Subscriber Discussion

500 Camera College System Going From Analog To IP

This is a college campus where we have roughly 500 analog cameras with digital recorders and nearly 50 IP cameras on 2 servers (ONSSI NetGuard EVS). Has anyone any advice on transitioning from analog to IP? Of course, any new buildings will be equipped with IP cameras, but what to do with the analogs is another story.

The IP cameras we have are getting older (3-5 years). Do I need to replace these as well? What is the life of an IP camera as far as technology goes? The analogs have been more cost effective since the recorders (depending on the environment) can last upwards of ten years. How long do you think it will take for analog to be pronounced dead? Should I be replacing every analog camera that fails with an IP camera?

I am an administrator responsible for security cameras, and I appreciate any insight into this situation.

Cynthia. Good question.

There are a number of factors in play.

The simplest solution is to keep the analog cameras on the existing recorders and manage the two platforms separately (i.e. use OnSSI client for IP cameras, use DVR client for analog cameras). Obviously, this has downsides from an operational standpoint.

In terms of transitioning off the DVRs, there are two general options - use encoder appliances or hybrid DVRs. Since you are using OnSSI, encoders are likely best (e.g., 16 channel ones - here's an example of a recent low cost one). If you get encoders, then you will need to connect / stream video from the encoders to a server. The good news is that such servers should be able to handle 100+ cameras, meaning less boxes to manage. On the other hand, depending on where you physically place the servers, this could increase network load.

In terms of replacing the IP cameras, it depends what the specifications of them are. If they are 3 - 5 years old, I suspect they are SD MPEG-4 cameras with poor low light and WDR. They probably should run for many more years but you'd like save some storage costs and improve image quality by switching.

As for replacing analog cameras with IP, the two key factors are (1) budget for higher cost of IP cameras and (2) spending for cabling / networking to connect the IP camera. For the later, either pull new category cable or add in EoC adapters.

This just touches on some of the issues.

Please provide feedback and more color and we can expand our advice.

Hello Cynthia,

I think the main question you need to answer is "what's the goal that my cameras should reach?". Probably it is to "see" well the critical events that you want to verify and/or prevent, giving images where the level of details/information that you need is enough for reaching the purpouse for what they had been installed. For example, if you want to see and recognize the face of the person who did something wrong, or to see a plate, or to see well also by night, whatever.. Not necessarily an old analogue or IP camera should be replaced if it already does what you need. Of course, it's highly possible that an old camera in outdoor without WDR or other newer features may not reach this goal. So for the cameras where this goal was already fullfilled (ex. indoor?), analogue or IP that they are, maybe you don't need to replace them now. For the others that don't, of course a newer adequate camera with new good features could surely help more. An old camera you can even always centralize it by an encoder if analogue, or by a timely "universal" VMS if IP (if you have already OnSSi): so you can time by time make the evolution according to your priorities.

But beware that to move from analogue to IP does not mean just to replace a camera: when you have 500 IP cameras you need before all a very careful and very well structured project for the network infrastructure! This is the first part of the plan you should take care, according to me: if you think you will go there, first of all think about the infrastructure able to sustain it all. It's not nice then to have 500 wonderful brend new perfect IP cameras, but then showing 1 frame per hour at the control room...;))



"Has anyone any advice on transitioning from analog to IP?"

The *easiest* thing to do is probably add cheap IP encoders to your analog cameras. You'll get the exact same image as you already have, but it will now look like an IP camera from the perspective of your OnSSI server. The tradeoff to this is that you'll be spending a chunk of money on encoders that will have little to no actual or resale value when/if you eventually upgrade your analog cameras to higher-res IP cameras.

"The IP cameras we have are getting older (3-5 years). Do I need to replace these as well? "

No, you don't need to, as long as they are still functioning satisfactorily. 3-5 years is just old enough that that might be at the lower end of the scale in terms of resolution, dynamic-range and low-light capabilities. It's possible that you paid $800-$2000 for each of these cameras 3-5 years ago and you could replace them with better models for ~$300-$500 today. Note that if you increase the cameras resolution, you'll also increase your stroage usage. But, these older cameras might not support h.264 video streaming and might use an older less efficient video compression technology, so it's also possible you could replace them with better units AND increase your video retention time. Hard to say without more details. In any case, it would be reasonable to expect them to have a 10 year lifespan.

"Should I be replacing every analog camera that fails with an IP camera?"

Many people, including myself, would say "yes". But this is driven by your own needs and budget. Your cost is most likely not just the cost of the camera, but the cost of cabling as well. If your analog cameras are running over coax, you'll need to replace the coax or use and Ethernet over Coax converter (at additional cost). Again, it might be cheaper to replace a batch of cameras at once, hire a cabling contractor for a couple of days and replace a chunk of your cable plant all at once. If the previous installer had the foresight to run Cat5 cable and use baluns to carry the analog video back, then your upgrade would be much easier.

"The tradeoff to this is that you'll be spending a chunk of money on encoders that will have little to no actual or resale value when/if you eventually upgrade your analog cameras to higher-res IP cameras."

An alternative that may be worth investigating is the use of budget H.264 DVRs as encoders. Whether your systems will support this, I don't know, but I've successfully experimented with using a Dahua-built DVR as a multi-channel encoder for a 3xLogic Vigil NVR. Vigil has supported generic RTSP support for a couple versions now, allowing the user to define custom RTSP strings; Dahua DVRs output an RTSP string identical to their cameras - put these two facts together, you essentially configure each channel on Vigil as if it were a Dahua camera, simply replacing the "channel=xx" parameter with the DVR channel number.

If your systems support this capability, this is not only cost-effective (as a four-channel, D1@30fps DVR can probably be had for under $200, or 32-channel for well under $1000), but when you're done with the DVR, you can put a hard drive in it, and it will have substantially more resale value than a standalone encoder.

First of all, thanks for the quick response. Please keep in mind that I only know what I pick up from my installers in the field which is the main reason I am registered for the upcoming IP camera course. With that being said, I find it all pretty overwhelming when it comes to the new technologies, the many choices and the network integration in security video.

We do use encoders and fiber transceivers in some areas. Also, we have a number of hybrid DVRs (Pelco) which accept 16 analog and 2 IP cameras. We have not utilized the IP channels yet because, apparently, the cameras supported are low MP, and we just haven't found the right fit. Now that you mention it, this would be a great way to transition as long as the supported cameras fit the bill. We have around 10 of these recorders, so that gives us 20 IP channels we can utilize without extra expense (except for the camera and cabling costs, of course).

Right now we have three different VMSs: Integral RemoteView, Pelco DX Client and ONSSI NetGuard EVS. It would make some people very happy if these were combined into one VMS whether we transition or not. Do you have any suggestions for that as well?

I really appreciate all the help from IPVM and members. I am so glad that I joined up, and I'm looking forward to the course. Please keep those pearls of wisdom coming!

"Also, we have a number of hybrid DVRs (Pelco) which accept 16 analog and 2 IP cameras."

Oh boy, you have them? :(

As I recall, they also have a throughput limitation (discussed here).

It sounds like your system is a bit of mess, with old Integral boxes (now owned by Pelco), some newer Pelco DVRs and OnSSI.

Putting them into 1 interface can be done, it's just a matter of cost :) and the cost can get very high.

I'd start by talking with Pelco. They'll need to determine how old those Integral recorders are and whether they are still supported and/or integrate with their current software/clients.

If you go to someone else, they will likely charge you additional software license fees for each camera they integrate.

On the other hand, you'll be stuck with all Pelco :)

One other question: Are you looking or able to replace the oldest DVRs (the Integral ones)?

I know Milestone can acquire video flows from some Hybrid DVR (but no clue if possible with Pelco's ones.. maybe yes?..). So being ONSSI Milestone-engine-based, maybe it is possible to take the flows from the DVR and centralize them into ONSSI? If you are satisfied of ONSSI features, this way you could try to centralize all in just one platform: you would use the old DVRs as encoders (so you would save money not buying new ones to acquire the analogue, using what's already available).

Is it possible?

p.s. Of course this way you will need to buy further licenses of ONSSI. But, if working, I think it would be the cheapest solution for first of all centralizing it all in a single platform.

We are an integrator in Florida and we recenlty did a college that had the same problem you did; lot's of disparate video surveillance systems (they had up to 4 different ones) that made use and management a nightmare. They did not have as much analog cameras as you did, so when we got them onto one platform they just replaced their analog cameras with IP and then added their existing IP cameras.

One thing to consider is some VMS systems have their own brand of encoders, and sometimes they will let you convert each analog channel to an IP camera license so it's no additional cost on that part when you convert.

Actually any VMS that supports analog to IP encoders should permit you to use your existing analog cameras along with IP cameras without having to replace them, and convert the analog cams to IP as the old cameras die off and you replace them with IP cameras. Whatever VMS you look at, you should ask what their theory and pricing scheme is on conversions to see what works best for you.

Another option for encoders would be to use HD-SDI. This would allow you to upgrade your existing analog cameras to HD-SDI 1080p cameras potentially reusing the coaxial cables you have in place and connecting them to the HD-SDI encoder. The HD-SDI encoders would replace your DVR's and would connect to your VMS via ONVIF Profile S. This HD-SDI encoder is listed on both the Milestone and ONSSI supported device list. There are also HD-SDI fiber transceivers that could replace your existing ones as well. Just another option.

Jeez, Keith, where could someone possibly find an HD SDI encoder? I mean, who makes them? Oh, wait, your company does.

Please make your background and interests explicit in recommendations.

The problem I see, and correct me if I am wrong, is that HD SDI encoders are pretty damn expensive. Yes/no? For instance, your 8 channel HD8R? What does it cost?

Overall, though, since she Cynthia has 10 relative new Pelco DX boxes, I doubt they are going to go anywhere.

Considering the costs of HD-SDI this is almost certainly more expensive than using an EoC adapter and a decent camera. How is this any kind of a benefit relative to all the other options on the market? It's probably even more expensive than just replacing the cables and using a standard IP camera.

I think it is a disservice to omit HD-SDI from the discussion as it is a viable option for many security applications, new or retrofit. All the components traditionally available for analog systems are also available for HD-SDI including DVRs, matrix switchers, fiber converters, encoders, and etc. When replacing analog infrastructure it can be the most cost-effective, but this can only be determined by taking into account the total cost per channel (hardware, software, installation, and maintenance). Also, HD-SDI pricing has come down significantly over the past year, so it should not be ruled out.

"I think it is a disservice to omit HD-SDI from the discussion as it is a viable option for many security applications, new or retrofit."

Even when the end user already has ~10% IP cameras? How does that make sense?

"Also, HD-SDI pricing has come down significantly over the past year, so it should not be ruled out."

So has the pricing for IP cameras, so HD-SDI is still at a relative price disadvantage.

I don't see any concrete data in your post, just a bunch of stuff that says it should be considered, and a comparison to analog product availability. To me, it seems the OP here is interested in some kind of a modern integrated CCTV system, so your positioning of HD-SDI should be relative to current-generation IP technology, not outdated analog components.

HD-SDI is a great potential option for little low-demand systems where the user wants to go a more DIY route, as long as their cable plant supports it. The only benefit HD-SDI really offers is cable re-use, if you can't guarantee that the existing cables are going to support it, it is barely worth considering.

You're also still very very limited in available product. No cameras greater than 1080p on the market. Limited (if any) PTZ options (has anyone released standards-compliant PTZ's yet? And if so, at what price?). No panoramic cameras or other "specialty" stuff.

IP is proven as a CCTV transport. You can argue the merits of IP for CCTV, and any other number of things, but you can't deny it has a foothold that is not going away, and will continue to evolve. HD-SDI does not seem similarly stable to me, it is a broadcast TV standard wedged into the CCTV world, with no clear benefit or CCTV-centric product roadmap. Limited product selection, high-cost, no clear support from major software platforms, no major brands really onboard with it. Suggesting something like that for a 500+ camera system is borderline cruel, IMO.

Keith, disclose the exact pricing of your products so we can properly evaluate their 'cost-effectiveness' or refrain from commenting. We do not need your or any other manufacturers vague marketing claims in our discussions.

Thanks again for all the responses.

To answer your question, Mr H, I was replacing the Intregal recorders with Pelco ones as they die. I guess I'll need to revisit that strategy.

Replacing the Integral recorders with Pelco is certainly an option to consider. At that point, though, you may just as well standardized entirely on Pelco and get rid of the limited number of cameras on OnSSI.

I am not saying this is the best scenario but if you really want to get everything on a single user interface, this is one straight forward approach.

Alternatively, you can add in PSIM software but that might add $100,000+ just for that overlay layer.

Cynthia, before thinking about centralizing or replacing where and what I repeat that according to me your first task should be to evaluate carefully camera by camera if what you have now is working or not accordingly to your needs and expectations. You know, us Italians are fond of antiquities and we always think more than twice before throwing off something..;))

Jokes asides, "needs and expectations" is definitely a relative point and it comes out from the answer to the question "what do I ask/need from my video surveillance system today?".

Try for example to get information internally about how many times it happened a critical event (and what is for the College "critical event" or "event-of-interest" it's definitely another focus point to recall and fix) and the video surveillance helped to solve/limitate/manage it? Or on the contrary, how many times "you" happened to say "Ah, if we had had more resolution that face/plate/detail would have been more recognizable..." or "Ah, if it had been more light...." or "Ah that camera was too old and flickering, if it had been better...." or whatever recalling to you if the system was effecitve or not in the past? And make camera by camera a check if it able to give enough quality of image (resolution, contrast, lighting, dynamics, ..), according to the specific environment it is put and the specific purpouse it should have.

After that you will have a report where you will have 500 results yes/no.. It may even happen that surprisingly you will read 50 "no" ans 450 "yes". In this case you may decide to think about centralizing all, by acquiring the analogue in a way or in another. But if instead, for example, you will have 400 "no" and 100 "yes" (or "yes, more or less.."..), so the solution is maybe actually to replace them almost all; but this, as written before, means first of all to make an accurate network infrastructure project.

Consider that anyway to make a transition analogue/IP has a cost. If you decided that anyway you will go to IP sooner or later, think well about maybe finding the budget to do it all now; because anyway if you buy now encoders or whatever to keep the analogue today, when tomorrow you will replace anyway all to go to IP these encoders will be useless then. So, it's like when you do a financing in a bank: you pay after, but more. If "tomorrow" (because of the report before) means 4-5 years, it has surely more than a sense; if "tomorrow" means probably next year, I think not and maybe the best solution would be to clench your teeth using this one that you have until you don't have the budget to do it once and all. IMHO..;)



John, is it possible that I read here time ago a report about how to evaluate a camera according to the real needed features (resolution, light, ..)? Do I remember well?.. If so, maybe that report could be useful to Cynthia to evaluate the effectiveness of the actual cameras?..

Since you mention NetGuardEVS I believe you are likely several revisions back on your OnSSI software. If that was updated and if you are using some Pelco DX8100 DVR, those can be pulled right into ONSSI as encoders. One other thing that you need to consider is how many PTZ cameras you have and how they are controlled. Some encoders cannot handle coaxitron and other "up the coax" control. You need to be sure when you are evaluating this you know exactly what control you need as some integrators might suggest a low cost encoder but that encoder may not be able to control the PTZ (depending on the supported protocol), and then you may be adding code translators and everything else, which you may find your initial bargain is not a bargain after all!

There is no easy answer here but I would suggest that you develop a complete phasing plan which will require you to work with a design firm or a trusted integration partner. This phasing plan will take into account all of the things I mentioned previously as well as fitting the conversion into your budget...or at least allowing you to budget for it.

FWIW, I hate the idea of encoding that many streams. It seems like a waste. Even with what I would consider a cost effective encoder and license scheme you would be in the $100,000 range and that does not include any servers or storage, but replacing all of this cameras and dragging new cable, switches, etc. would easily be 5-7 times that amount depending on the camera model, existing network switching, etc.

Hi Cynthia,

I admin an OnSSI platform that I virtualized with VM and run just under 300 cameras. I tell folks we do not handle any converters or other nonsense. Just straight IP cameras with direct cat runs to telco closets. Usually you get people wanting to use encoders and such because it may save them a few dollars but like anyone will tell you it involves additional gear to troubleshoot later and the cameras probably end of life anyway. Using old cameras will just cost you more in storage being they wont do 264 most likely.

We leave the analogs alone and let people migrate to IP when they are ready. We have a base and backup as well as 5 recorders so far which run smooth as silk on vm.