Network Usage Standard Data
Document PurposeProvides standard data on network traffic and system resource usage that occurs when introducing SHIELD Gate in the customer's internal network environment. This can be used for infrastructure capacity estimation, introduction review, comparison with other solutions, and other purposes.
1. Measurement Overview
1.1 Measurement Environment
| item | Content |
|---|---|
| Measurement Environment | Cloud Virtualization Development and Testing Environment (8 Core, 32GB RAM) |
| Number of concurrent users | 10 people |
| Measurement Site | Naver Map, Naver Sports, video site |
| Measurement Items | Network transmission and reception traffic, CPU usage, memory usage |
1.2 Measurement Scenario
Measured using three representative scenarios reflecting actual user behavior.
| scenario | Usage Behavior | Traffic Characteristics |
|---|---|---|
| Scenario 1: General Web Surfing | Exploring Portal and News-type Pages (Continuous Scroll, Page Navigation) | Static + Image Content, Scroll Refresh |
| Scenario 2: Map site | Map scrolling, zoom in/out, place search | Dynamic Content + Tile Image |
| Scenario 3: Video Playback | Video Streaming Playback | Continuous Streaming Traffic |
1.3 Precautions for Utilizing Measurement Results
⚠️ This data isStandard Measurement Environment Reference Valuesand, in actual customer environments, differences may occur due to the following factors.
- User-specific usage patterns (page stay time, scroll frequency, video viewing ratio, etc.)
- Types of Content on the Accessed Site (Static Page vs Dynamic Page)
- Video Playback Quality Settings (SD/HD/FHD/UHD, etc.)
- Screen Transmission Frame Rate Settings (12fps / 24fps / Auto)
- Network Environment (Bandwidth, Latency)
Each scenario value may change the relative ranking between scenarios if the above conditions change.
2. Network Usage by Scenario (10 Concurrent Connections)
SHIELD Gate(RBI) compresses and transmits the screen rendered by the server. Therefore, network usage is not based on the original size of the accessed site but ratherChange in Screen and Compression EfficiencyIt depends on the left and right. Continuous scrolling during general web surfing has the highest traffic as the entire screen is continuously refreshed, while video playback is measured relatively lower as changes are concentrated in the playback area.
Measurement result traffic size:General web surfing (75.18) > Video playback (50.05) > Map (32.70).
2.1 Scenario 1: General Web Surfing
| Measurement Items | Measurement value |
|---|---|
| Network Average | 75.18 Mbps |
| n Reception (RX) | 80.23 Mbps |
| n Transmission (TX) | 70.12 Mbps |
| Total Pod CPU | 0.99 cores |
| Total Node CPU | 3.09 cores |
💡 When scrolling continuously, the entire viewport is updated every frame, resulting in the largest amount of screen change. Conversely, during page stay (reading), there is almost no traffic. This figure is based on the section where scrolling is concentrated.
2.2 Scenario 2: Map Site (Naver Map)
| Measurement Items | Measurement value |
|---|---|
| Network Average | 32.70 Mbps |
| n Reception (RX) | 35.11 Mbps |
| n Transmission (TX) | 30.29 Mbps |
| Total Pod CPU | 1.22 cores |
| Total Node CPU | 4.15 cores |
💡 The map loads tile images when scrolling or zooming, but the average traffic is low. However, the interaction frequency is high, so CPU usage is higher than regular web browsing.
2.3 Scenario 3: Video Playback
| Measurement Items | Measurement value |
|---|---|
| Network Average | 50.05 Mbps |
| n Reception (RX) | 60.80 Mbps |
| n Transmission (TX) | 39.30 Mbps |
| Total Pod CPU | 2.12 cores |
| Total Node CPU | 6.03 cores |
💡 The video focuses on changes primarily in the playback area and has continuous movement, resulting in high compression efficiency, so the traffic is lower than regular web browsing. On the other hand, due to continuous screen updates, CPU usage is the highest among the three scenarios. There may be variations depending on the quality (SD/HD/FHD).
3. Per Capita Conversion and Dosage Calculation
3.1 Per Capita Usage by Scenario
The measurements of the top 10 peopleBased on 1 userIt is the value converted to __PH_0__.
| scenario | Network Average (Per Capita) | Receiving (RX) | Transmission (TX) | Pod CPU (per person) |
|---|---|---|---|---|
| General Web Browsing | about 7.5 Mbps | about 8.0 Mbps | about 7.0 Mbps | about 0.10 cores |
| Map Site | About 3.3 Mbps | about 3.5 Mbps | About 3.0 Mbps | about 0.12 cores |
| Video Playback | About 5.0 Mbps | About 6.1 Mbps | About 3.9 Mbps | about 0.21 cores |
3.2 Determination of Maximum Load Criteria (Conservative)
General web surfing (infinite scrolling) has the highest traffic among measurement scenarios,Conservative Standards for Infrastructure Capacity Estimationcan be utilized. (Approximately 7.5 Mbps per person)
| Number of simultaneous connections | Expected Network Average | Recommended bandwidth (30% margin) |
|---|---|---|
| 10 people | about 75 Mbps | About 98 Mbps |
| 50 people | about 375 Mbps | About 488 Mbps |
| 100 people | About 750 Mbps | about 975 Mbps |
| 200 people | about 1.5 Gbps | Approximately 1.95 Gbps |
| 500 people | About 3.75 Gbps | about 4.88 Gbps |
💡 Recommended Bandwidthis the value added with a 30% margin over the measurements to respond to peak traffic and network fluctuations.
3.3 General Work Mixed Standards Calculation (Recommended)
In actual office environments, web surfing is predominant and continuous scrolling occurs only in certain sections, so the following mixing ratio standards are recommended as realistic estimates.
| Usage Pattern | ratio | Average Traffic Per Person |
|---|---|---|
| General Web Browsing | 70% | 7.5 Mbps × 0.7 = 5.25 Mbps |
| Map/Search and Other Interactions | 20% | 3.3 Mbps × 0.2 = 0.66 Mbps |
| Video Playback | 10% | 5.0 Mbps × 0.1 = 0.50 Mbps |
| Mixed Average (Per Capita) | about 6.4 Mbps |
| Number of simultaneous connections | Mixed Pattern Expected Traffic |
|---|---|
| 10 people | about 64 Mbps |
| 50 people | About 320 Mbps |
| 100 people | about 640 Mbps |
| 200 people | about 1.28 Gbps |
| 500 people | About 3.2 Gbps |
💡 This mixing ratio is a typical assumption for general office environments, and adjustments to the ratio may be necessary based on the actual work characteristics of the client company.
4. CPU/Memory Resource Usage
4.1 System Resources by Scenario (10 Concurrent Connections)
| scenario | Total Pod CPU | Total Node CPU | Per Person Pod CPU |
|---|---|---|---|
| General Web Browsing | 0.99 cores | 3.09 cores | about 0.10 cores |
| Map Site | 1.22 cores | 4.15 cores | about 0.12 cores |
| Video Playback | 2.12 cores | 6.03 cores | about 0.21 cores |
4.2 Resource Impact of Frame Rate Settings
The CPU usage varies depending on the screen transmission frame rate. Below isNaver Map Standard, 2-person Automated TestThe result.
| Frame Settings | version | CPU Usage | Memory Usage | Note |
|---|---|---|---|---|
| 12fps | existing | 1.5 cores (40%) | 1,026 Mi (32%) | - |
| 12fps | Improvement | 1.5 cores (38%) | 958 Mi (32%) | Difference is minimal |
| 24fps | existing | 1.6~1.7 cores (44%) | 1,027 Mi (32%) | - |
| 24fps | Improvement | 1.7 cores (50%) | 956 Mi (32%) | CPU slight increase |
| Auto | existing | 1.7 cores (46%) | 1,026 Mi (32%) | - |
| Auto | Improvement | 1.8 cores (50%) | 1,018 Mi (32%) | CPU slight increase |
💡 The higher the frame rate, the better the perceived quality for users, but CPU usage also increases (this is normal). In a typical work environment, 12fps provides sufficient usability.
5. Frequently Asked Questions (FAQ)
**Q1. In what environment was this measurement obtained?**A. Measured based on 10 simultaneous connections in a cloud virtualization development and testing environment (8 Core, 32GB RAM). Actual figures may vary depending on the bandwidth and usage patterns of the customer's internal network.
**Q2. How is traffic calculated when there are N people?**A. Convert the average traffic per person × number of simultaneous users. For a typical office environment, use the mixed standard of Section 3.3 (approximately 6.4 Mbps/person), and for conservative estimates, use the maximum load standard of Section 3.2 (approximately 7.5 Mbps/person). It is recommended to add a 30% margin for peak response.
**Q3. How should we determine the criteria in environments where the video playback ratio is low?**A. In a typical office environment, the mixed criteria in Section 3.3 is realistic. For conservative estimates that need to capture traffic the most, please use the maximum load scenario of **general web surfing (continuous scrolling) criteria (Section 3.2)**.
**Q4. What is the reason for the difference between receiving (RX) and transmitting (TX) traffic?**A. In all scenarios, reception (RX) is measured to be greater than transmission (TX). This is largely due to the fact that in the RBI environment, the screen data rendered by the server is transmitted predominantly towards the user.