Right, our site manager is wanting to get another camera installed on our CCTV system, and I think it is high time we move away from the analogue system of old (with it's roughly 0.4MP resolution per camera), and move to a proper IP CCTV based system.
Ok, i'm not suggesting we do this in one giant lump, as buying all the video servers, storage server and PTZ IO video server would make my boss wish to bury me. What I'm thinking is that I want to get something and show her how good it is, compared with the old system.
So... What do people suggest? I know I can get a camera on buy and try, and software on free trials and the like. But what if my boss goes for it? I want to then buy the equipment to start recording things properly.
What I really want to know, is what should I plan to buy, if I want a system that can take a few cameras at first, but potentially scale up to 32 cameras in the future, as we slowly but surely convert them over.
Because as it stands, to support that amount of data, the system would need to scale to around 4TB of storage...
Suggestions, ideas etc...?
Also, it is more than likely my boss would want such a system to be a 'managed service' ie. done by a company. Our local CCTV/security company seems to dislike IPCCTV with a vengeance, so any suggestions as to a company that would do this?
In a previous job I've successfully used Axis IP Cameras together with a standard server/XP workstation to record the various cameras. The Axis software's a bit clunky but it does work; the cameras come in either wired or wireless variants, with external housings if you want to put them outdoors.
The network loading isn't noticeable (or at least wasn't in the installation I put in - around 6 or 7 wired cameras) and I set it up so that reception had a scrolling display of the main cameras, together with a "public" monitor showing you that you were on camera as you walked through the main entrance. The server console allowed viewing of all cameras simultaneously, along with full recording control, and users with the client software could view all/any cameras depending on their assigned user rights. All displays were on standard VGA TFTs, no specific video hardware necessary.
The whole system worked perfectly! (Still does, to the best of my knowledge.)
If you want further info, PM me for details and I'll help as much as I can.
The video managment software bundled with the Linksys cameras will support upto 16 cameras without having to purchase additional software/licences so with those its only the cost of additional cameras and storage you have to worry about.
Thanks for that. But I'm quite happy with that sorta side of things, I'm more worried about buying a 'server' that would scale storage-wise for it. For example, would a NAS box do the job? Or would the I/O be too much for it to handle for standard 7200 rpm SATA disks? Would I have to go for more expensive SAS disks and a more expensive server? etc...
I want to be able to do this cheaply, but reliably and it has to be scalable. (ie. I don't want egg on my face X years down the line when they add camera 26 and find it doesn't work, they've ran out of space etc...).
If you're using Gig-E, then NAS should be fine. I have recently used the D-Link setup with the D-ViewCam software for a restaurant/bar/dining hall. I just built them a server with 2TB of RAID5 storage, so there was no external storage to worry about at all. I don't know about having this managed by a third party, but it's really easy to get working as long as you can get the motion sensing set up. Then it only records when something is in the view of the camera, saving a bunch of recording space. You could easily, and pretty cheaply do 3TB of RAID5 storage, and it sounds like you would only need the 4TB if you do 32 cameras down the road, in which case, you would have time to save up for SAS. I built the server with just a core 2 duo for less than $2000.
The last time I investigated (regarding Axis cameras) it was a balance between the number of cameras recording to the same place and the desired frames-per-second. If you're prepared to sacrifice fps you can have more cameras writing to the same device and vice versa. We have 16 analogue cameras saving to a blackbox camera server with 500Gb of storage which lasts (on average) 3 weeks before data gets overwritten.
During lesson times (assuming you're filming corridors and the school grounds) you can switch the cameras from continuous to movement-based recording to save diskspace.
If you use movement triggers outside at night to conserve disk space (and thus have a longer period of time on disk), be aware of spiders (webs near/on cameras), strong shadows and blowing tree branches that will keep the camera filming.
Have a look at some of the stuff from Dedicated Micros, we have 2 cameras at the last place I worked at that you could read what was on the side of the passing trucks on the main round nearly a mile away!
My current recommendation would still be ACTi megapixel IP cameras; best value for money and free video recording software for up to 64 cameras.
They perform very well and have a good product roadmap which should prevent future egg-on-face.
Currently their free software only manages their own brand cameras, but others such as LuxRiot and Icode will manage ACTi cameras plus others such as Axis.
I know that others have had SCL build them suitable good value servers with lots of TeraBytes.
This new video NVR from QNAP might suit some folk - a dedicated video recording NAS box off-the-shelf.
If your local CCTV installers are IP-shy then installation and service can be carried out by anybody that you trust with a ladder and wiring. As an IP solution on YOUR network you will need to do the lion's share of the setup anyway.
This is a hard one to answer fully without knowing the application but here goes:
There is 2x thoughts on the recording solution didcated appliance devices NVR with embedded OS or Windows based software installed on server architecture. There are + & - on both.
Cameras, a large array out there and this makes choosing the recording platform critical. I found that very rarely can you use one manufacture for all so need multi vendor support. Megapixel is great but this brings with it storage and bandwidth issues. I would like to advise customer that they should put in nothing else, but cost can sometime be the issue.
I would look at changing out the cameras that maybe you want more resolution on and then use codecs to bridge the analogue to the IP recording solution.
we have installed the Sirrus solution NVR's these are ruggerdised built on XP Embedded and their new 4u chassis supports 8x drive bay. You can install as many as you require max 24 cameras per server. Various viewing clients.
the other software we are looking at and trialing at the moment is Exacq this is also multi cameravendor support and will expand 100,s to 1000 camera installs. This is installed on a linux or windows server platform.
Sony, Axis, IQEye, Vivotek, Dynacolor, Panasonic, have installed these product previous. Have just started playing with the Acti's as mentioned in other quotes these are well priced but i have issue with current sirrus NVR new version out early next year to support H264, MPEG4 etc Acti deliver over RTP/RTSP ok with Exacq
if you want any other info then more than happy to answer your questions
We keep having Security firms in, and they all say the DM http://www.dedicatedmicros.com/uk units we have are the best in the industry, at which point I usually complain about them, which they usually blame on poor training (yeah, but it's not ICT, so I got no training, it's only mine, because they put it in my secure cabinet). We have 5 of these Product Details - Dedicated Microcomputers Limited all with 16 cameras, they come with 300gb-500gb disks (or it may be 1tb now), if they are configured right (our aren't) it will copy/move the video across to another server (they suggest NAS boxes).
With 16 cameras all recording, at a decent resolution, we get about 1 week on a 300gb disk, so it needs another storage system. I think you can have the extra disk directly connected, but I'm not sure.
Best get hold of Chubb or similar, get them in to quote. it might be good to check with your LA, as they might have a deal with some security company.
Pay for the training, pay for the annual maintenance (as this way they come in and clean your cameras).
Make sure you know where you want cameras, or at least what they are for.. we have 80 cameras, which are a missmatch of security / student surveillance, and as a consequence, they're not much use at either. We should have had cameras pointing towards each external door - one in, one out, but the only one that is any where near is the one in the reception/foyer, which is so high people can walk underneath it, and not be seen.
We have 5 external cameras that loop round, and are completely useless... too slow, and you miss stuff that is happening, too fast and you can't see what happened..
Oh, and those DM boxes all have network connections, so you can view it remotely, via software or web/java, apparently we could set a website up to auto-load images... but I never have enough time..
Thanks for your advice. But I have now been blind-sided by my manager, after our local security firm told her that IP CCTV cameras were unreliable and whenever a server went down, we'd lose CCTV.
Ignoring the fact that my network infrastructure has had 100% uptime (outside of configuration changes, which amounts to a total of about 2 hours across the entire switches over a year and a half). Or the fact that the servers I run have had an average of a 98% uptime (2% includes downtime for maintenance, and power cuts etc...). Or the fact that the analogue cameras we have installed already have had to be repaired multiple times (giving a roughly 94% uptime), or that we had to spend thousands of £££'s when my manager moved office and wanted the DVR's moved up with her. Or the fact that on the occasions we've needed to identify someone's face in a) a criminal damage situation and b) a harassment situation, we have been unable to do so both times, due to low quality footage. Or ignoring the fact that installing each new analogue camera costs us between £800 and £1000 depending on location.
Oh well... We might as well go back to chalk and slates then, as they don't 'go down'.