- Session 2
- Session 3
- Session 4
- Target
File Transfer
Complete Transfers Across Long Distances and Changing Network Conditions.
As distance increases, for example between Seoul and New York, RTT increases, and even links with the same bandwidth can use their actual capacity differently through a single connection. Packet loss also varies across WAN and Internet environments, so maintaining throughput and completing the transfer under changing network conditions matters more than a momentary peak speed measured under ideal conditions.
INNORIX applies Parallel Transfer and Transfer Optimization designed for long-distance environments while managing Transfer State and Recovery together. When network conditions change during a long-distance transfer, progress is preserved, and if the transfer is interrupted, it can continue from the existing state through completion.
LOCAL
- Server A
- Low RTT
- Server B
LONG DISTANCE
- Seoul
- High RTT
- New York
Network Condition
- Stable
- Loss
- Recover
- Loss
- Complete
File Transfer
Treat Network and Storage as One End-to-End Performance Path.
A 10 Gbps network does not mean the source can read data or the target can write it at the same rate. Actual file transfer throughput is affected by the slowest part of the path across source read, transfer processing, network, and target write.
The file profile also changes performance along the same data path. A single large file depends heavily on sustained network and storage throughput, while many small files add File Discovery, Open/Close, Metadata, and Queue Processing to the performance equation. INNORIX operates transfers based on the actual throughput of the complete data path, including storage I/O, file size, and file count, not on a single network number.
Data Profile, Primary Performance Factors
- Source Storage — 600 MB/s
- Transfer Processing
- Network — 10 Gbps
- Target Storage — 450 MB/s
File Transfer
Operate Multiple Networks and Multiple Transfers Within Total Capacity.
When sources and targets have multiple network interfaces or available network paths, Multi-Network Transfer can use capacity across those paths for data movement. At the same time, an enterprise network is not dedicated exclusively to file transfer. Applications, databases, user traffic, and other services may share the same infrastructure.
INNORIX uses Parallel Transfer, Dynamic Concurrency, and Queue management to control not only the throughput of an individual transfer but also the total capacity consumed by multiple simultaneous transfers. The goal is to operate high-speed transfer alongside other workloads within the existing enterprise infrastructure rather than isolating it on a separate network island.
Operation, Role
- Source
- Network 1
- Network 2
- Network 3
- Target
- Applications, Database, User Traffic, Services
- Shared Capacity
- Queue, Concurrency, Transfer
File Transfer
Make Speed and Recovery Part of the Actual Completion Time.
If a high-speed transfer fails midway and the entire file must be sent again from the beginning, actual job completion time can increase dramatically even when peak throughput is high. In long-distance transfers involving hundreds of gigabytes or terabytes, Recovery becomes as important a performance factor as throughput itself.
INNORIX combines High-Speed Transfer with Resume, Retry, and Recovery so transfers can continue based on data that has already been delivered. Rather than treating speed and reliability as separate features, it operates around the total time required for the transfer to actually complete.
Large Transfer
- 0% – 82%
- Interrupted
- Resume
- 82% – 100%
File Transfer
Apply One High-Speed Capability Across Different Data Movements.
High-Speed File Transfer is not a separate accelerator limited to a specific point-to-point transfer. It is a common performance capability that can be applied across different data paths between servers, cloud environments, object storage, data centers, branches, and AI compute.
The data profile is not limited to a single large file either. The same performance capability can be used for Large File, High-Volume, Mixed Dataset, Cloud Migration, Distribution, and AI Data Delivery. When needed, transfers can also be automated through Schedule, File Event, API, and Flow rather than being limited to manual execution.
Area, Examples
- Source
- INNORIX
- Server
- Cloud
- Object Storage
- Data Center
- Branch
- AI Compute
File Transfer
Validate Transfer Performance Under Real Conditions.
A single number such as "up to 10 Gbps" is not enough to determine transfer performance in a real environment. Results can vary substantially depending on bandwidth, RTT, packet loss, file profile, storage, and concurrency conditions even when the same product is used.
High-Speed Transfer should therefore be evaluated not only by peak speed but also by how much of the available network capacity becomes actual file throughput under specific conditions. Improving low throughput on a 1 Gbps environment and making the most of already deployed capacity on a 10 Gbps or faster network are different in scale, but they address the same underlying problem.
Test Condition, What to Measure
Transfer Efficiency
File Transfer
See Transfer Performance and Completion Results on the Same Platform.
In enterprise environments, confirming that a transfer completed successfully is as important as achieving high speed. INNORIX lets operators view progress and throughput in Runs and Monitoring, and verify individual files and final transfer results through File Status and Receipt.
The same operating model applies when High-Speed Transfer is connected to Automation and Flow. Performance, Recovery, and Completion Result can be managed as one Transfer Operation whether the transfer starts from a Schedule, Event, API, or a previous Flow.
RUN-1842
File Transfer
Turn Network Speed into Actual File Throughput.
Consider bandwidth, RTT, packet loss, storage I/O, and file profile together, and use Parallel Transfer, Dynamic Concurrency, and Long-Distance Optimization to turn the available capacity of your existing network into actual data movement.
Apply the same High-Speed Transfer Capability across Large File, High-Volume, Server, Object Storage, and AI Data Delivery, and operate Recovery and final results on the same platform.
Every one of these capabilities,on a single platform.
The capability on this page is one of the many file transfer features INNORIX Platform provides. From file transfer to automation and unified operations, try every feature free.
Start FreeConsidering high-speed file transfer?
We'll help you review a high-speed file transfer setup that fits your current network and transfer environment.
- Review of long-distance and global transfer environments
- Performance configuration suited to your network
- Guidance on applying it to your existing infrastructure