A PoE surveillance project needs more than enough camera ports. Each camera contributes a continuous video stream, while switches, uplinks, recording systems, viewing clients, and ordinary business traffic share the same network. If the design is based only on port count, a system can appear complete yet suffer from congestion, delayed live view, dropped packets, or difficult troubleshooting.
This guide presents a practical planning method for PoE IP camera bandwidth and network capacity. It uses confirmed information from SWGJVISION’s product and solution pages, but it does not treat any unconfirmed bitrate as a product specification. The numerical values in the formulas and worked example are clearly labeled planning assumptions. For a quotation or deployment, use the bitrate, frame rate, stream profile, audio setting, and actual operating mode stated in the selected camera’s current datasheet.
Estimate the aggregate camera load by adding the expected bitrate of every active stream, convert the total to a common unit, then reserve headroom for protocol overhead, bursts, viewing, recording, maintenance, and normal business traffic. Check this result separately against each access-switch port, switch backplane, uplink, recorder interface, and VMS server interface. A useful planning expression is:
Required capacity = (sum of active stream bitrates × overhead factor × growth factor) + shared network traffic
This is a planning method, not a SWGJ performance guarantee. PoE solves the power-and-data cabling problem; it does not automatically make every uplink large enough. SWGJ’s PoE IP Camera page confirms that its PoE cameras transmit power and data over one Ethernet cable, cite IEEE 802.3af/at compliance, and state operation up to 100 meters from the switch. Those facts help define the physical design, while the actual bandwidth calculation must come from the selected configuration.
Before calculating, establish what the supplier actually describes. SWGJ’s PoE IP Camera page identifies one-cable power and data, HD monitoring, 4K options, night vision, motion detection, H.265 compression, weatherproof construction, and remote live viewing. It also states that the PoE models can work up to 100 meters from the switch and reference IEEE 802.3af/at standards. These statements are useful planning inputs, but “4K” is an image-resolution label, not a fixed bitrate.
The broader IP Camera range is described as supporting Ethernet and WiFi access, remote viewing, NVR integration, local or cloud recording, and ONVIF interoperability. For a wired PoE design, use the specific Ethernet model and its stream settings rather than applying one average to every camera. Different resolutions, scene movement, compression settings, frame rates, and audio choices can produce different traffic patterns.
SWGJ’s IP Network Camera solution positions PoE CCTV for commercial, industrial, and residential projects, including factories, warehouses, offices, schools, stores, parking lots, and public infrastructure. Its Applications page lists oil wells, piers, navigation, chemical industry, large stadiums, construction sites, storehouses, and parking lots. These use cases can have different scene complexity, but neither page supplies a universal bitrate.
SWGJ describes Embedded NVR products as supporting multi-channel recording, H.265 compression, remote access, and IP camera integration. Its Video Management System page describes real-time viewing, recording, playback, and device management. Include the selected recorder or VMS host as a network endpoint, without treating channel count as a throughput guarantee.
A reliable bandwidth plan starts with an inventory that connects each camera stream to the physical network path carrying it. This makes it easier to separate confirmed device settings from provisional design assumptions and to identify the busiest switch, uplink, recorder interface, or VMS host before procurement. The inventory should remain part of the project record and be updated when camera profiles or viewing requirements change.
Create one row for every camera or camera profile. Record resolution, frame rate, codec, target or maximum bitrate mode, keyframe interval if applicable, audio status, substream status, and whether the stream is continuously recorded or viewed only on demand. Also record the physical switch, switch port, uplink path, recording endpoint, and expected number of concurrent viewers. This inventory prevents a high-resolution camera from being hidden inside a low-resolution average.
Use the camera’s configured bitrate whenever the device exposes one. If the camera uses variable bitrate, record both the expected average and the maximum or observed peak. For scenes with moving vehicles, crowds, foliage, rain, or changing illumination, the peak is especially important. If no reliable value is available yet, mark the number as an assumption and replace it during commissioning with a controlled observation.
A primary stream used for recording is not necessarily the only traffic generated by a camera. A lower-resolution substream may serve a browser wall or mobile preview, while a primary stream may be requested for investigation. The same source can therefore create more than one flow at a time. Count each simultaneous flow at the point where it crosses the relevant uplink.
Avoid counting the same traffic twice. A camera-to-access-switch flow consumes the camera port and the switch’s internal forwarding capacity. When it leaves that switch toward a recorder, it consumes the uplink. At a recorder-facing link, count all streams arriving there, plus any outbound live-view or playback traffic that uses the same interface. Draw arrows for each traffic direction so the busiest link is visible before hardware is ordered.
PoE planning has two independent constraints: data capacity and electrical power. The SWGJ page confirms one-cable delivery and 802.3af/at compliance, but a port can be data-capable while the switch’s total PoE budget is insufficient for all attached devices. Check the camera’s required class or wattage against the switch budget, including startup and night-illumination behavior where applicable. This article focuses on bandwidth; the power check should be documented alongside it rather than substituted for it.
Once the camera inventory is complete, convert the listed stream values into a common unit and calculate the load at each shared network point. The purpose is not to create a universal bitrate claim, but to make the project assumptions visible, repeatable, and easy to replace with measured values during commissioning. Keep camera traffic, viewing traffic, and ordinary shared traffic separate until the final capacity check.
Bitrate is commonly shown in Mbps, while interface capacity may be listed in Gbps. Convert before adding. For a simple planning method, sum the active stream rates:
Camera load (Mbps) = stream 1 + stream 2 + ... + stream n
Then apply allowances:
Designed load = camera load × (1 + protocol allowance) × (1 + growth allowance)
The allowances are engineering choices, not SWGJ specifications. A project team may choose different values after reviewing packet captures, service-level objectives, and the behavior of its switches and recording software. Keep the assumptions visible in the design document so another engineer can reproduce the result.
The following is an illustrative planning method, not a claim about SWGJ camera bitrate. Assume a small warehouse has 12 PoE cameras. Ten cameras are assigned a planning bitrate of 4 Mbps and two cameras are assigned 8 Mbps because their scenes are expected to be more detailed. Assume one primary stream per camera, a 15% allowance for protocol and traffic variation, and a 25% allowance for future expansion and operational bursts.
| Item | Planning assumption | Calculation | Result |
|---|---|---|---|
| Ten standard streams | 10 × 4 Mbps | 40 Mbps | 40 Mbps |
| Two higher-load streams | 2 × 8 Mbps | 16 Mbps | 16 Mbps |
| Base camera load | — | 40 + 16 | 56 Mbps |
| Protocol/variation allowance | 15% | 56 × 1.15 | 64.4 Mbps |
| Growth and burst allowance | 25% | 64.4 × 1.25 | 80.5 Mbps |
The resulting 80.5 Mbps is the design figure for this example’s camera flows. It is not a product rating and should not be copied into a bill of materials without validating the real camera profiles. If two operators simultaneously request additional streams, or if playback crosses the same uplink, add those flows to the calculation rather than quietly consuming the reserve.
A second useful check is conversion to storage-side throughput. Divide Mbps by eight to estimate MB/s for a rough planning view: 80.5 Mbps ÷ 8 is approximately 10.1 MB/s. This conversion is only a bandwidth-to-byte-rate planning method; recording retention also depends on motion, schedule, codec behavior, container overhead, and the recorder’s implementation. Do not infer storage capacity from this article alone.
The aggregate figure is only useful when it is checked against the actual path taken by the streams. A camera port, access-switch fabric, uplink, aggregation link, recorder interface, and VMS host can have different constraints. Review each layer separately, because a system may pass one check while failing at a shared uplink or recording endpoint.
Each access port must support the camera’s negotiated link and traffic profile, while the switch must forward all attached streams without exhausting its switching capacity. For a switch with several cameras, sum the flows that leave through the uplink. A port count can be adequate while the uplink is undersized, so evaluate both independently. Confirm whether the switch supports the required PoE standard and total power budget as a separate test.
The uplink is often the first shared bottleneck. Use the busiest expected interval, not merely the average across a full day. Include primary streams, substreams requested by live-view clients, recorder traffic, playback, management, and other applications that share the path. Apply the chosen headroom policy after the flows are summed. If a link is near saturation during a burst, latency and packet loss can affect many cameras at once.
For longer runs, the SWGJ PoE page’s 100-meter statement is a useful physical reference, but it does not approve every cable route or guarantee a particular throughput under site conditions. Verify cable category, termination quality, electromagnetic environment, intermediate equipment, and the actual switch interface. Where a project needs a longer path, treat the transport design as a separate engineering decision.
The selected Embedded NVR or VMS host should be checked for incoming throughput, simultaneous recording behavior, live-view output, playback, and interface speed. Channel count is a device organization feature; it is not by itself a bandwidth calculation. If a recorder accepts streams from several access switches, add those streams at the recorder-facing interface and leave room for expected live viewing and maintenance.
Similarly, the Video Management System description confirms functions such as real-time viewing, recording, playback, and device management, but it does not publish one universal throughput for every deployment. Obtain the host sizing guidance for the chosen software release and test the intended number of streams, users, and recording profiles in a pilot.
Start with a design worksheet that names each camera, stream, switch, uplink, recorder interface, and assumption. Mark every number as confirmed, measured, or provisional. This simple status field prevents a planning estimate from being mistaken for a vendor specification during procurement.
Next, validate the camera profiles on a representative sample: a quiet scene, a high-motion scene, day and night conditions, and any scene with lighting changes. Measure average and peak traffic at the switch and at the recorder-facing link. Compare observations to the worksheet, then update the allowance if the measured behavior differs materially from the assumption.
During acceptance testing, review interface utilization, errors, discards, packet loss, latency, link negotiation, PoE events, and recorder health. Test the busy case: concurrent live viewing, a playback request, and ordinary business traffic. Do not use a single successful camera preview as proof that the whole path has capacity.
Finally, document an expansion trigger. For example, when the designed load reaches the team’s chosen utilization threshold, require a new uplink review before adding cameras. Keep spare port and PoE budget, but do not count unused ports as bandwidth reserve. Recalculate whenever resolution, frame rate, codec, recording schedule, audio, viewing behavior, or camera count changes.
For project-specific confirmation, compare the worksheet with SWGJ’s Products, review the relevant Applications, and request current datasheets or test guidance from the manufacturer before final deployment.
The following questions address common planning mistakes when designing a PoE IP camera network. The answers describe a method for checking the selected configuration; they do not replace the current camera, switch, NVR, or VMS documentation.
There is no single universal value. The required rate depends on resolution, frame rate, codec, scene complexity, bitrate mode, audio, and simultaneous streams. Use the configured or measured bitrate for the exact camera profile. Any number used before that confirmation must be labeled as a planning assumption.
No. PoE combines electrical power and data on one Ethernet cable; it does not create extra data capacity. Check the camera port, switch forwarding capacity, uplink, recorder interface, and any shared path independently.
Yes, whenever they can be active at the same time. Count every stream at the link where it travels. A substream used for previews may be small, but multiple viewers can still create meaningful aggregate traffic.
No. SWGJ’s PoE page states operation up to 100 meters from the switch, which is a useful product planning reference. The final result also depends on cable quality, termination, environmental conditions, intermediate equipment, and negotiated link performance. Validate the installed path.
Request the current camera datasheet and configured stream limits, codec and frame-rate options, PoE class or power requirement, interface details, and any recorder or VMS throughput guidance. Then compare those facts with measured pilot traffic and the site’s growth assumptions.
This article was prepared after reviewing SWGJVISION pages for PoE IP Camera, IP Camera, Embedded NVR, Video Management System, Solution, and Applications. Product capabilities and distances are reported only where those pages explicitly described them; all numerical bandwidth examples are planning methods.
Related Articles: