Patent 10 / Protocol-Agnostic File Transfer Hub
01 / 11 US10701135B1
↑↓ navigate  ·  all patents →
Siten Sanghvi  ·  Granted Jun 30, 2020

Any protocol in.
Any protocol out.
AI picks the route.

An intelligent hub appliance that ingests data over any network protocol, uses machine learning to select the optimal destination server in real time, and converts the output to whatever protocol the destination requires — with the sender kept completely in the dark about where data went.

US10701135B1Patent
Jan 7, 2020Filed
6 monthsTime to grant
Multiple ClaimsScope
2 CitationsForward citations
SCROLL TO EXPLORE
Visual patent explainer
02 / The Problem

Every protocol needs its own gateway. Every route is static. No one optimizes.

Enterprise file transfer systems are built around specific protocols. Adding a new source or destination requires new gateway infrastructure. Routing decisions are pre-configured and never adapt — a server at 90% capacity receives the same traffic as a server at 20% capacity.

Protocol silosEach protocol stack requires separate integration — no unified entry point
Static routingTraffic distribution decisions made at setup time, never updated dynamically based on server state
Invisible failureA degraded destination server continues receiving traffic until an administrator manually reroutes
03 / The Invention

One hub. Any input. AI-optimized output.

The intelligent hub sits between all data senders and all destination servers. It detects the incoming protocol, profiles the sender and recipient, uses ML to select the best server in real time, converts data to the right output protocol, and deliberately hides the destination from the sender.

Transceiver — Speaks Every Protocol

Receives incoming data using any transmission protocol, splits it into addressable groups, and transmits those groups using a second protocol selected by the intelligence layer. One transceiver replaces an entire fleet of protocol-specific gateways.

Interrogator — Profiles Every Packet

Identifies the sender, intended recipient, and characteristics of the incoming data before any routing decision is made. This profile is the input to the machine learning model that selects the destination.

Intelligence Engine — Optimizes in Real Time

Ranks available destination servers against multiple criteria — capacity, bandwidth, latency, security profile — and routes each data group to the current best server. Rankings are continuously updated from historical performance data.
04 / Architecture

One entry point.
Infinite destinations.

Every sender connects to the same hub regardless of what protocol it uses. The hub handles all classification, routing, and translation internally. From the sender's perspective, data disappears into the hub. The sender never learns what happens next.

System diagram — US10701135B1
FTP Sender
HTTP Sender
SFTP Sender
↓ any protocol
Intelligent Hub
Interrogate → Route → Translate → Send
↓ optimal protocol
Server A
Score: 94
Server B
Score: 71
Server C
Score: 58
05 / Protocol Conversion

FTP in.
HTTPS out.
Zero reconfiguration.

The data translation processor converts incoming data from its original protocol to whatever protocol the selected destination server expects. No changes needed at the sender or receiver — the hub handles all translation invisibly.

The sender receives no information about what protocol was used for delivery or which server received the data — network topology remains opaque to the outside.

Protocol translation — US10701135B1
06 / AI Server Selection

The best server
for every packet.
Every time.

The intelligence engine continuously scores all available destination servers against a weighted model combining available capacity, network bandwidth, response latency, and security classification. The data goes to the current leader — not the historically cheapest or the default.

As each server's real-time state changes, scores shift. Traffic follows automatically, with no manual intervention required.

Live server ranking — US10701135B1
07 / Network Obfuscation

Senders see the hub.
Nothing else.

By design, the hub returns no information to senders about the destination server, the output protocol, or the routing decision. This "network obfuscation" is an explicit feature — it protects the destination topology from reconnaissance by compromised or malicious senders.

Even if a sender is breached, attackers cannot infer destination IP ranges, server names, or protocol details from the data transfer alone.

Information visibility model
FieldSender seesHub knows
Ingress (incoming transfer)
Source protocol FTP FTP
Sender identity Own ID Sender profile
Egress (delivery)
Destination server Server Alpha (ranked #1)
Output protocol HTTPS/TLS 1.3
Routing reason Score 94 — capacity + latency
08 / Data Splitting & Delivery

One stream.
Many destinations.
Parallel delivery.

The transceiver splits incoming data into groups and can route different groups to different destination servers simultaneously. This enables load distribution across a server farm even from a single incoming transfer — not just routing, but parallelizing delivery.

The intelligence engine determines the split — how many groups, which servers receive each, and which output protocol each delivery uses — all optimized for current network conditions.

Parallel delivery model
One incoming transfer
Split by intelligence engine into N groups
Group 1
→ Server A
Group 2
→ Server B
Group 3
→ Server C
Each group → different protocol · different destination · simultaneously
09 / Applications

Wherever data
crosses a protocol
boundary.

The intelligent hub adds value wherever enterprises have multiple data sources speaking different protocols, or multiple destinations that need different protocols — without requiring changes to either end.

Use cases — US10701135B1
Express
Unified File Transfer Gateway Replace per-protocol gateways (FTP server, SFTP server, HTTP endpoint) with a single hub that accepts all three and routes intelligently to backend systems.
Express
Real-Time Adaptive Load Balancing Unlike traditional round-robin, the hub continuously re-ranks servers and reroutes traffic when a server degrades — without any configuration change or restart.
Inferred
Secure Data Ingestion Pipeline Network obfuscation prevents senders from mapping the destination infrastructure — useful for financial institutions ingesting data from external counterparties without exposing internal topology.
Inferred
Multi-Cloud Data Distribution Route the same incoming data to different cloud storage providers in parallel, converting to each provider's preferred protocol — S3, Azure Blob, or GCS — from a single hub.
10 / Citations

Cited by 2 patents across
two sectors

US10701135B1 has been cited by teams at Micron Technology and Shanghai Jiao Tong University — spanning semiconductor memory systems and industrial protocol research. Citation data verified via Google Patents, June 2026.

Forward citations confirm the hub architecture resonates well beyond the original filing's direct use case — any system that handles multi-protocol data routing is in scope.
Forward citations — US10701135B1
US11088980B1 Aug 10, 2021
Single message management platform
Micron Technology, Inc.
CN119155358A Dec 17, 2024
Low-time-delay high-certainty industrial protocol conversion architecture and algorithm
上海交通大学 · Shanghai Jiao Tong University
11 / Timeline

Patent Lifecycle

Jan 7, 2020
Filed
Application US16/737,058 filed
6 months
Jun 30, 2020
Granted
US10701135B1 granted — published with grant (B1). One of the fastest grants in this portfolio.
~19.5 years
Jan 7, 2040
Expires
Est. expiration (subject to maintenance fees)
End / Patent 10