How to Choose IP Camera Bandwidth for Network Planning

Why Bitrate Drives Your Network Plan

Network planning for a video surveillance system starts with the data stream. The resolution you select on the camera menu is only one part of the equation. The actual bandwidth a camera consumes depends on codec, bitrate, frame rate, and scene complexity. A 4-megapixel H.265 camera in a low-activity area uses far less bandwidth than a 2-megapixel H.264 camera watching a busy warehouse. The encoder compresses the image based on motion. Still, a camera pointed at a static wall at 30 frames per second may use only a fraction of the bandwidth of a camera tracking a forklift in a loading dock.

Most installers make the mistake of calculating bandwidth based on resolution alone. That leads to undersized switches and slow NVR writes. You need the real numbers from the camera specifications. Resolution defines the spatial detail, but the bitrate defines the data volume. A high-resolution image with low motion compresses efficiently. A low-resolution image with heavy motion or complex backgrounds can generate a higher bit rate.

What You Need Before You Calculate

Before you run the math, gather these items from the product data sheets and your site survey:

1. Camera model and firmware version.
2. Codec selection: H.264, H.265, or H.266.
3. Maximum and typical bitrate for each camera.
4. Frame rate setting, usually 15, 25, or 30 frames per second.
5. Number of cameras per switch and per NVR.
6. Planned storage duration and number of days.
7. Whether the system uses fixed bitrate or variable bitrate.

Variable bitrate is standard for most surveillance cameras. The encoder adjusts the bit rate based on motion. A static office wall uses very little data. A parking lot with moving cars uses more. Your calculation should use the maximum sustained bitrate, not the average, because network saturation matters most during high-motion periods. Check the data sheet for the “maximum” or “peak” figure. Many cameras list a lower “typical” number that reflects a quiet environment. If you plan a system for a retail store or a highway interchange, the peak figure is the correct one to use for switch sizing.

Also verify the codec support on your NVR and VMS. While H.265 is common in modern cameras, older NVRs may only support H.264. If the NVR cannot decode the stream, you must either downgrade the camera or upgrade the recorder. This compatibility check prevents a scenario where the network is correctly sized but the storage side fails to record the stream correctly.

How to Calculate Total Network Bandwidth

The base formula is simple:

Total Bandwidth = Number of Cameras × Average Bitrate

Use the bitrate in megabits per second. If the data sheet lists kilobits, divide by 1000 to convert.

Example:

• 20 cameras
• 4 Mbps average bitrate per camera
• 20 × 4 = 80 Mbps total camera traffic

This figure is only the camera stream. You must add overhead for the Video Management Server, user logins, and control traffic. Add 10% to 15% for VMS traffic. In the example above, 80 Mbps plus 12 Mbps equals 92 Mbps. VMS traffic includes RTSP discovery, ONVIF commands, and the continuous stream of metadata used for search functions. While individual user logins consume little data, the constant background communication between cameras and the VMS adds up over a large install.

Next, apply network headroom. A network running at 100% capacity will drop packets. Plan for 20% to 30% spare capacity. In the example, 92 Mbps plus 25% equals roughly 115 Mbps. This is the minimum switch capacity you need for the camera segment. If you use a 1 Gbps switch, 115 Mbps leaves plenty of room for other devices. If you use a 100 Mbps switch, the system will fail immediately.

Consider the switch architecture. If all 20 cameras are on a single switch, that switch needs to handle the total load. If you split the cameras across two 8-port switches, each switch handles only 10 cameras, or roughly 40 to 50 Mbps of camera traffic. This approach also simplifies troubleshooting. If one switch fails, only half the cameras are affected. It also allows you to match the switch power budget to the camera power draw more precisely.

How Storage Affects Your Bandwidth Plan

Network bandwidth and storage throughput are different problems, but they share the same NVR or VMS hardware. A system can have a perfect network and still fail if the NVR cannot write all the streams simultaneously. The network carries the data to the NVR, but the NVR must process and store it on disk.

Check the NVR write throughput. It is measured in gigabits per second. For 20 cameras at 4 Mbps, the total write load is 80 Mbps. A single NVR with a 10 Gbps write port can handle this easily. A 1 Gbps port will struggle if all streams peak at the same time. The write speed of the internal hard drives also matters. If the NVR uses multiple hard drives in a RAID configuration, the effective write speed depends on the drive type and the RAID level. A 1 Gbps network port is useless if the drives can only write 500 Mbps.

Storage capacity is a separate calculation. Multiply the total bitrate by 86,400 seconds (one day) and the number of days of storage. Convert from megabits to bytes for the drive size.

Example:

• 20 cameras × 4 Mbps = 80 Mbps
• 80 Mbps × 86,400 seconds = 6,912,000 Mbit per day
• 6,912,000 / 8 = 864,000 MB per day
• 864,000 MB × 30 days = 25,920,000 MB
• 25,920,000 / 1,024 / 1,024 ≈ 25.3 TB

This is the raw storage needed. Add 15% to 20% for filesystem overhead and indexing. You need roughly 30 TB of usable storage for that example. Filesystem overhead includes the space required for the file structure itself. Indexing overhead is the space used to build the metadata that allows you to search for motion events or specific timestamps. If you use motion-based recording instead of constant recording, the storage requirement drops, but the calculation becomes more complex because the peak bandwidth is harder to predict.

Common Mistakes in Bandwidth Planning

Installers and buyers make these errors regularly.

1. Using marketing resolution instead of actual bitrate. A 4K camera does not automatically mean a 4K bandwidth requirement. The bitrate setting matters more. A 4K camera set to a low bitrate for long-term storage may use less bandwidth than a 2K camera set to a high bitrate for detailed inspection.
2. Ignoring PoE limits. A switch port that delivers 25 watts may limit the number of high-power cameras it can support. Check the power budget separately from the data budget. A PTZ camera with built-in heaters and IR illuminators can draw 25 watts or more. If a switch has a 32-port configuration and each port is limited to 8 watts, the total power budget is 256 watts. If you connect 20 high-power cameras, you exceed that budget. The switch will shut down ports or cameras to protect itself.
3. Forgetting VMS overhead. The VMS is not passive. It processes metadata, thumbnails, and user requests. This traffic adds up. In a large system with 500 cameras, the VMS traffic can be significant. It can consume several percent of the network bandwidth even when no one is watching the screens.
4. Assuming one NVR handles everything. If the NVR write port is saturated, the network may look fine but the NVR will drop frames. This is a common issue in retrofit installations where the existing NVR was sized for a smaller system. Adding more cameras without upgrading the NVR leads to data loss.
5. Not testing with real footage. A static test image shows low bandwidth. A real site with motion shows the true load. Before commissioning, run the cameras for at least 24 hours with normal activity. This reveals the peak bitrates and helps verify the storage calculations.

Verification Before Commissioning

Before you cut the ribbon on the installation, run a verification test.

1. Record all cameras for 24 hours.
2. Check the NVR storage usage. Compare it to your calculated size.
3. Monitor the switch port utilization. Use the switch management interface or a network analyzer.
4. Check the NVR write throughput. Ensure it is below 80% of its maximum.
5. Test the VMS playback. Open multiple streams at once and check for lag.

If the network port exceeds 70% utilization during peak hours, increase your headroom or move cameras to a different switch. If the NVR write throughput hits 100%, the system will drop frames even if the network is healthy. Look for dropped frames in the VMS. If you see gaps in the timeline or stuttering playback, the NVR is overloaded.

Also check the power stability of the PoE switches. If a switch is overloaded, it may drop cameras intermittently. This is a subtle sign of a power budget issue. Monitor the switch logs for port disconnects. If you see repeated disconnects on specific cameras, check the power draw of those cameras against the switch port limit.

Quick Reference Table

The table below shows typical bandwidth ranges for common camera types. These are general guidelines. Always check the specific camera data sheet for exact numbers.

Camera Type	Codec	Typical Bitrate Range	20 Camera Total	30 Day Storage Estimate
2MP	H.264	2 to 4 Mbps	40 to 80 Mbps	12 to 24 TB
4MP	H.265	3 to 6 Mbps	60 to 120 Mbps	18 to 36 TB
8MP	H.265	6 to 10 Mbps	120 to 200 Mbps	36 to 60 TB
PTZ	H.265	4 to 8 Mbps	80 to 160 Mbps	24 to 48 TB

Note: Storage estimates assume constant recording. Motion-based recording can reduce storage but makes bandwidth planning harder because peaks are harder to predict.

Final Check

Your bandwidth plan is complete when you have three numbers:

1. Total camera bandwidth in Mbps.
2. Total VMS and control overhead in Mbps.
3. Total switch and NVR throughput in Gbps.

If your switch is 1 Gbps and your total camera traffic is 400 Mbps, you have headroom. If your switch is 1 Gbps and your total camera traffic is 950 Mbps, you will see dropped frames.

Use the actual camera settings, not the defaults, when you write the final specification. A camera set to 30 fps at maximum bitrate uses significantly more bandwidth than one set to 15 fps at minimum. The choice is yours, but the math must match the setting. Document the specific bitrate and frame rate for each camera in your project files. This creates a clear record of what was installed and what was expected. It also makes it easier to troubleshoot issues later. If a camera starts using more bandwidth than expected, you can compare the current setting against the documented baseline.