INNORIX | 30% Packet Loss

1. An Uncontrollable Communication Wasteland: A Desperate Network Emulation Where 30 Out of 100 Packets Disappear

A harsh data environment with 30% packet loss, where conventional transfer protocols effectively cease to function

Item	Config Value	Meaning	Network Impact
Packet Loss Rate	30%	3 out of 10 lost	Communication becomes impractical
Loss Pattern	Random / Burst mixed	Continuous + intermittent loss	Maximized recovery difficulty
RTT (Latency)	100 ~ 500ms	Long-distance / satellite assumption	Increased retransmission delay
Jitter	50 ~ 200ms	Response variability	Unstable transfer speed
Bandwidth Limit	10Mbps ~ 100Mbps	Limited link environment	Performance Bottleneck
Packet Reordering	Partial occurrence	Out-of-order packets	Increased receiver processing load
Test Method	tc qdisc / network emulator	Forced environment control	Reproducible experiment
Expected Result	TCP performance collapse	Congestion control malfunction	Speed converges to 0

Accumulated latency from retransmission requests drives transfer efficiency toward zero

Stage	Network State	TCP Behavior	Internal Mechanism Change	Result
1. Normal Transfer	Initial data send	Window expansion (Slow Start)	Gradual throughput increase	Speed rises
2. Packet Loss Occurs	Partial loss	Duplicate ACK received	Loss detection starts	Unstable transfer
3. Retransmission Trigger	Missing ACK / Timeout	Retransmit packets	RTT increases	Delay accumulates
4. Congestion Control	Repeated loss	Window size sharply reduced	Forced throughput decrease	Speed drops sharply
5. RTT Increase	Waiting time grows	Timeout expands	Waiting accumulates	Transfer stalls
6. Additional Loss	Continuous 30% loss	Repeated retransmission	Vicious cycle forms	Efficiency collapses
7. Window Collapse	Shrinks to minimum	Throughput limited	Slow recovery	Near unusable
8. Final State	Persistent loss	Repeated attempts only	Minimal actual delivery	Speed → 0

A proprietary INNORIX algorithm that selectively recovers only missing data without exhausting full bandwidth

Stage	Network Condition	INNORIX Engine Behavior	Processing Method	Result
1. Data Transfer	30% packet loss	Maintains full streaming	No interruption	Continuous transfer
2. Loss Detection	Missing blocks	Real-time tracking	Block-level identification	Precise detection
3. State Recording	Received vs missing	Transfer map generated	Bit/block management	Full visibility
4. Recovery Target Selection	Missing data exists	Select only lost blocks	Selective request	Eliminates redundancy
5. Retransmission	Only required parts	Parallel recovery	Distributed usage	Fast reconstruction
6. Flow Maintenance	Concurrent transfer	Continuous streaming	No interruption	Stable speed
7. Repeated Loss Handling	Ongoing loss	Same logic reapplied	Auto correction	Stable transfer
8. Final Convergence	All blocks received	Missing = 0	Fully reconstructed	100% recovery

Bandwidth optimization technology that maintains constant speed regardless of RTT

Item	Legacy TCP Transfer	Problem	INNORIX Method	Result
RTT Dependency	Directly proportional	Speed drops with distance	RTT-independent	Constant speed
Round-trip Structure	ACK-based required	Waiting accumulates	Asynchronous streaming	No wait
Transfer Window	Limited by RTT	Throughput constrained	Parallel streams	Throughput maintained
Loss + Delay	Retransmission delay grows	Performance collapse	Independent handling	Minimal impact
Bandwidth Usage	Drops with RTT	Inefficiency	Continuous utilization	Maximized
Speed Variability	Highly volatile	Unstable transfer	Stable rate	Stability ensured
Long-distance	Degrades over WAN/satellite	Not practical	Same performance	Global ready
Final Result	“Farther = slower”	Structural limitation	“Distance irrelevant”	Consistent transfer

A flexible engine that dynamically adjusts throughput in real time based on changing network conditions

Situation	Network Change	Legacy Response	Problem	INNORIX Behavior	Result
Bandwidth Drop	Sudden congestion	Speed collapse	Unstable session	Auto rate reduction	Connection maintained
Bandwidth Recovery	Congestion cleared	Slow recovery	Delay persists	Immediate expansion	Performance restored
Increased Loss	Higher loss rate	Retransmission spike	Speed collapse	Selective recovery	Flow maintained
RTT Increase	Latency rise	ACK delay	Transfer stalls	Async streaming	Minimal impact
Jitter	Irregular delay	Speed fluctuation	Instability	Buffering + tuning	Stability maintained
Short Disconnection	Temporary drop	Transfer failure	Restart needed	Auto reconnect + resume	Stability maintained
Repeated Variability	Constant change	Continuous instability	Requires management	Real-time tuning	Unmanned operation
Final State	Unpredictable	Unstable / failure	Operational risk	Seamless streaming	Stable completion

Overcoming poor network conditions through software without requiring costly infrastructure

Item	Legacy Method (Infrastructure Dependent)	Limitation	INNORIX Method (Software-Driven)	Result
Network Requirement	Low loss / low latency required	Environment constrained	Handles high loss/latency	Works anywhere
Circuit	MPLS / dedicated lines	High cost	Public internet	Cost reduction
Performance	Depends on infrastructure	Limited efficiency	Engine compensates	Stable performance
Loss Handling	Depends on network quality	Not fundamentally solved	Selective recovery	Guaranteed quality
Latency Handling	Affected by distance	WAN degradation	RTT-independent	Global consistency
Scalability	Requires line expansion	Cost/time increase	Software scaling	Instant expansion
Operational Cost	Continuous circuit cost	OPEX increase	Minimal infra	Cost optimized
Deployment	Complex network design	Requires experts	Plug-and-play	Fast adoption
Final Result	“Only works on good networks”	Limited	“Works on any network”	Infrastructure independent

Final proof that even under extreme loss, the delivered data matches the original at the bit level

Verification Item	Legacy Result	Limitation	INNORIX Method	Final Result
Packet Loss Impact	Possible data loss	Incomplete recovery	Full block recovery	100% received
File Count	Some failures	Hard to detect	Full comparison	Exact match
Total Data Volume	Approximate	Cannot detect corruption	Byte-level comparison	Identical
Data Content	Possible corruption	Hidden errors	Bit-level verification	Zero-byte error
Checksum	Partial validation	Low reliability	Full hash comparison	Perfect match
In-transfer Validation	Limited	Error accumulation	Real-time validation	Immediate correction
Failure Accumulation	Quality degradation	Reliability collapse	Independent of failures	Same result
Final Verdict	“Transfer complete” (uncertain)	Uncertainty	“Identical to original”	100% integrity confirmed