Axis bet that security belongs on the network.

Light hits the sensor. WDR and shutter settings tame the scene. The encoder compresses it, and Zipstream keeps the quality cheap. It travels over Ethernet, where PoE carries power on the same cable. A recorder stores it and an operator watches or reviews. The modern twist is edge analytics: the camera decides what's worth attention.

Close with: “A camera that understands what it's seeing, and only bothers you with what matters.”

Walk the five inputs per zone. Mix P series for coverage, Q-series PTZ where zoom and endurance matter, M series for dense indoor counts. Run analytics on the camera with ACAP so the recorder stays lean. Size the switches for Class 4 or PoE++ on the PTZs, and get bitrate and storage from Site Designer. Hand back a per-zone breakdown, not one number. Full version under system design.

It's a common language for security devices: discovery, streaming and PTZ (Profile S), advanced codecs and events (T), analytics metadata (M). The customer isn't locked to Axis's recorder; they can choose the VMS that fits. Axis co-founded it in 2008. Without it, every camera brand would be an island.

Start with their problem, not the camera. Then show three things: a backlit door where WDR keeps both sides visible; a person walking through while the face stays usable; analytics flagging a person or car with no settings to touch. Stop there. Then ask: “What would this have caught for you last month?” Credibility comes from the setup, not the gadgets.

Move from unit price to total cost and risk. Edge analytics mean fewer servers and licences. IK10 housings and LTS firmware mean fewer truck rolls over a 7–10 year life. Zipstream cuts the storage bill every year. And for government or enterprise, a system that fails an audit costs far more than the sticker gap. Then ask what the footage has to prove, and let that drive the design.

Acknowledge it: they're cheaper and ship fast. Then the three reasons: NDAA Section 889 and the FCC Covered List for anything federally funded; firmware provenance (own chip, signed firmware, Edge Vault); and an open platform that works with their VMS. Be honest: if they only want the cheapest boxes, Axis isn't the answer. For anything that has to survive an audit, it is.

Almost certainly through ONVIF, and with the major platforms through deeper integrations that use VAPIX for Axis-specific features. Ask which VMS and version, then check how deep the integration goes before promising a specific analytic. See the VMS table.

One cable for power and data. Installs are faster and usually need no electrician at each camera, and a UPS at the switch keeps every camera up in a power cut. The catch: Q-series and PTZ cameras need PoE+ or PoE++, so size the switch or cameras drop out at night when IR comes on.

Use STAR #1. The shape of a good answer: you put the customer's outcome ahead of the biggest invoice, you explained the trade-off plainly, and the honesty paid off later (a referral, a follow-on deal, or simply trust). Finish with what you learned.

For a sales engineer, prep is the selling. Block focused time on the customer's use case, then demo with it fresh. Skip prep and you sound like a salesman; only prep and you never close.

Use STAR #2. Show that each person owned a piece (technical credibility, the relationship, the commercial close) and that the deal was won because they lined up. That's Act as One.

Axis invented the network camera in 1996 and has built on it ever since. It now makes cameras, audio, access control, intercoms, radar and environmental sensors, all as network devices, and sells them through integrators. It competes on image quality in poor light, lower bandwidth and storage, cybersecurity, and open standards, so its products fit into whatever system the customer already runs.

Zipstream runs inside the camera. For each frame it finds what matters (people, objects, motion) and compresses that area lightly, and it compresses the empty background heavily. You keep the forensic detail and stop paying to store detailed images of a blank wall. Dynamic GOP sends fewer full frames when nothing is happening, and the storage profile adds B-frames for about another 50%. It stays standard H.264, so no special player is needed.

You don't need to rebuild the network. Axis cameras send several streams at once: record 4K to local storage for forensics and send a low-resolution live stream to the remote operator. Add Zipstream, average bitrate for a predictable budget, and AV1 on ARTPEC-9 cameras. It's also worth asking whether they need more pixels or better low-light performance, since a larger sensor often helps more.

Signed Video. The camera calculates a checksum for every frame, signs groups of checksums with a private key created at the factory and stored in Axis Edge Vault, and embeds the signatures and public key in the H.264/H.265 stream. At playback, the checksums are recalculated and compared, so any changed pixel is caught. The video isn't encrypted, so any player can still open it.

It depends on what they need to know. Lightfinder gives colour video in low light, which helps identify clothing and vehicles. Thermal needs no light at all and detects people who fill only a few pixels, even partly hidden by bushes or mist, with few false alarms. It's for detection, not identification. Radar gives exact position, ignores weather and insects, and can steer a PTZ. Radar-video fusion combines early detection with classification up close.

PoE over coax reuses the existing coaxial cable for power and data, up to 500 m depending on cable quality. For elevators, PoE over 2-wire uses spare wires in the traveling cable, up to 220 m. Standard PoE reaches 100 m. Pulling new cable is often the most expensive part of an upgrade quote.

Zone/Division 1 is the higher-risk area, where explosive gas or dust is likely. The camera sits in a heavy stainless-steel enclosure designed to contain an internal explosion, and only certified staff can install it, because a scratch on the flame path ruins the protection. Zone/Division 2 covers the larger, lower-risk area around it. The electronics are designed never to spark or overheat, so there's no heavy enclosure. These cameras are lighter, cheaper and easier to fit.

No. Blur comes from shutter speed. Frame rate only sets how many still images you get per second. A faster shutter reduces blur but lets in less light. If anything, lowering the frame rate saves bandwidth without making each image blurrier.

With sensors that produce data but no footage or recordings. The acoustic sensor recognises screams, shouts and breaking glass but never streams or stores audio, only metadata. Air quality sensors detect vaping and smoking. Radar can be used where cameras aren't allowed. Network audio covers announcements and emergency messages.

Point to total cost and risk, not the unit price. Storage costs more over time than cameras, and Zipstream reduces it. Cybersecurity: Edge Vault, signed firmware, a public hardening guide and vulnerability policy. Evidence that can be proven with Signed Video. Openness: ONVIF, VAPIX and SIP mean no lock-in. In the US, NDAA compliance matters for any federally funded project. And one partner can supply video, audio, access and sensors on the same network.

The mixer, amplifier, equaliser and the sound engineer needed to set them up. The speaker, amplifier and DSP are one pre-tuned unit, and the signal stays digital all the way, so there's no noise from long cable runs. It also tests itself: a built-in microphone listens for a calibrated test tone.

Reliability costs delay, and it only pays off if late data is still useful. A retransmitted live-video packet arrives after its moment has passed. UDP also supports multicast, which TCP can't, so any one-to-many streaming setup needs it.

TCP delivers data in order, so one lost packet makes everything behind it wait for the retransmission. On live video you see a freeze, then a jump, and the delay doesn't recover.

It works: RTSP interleaved sends media over the TCP control connection and gets through firewalls cleanly. But they lose multicast, so bandwidth grows with every viewer, and live delay becomes less predictable. That's fine for a few operators. For a big control room, show them the bandwidth numbers first.

About 130 TB. 1 Mbps for a day ≈ 10.8 GB, so 4 Mbps ≈ 43 GB per camera per day; × 100 = 4.3 TB/day; × 30 ≈ 130 TB. Then say that Zipstream means real bitrates usually run well below the cap, and the storage profile can roughly halve it again.

Footage is evidence, and evidence needs trustworthy timestamps. If cameras' clocks drift apart, piecing an incident together across several cameras becomes unreliable. NTP runs over UDP, so it fails silently: nothing visibly breaks, the clocks just drift.

Ping proves layer 3 works, so move up the stack: credentials; RTSP 554 or the RTP ports blocked; the VMS on another subnet with no gateway set on the camera; UDP 3702 filtered if discovery is the problem; the wrong stream profile. Try RTSP over TCP: if video starts, it's a UDP path issue.

At work you're on the LAN. From home you hit the site's public IP, and the camera only has a private address behind NAT, so the request never reaches it. Fix: a VPN or Axis Cloud Connect. Port forwarding is a last resort (HTTPS, non-default port, no default credentials). Never forward bare RTSP to the internet. More.

Same subnet: switch. Anything else (another VLAN, building, the internet): router, and the camera needs a default gateway.

In order: Zipstream (night is where it pays: about 5.4 → 0.9 Mbps in Axis's demo), average bitrate for a fixed budget, dual streams (full resolution to disk, low resolution to remote viewers), don't run 30 fps where 15 will do, and AV1 on ARTPEC-9 cameras if the VMS supports it.

Network 10.20.30.0, broadcast 10.20.30.255, usable hosts .1 to .254.

No. With a /24 the third number is part of the network, so .10 and .11 are different networks.

62. 32 − 26 = 6 host bits, 2⁶ = 64, minus network and broadcast.

The address has to stay stable for the VMS, firewall rules and DNS. A camera that gets a new lease effectively disappears.

Lighting · distance and field of view · pixel density · weather and abuse · VMS. Check.

Q premium: evidence-grade image, PTZ, thermal. P the professional workhorse for most commercial sites. M compact and cost-first for dense indoor counts. Check.

Not directly; they're different /24 networks. It needs a router between them and a default gateway on each device, and any firewall in between has to allow the traffic.

80 HTTP · 443 HTTPS · 554 RTSP · 3702 WS-Discovery · 5060/5061 SIP. Bonus: 1900 SSDP, 5353 mDNS, 123 NTP.

A lost UDP packet costs one frame; a lost TCP packet stalls everything behind it while it's resent. Live operators would rather skip a frame than freeze. UDP also enables multicast.

25 · 125 · 250 px/m (observe is 62.5, inspect 1,000). Check.

Low-light noise makes every frame look different, so a plain encoder spends bits on grain. Zipstream compresses the static background hard. About 5.4 → 0.9 Mbps in Axis's demo.

S basic streaming and PTZ · T advanced streaming (H.265, events, HTTPS) · M analytics metadata.

NDAA/FCC compliance · firmware and supply-chain trust (own chip, signed firmware, Edge Vault) · open platform with any VMS.

Five inputs → P for coverage, Q-PTZ where zoom matters, M indoors → ACAP analytics on the camera → PoE Class 4/PoE++ for PTZ → Site Designer for bitrate and storage → per-zone breakdown.

An example in Always Open: “Axis co-founded ONVIF and works with whatever VMS the customer already runs, so nobody's locked in. That's the kind of honesty I want to sell with.” Then say it in your own words.