Envysion recently introduced support for ACTi IP cameras.  In particular our customers are demanding a mid to low priced camera that offers megapixel resolution.

Our customers needed a reliable, high resolution camera option at the lowest possible price.  ACTi’s megapixel cameras supply good quality images that exceed our customer’s requirements.  Our customers also like the several flavors of dome cameras which ACTi offers in megapixel resolutions.

We liked the easy integration with ACTi’s software and the near ubiquitous Power over Ethernet support.   For software integration, all we really need is the RTSP protocol for video transmission and a simple web interface that we can create scripts to auto-provision from our Hybrid NVR’s.  We also received excellent support from their development team.

Thanks ACTi!

Over at IP Video Market Info John Honovich has a good discussion going regarding megapixel without IP.  The topic might also be labeled: “What’s the best way to retrofit an existing analog system for megapixel video?”

Currently megapixel requires IP cameras which can be a costly upgrade if you are all analog today.  IP is very different from an install perspective and requires new cabling and complex configuration of IP addresses, cameras and NVR’s.  Or does it?

IP cameras can be auto configured with a good vms system. we at Envysion implemented auto-discovery and configuration in less than 30 days for Axis IP cameras. Using zeroconf and a simple http api we can now do it in just a couple days and push the software update into production automatically.

Ethernet can run over coax (IEEE 10Base2 is based on rg58 (50ohm), but with tweaks can run on rg59 (75ohm). There are a number of products that do this now, but they are currently a bit clunky and perhap too expensive. But that can be fixed. Just need some IP cameras with an integrated Ethernet over rg59 interface to get the cost down and provide a clean install.

Remember those images where if you “defocus” your eyes just right you can see a 3D image jumping out at you on the page?  These “stereograms”

A new imaging chip from Stanford University uses a similar principle to capture an image which includes depth information.  It sounds like they are looking at digital photography, but one can also imagine this technology trickling it’s way into surveillance video.

Business data (meta-data) needs to be constantly scanned and reported onto to maximize it’s value.   It’s important that there are many ways to scan, slice and dice this data because each business is different and the information sought also frequently changes.

Storing video and business intelligence separately allows Envysion to have more flexibility as well as reliably scale to larger and at lower cost that systems which store business data embedded into the same database as video.  Storing business or “meta” data in it’s own, separate database really helps give one a lot of flexibility to run all kinds of analysis using regular database and analysis tools as opposed to specialized tools that have to deal with a proprietary video database format.  In addition, the meta data is much, much smaller than video data, so the business information database is much less costly to scale up to support a lot of data.

Salient Systems makes a good argument that video management systems should store video and meta data separately for reliability’s sake in their whitepaper on Modular vs. Dependant Design.

Envysion is having so much growth that we’ve recently had to expand our database capacity to store the ever higher volume of event data flowing in.  (We still store video and the video database at the remote site)

Adding database capacity isn’t such a big deal if one just goes out and buys some “big iron” servers, a big storage area network system and pays a huge license fee for the latest Oracle enterprise database software.  This is the traditional enterprise way to grow a “data warehouse”.  This is sometimes called “Scaling Up” and it is extremely expensive, often taking up the a huge slice of any IT organization’s budget whenever there is truly a large amount of data stored in databases.

Instead, Envysion is using cutting edge, cost savings methods to grow our hosted service.  By “Scaling Out” our data across multiple commodity market server hardware and using the MySQL database, we’ve been able to keep costs low, keep performance high and still grow big, like 100’s of millions or even billions of rows big.  This is the way of highly scaled and efficient web applications.

Why does this matter?  Simple!  The costs of a scale out is at least an order of magnitude lower than the big general purpose data warehouse solutions out there.

But doing a “scale out” isn’t so easy.  While a “scale-up” design utilizes a small number or even just one, single, ever larger growing (in both size and dollars!) database.  The scale-up design is somewhat simple from an application perspective.  Conversly, doing a “scale out” really requires that one know how your users want to see your data.  You also need to know your application and data well enough to distribute the data across many low cost servers.  In other words, there’s not yet just one simple way to “shard” out your data across many database servers.  (although many innovative companies such as codeFutures dbShards and Dataupia are working on this as we speak)

When I hear talk about the virtues of “Software Only” video systems, I kinda think, huh?  Since when is a camera made of software?  I think what is really being talked about are the virtues of a multi-vendor solution as opposed to a single vendor solution.

Multivendor network video systems are a great direction and the wave of the future because they deliver more value to the customer.  They give customers more options to put together solutions which better fit their needs and better prices

Today’s IP video solutions which promote “software only” are really saying, buy your choice of cameras (mix and match for your needs) and your choice of recording hardware (PC’s made by Dell, HP, etc..), and buy your software from us.  That’s a huge leap forward from the analog video surveillance market which has really only had competition on the camera front and locked in customers to buying recording hardware from the same suppliers as the software.

What seems to be taking some time however is the increased cost of the IP cameras is worth the benefits of a multivendor recording solution. It would be great if there were some open standards DVR’s out there, but alas, they don’t seem to exist.  If and when this happens, I can see a lot of value being created for customers as very cost effective DVR’s can be sold as part of a multivendor solution.  Imagine being able to buy whatever DVR suits your needs and cost requirements and being able to link that DVR with the advanced software platforms out there today, such as Milestone, OnSSI and of course, Envysion Video.

Having worked in telecom and enterprise networks for 15 years I strongly believe multivendor networks are often the right thing for the customer.  There are so many things a network needs to deliver that to deliver the “best of breed”, multiple vendors often have to be used.  Of course, there’s the healthy pricing competition that comes along with having more than one vendor bid a solution, and continue to bid upgrades and growth of that system going forward.  As a customer and builder of network solutions, it’s no fun to be locked into a single vendor’s solution.  Price is a huge concern here when the vendor knows how much it will cost the customer to switch vendors.

In the telecom engineering world a huge focus area is on having inexpensive, robust devices out at the edge of the network which could be 100% remotely managed using automated processes.  All the edge devices (say a router that is at a customer’s location) must have a very consistent and methodical way of being configured and managed in order to scale up to serve millions of customers.

In the managed video world, probably the best inexpensive and robust  edge devices are “DVR appliances”.  But unlike a managed Ethernet Switch, an IP router or DSL modem, you can’t pick and choose your DVR box from a list of vendors and hook up to one’s centralized viewing, command and control system.

In the managed network equipment world, we have well documented SNMP interfaces, configuration files and scriptable command lines that can be integrated into centralized management systems.  But in the DVR world there are proprietary protocols, no remote access to important administrative features and a lack of scriptable or programmable interfaces to get at all the functionality of the DVR.

For central viewing, command and control we made our own Envysion Video service which is built from the ground up to scale to manage information events (mostly point of sale information) and video.  We also had to make our own DVR appliance (The Envysion EnVR) to meet our needs, but it’d be great if we could buy them from 3-5 vendors instead.  Maybe someday.

In the late 80’s and even early 90’s, IP networks and the products used to build them were very immature.  Companies were building network routers out of PC’s running off the shelf unix systems and the software being used to run the Internet backbone was literally, fresh out of the lab from last week.  Each box had hand-configured specialized configurations by engineers with a lot of tacit knowledge of how the system worked. Some vendors had proprietary systems that were easier to manage, but only worked with their own software, between their boxes and nobody elses.  This just didn’t fly very far in the network world where everything has to talk to each other.

Being a “network guy”, I see a lot significant parallels to this in the world of network video.

Arthur C. CLarke’s third “law” of prediction states:  Any sufficiently advanced technology is indistinguishable from magic.

However, these days it seems expectations have risen.  Sometimes it seems magic is indistingishable from technology.  As a technical person, have you ever been asked to “just make it work”?

How about just making a megapixel IP camera without storage sending it’s video over the uplink of a broadband DSL line?    At 7Mbps, that DSL seems like it should be able to handle it, right?  Well the uplink is only 896kbps and the IP camera is trying to use almost 3 times that amount.

Sure, with a little creativity, one can make that camera “work”.  Reduce the quality, the frame rate and well, the resolution.  It might not be megapixel anymore, but it’ll work.  Of course, the low cost solution that preserves the most options is to get some storage local to the camera.  Envysion’s system allows you to stream the video at a lower quality for review and then upload the video from the remote location to the Envysion datacenter or to your PC once you’ve found what you’re looking for.

Other solutions like VideoIQ’s camera with built-in storage might be attractive if one cannot use a recording device that is external to the camera.

If you have a network (and who doesn’t these days?) IP cameras sure are a lot easier to access than analog cameras.  But this ease of access also makes it easier for un-authorized users to access IP cameras .  Imagine if someone in the back alley could stream video of your back office using your wireless network.

Most IP cameras I’ve worked with lack a number of enterprise grade security features.  They need to implement SSL for encryption and they need to integrate into a central authentication system such as RADIUS or LDAP.   However, even without those technologies, there are a couple of basic things one can do to secure your cameras.

The first thing to do is to change the factory default passwords.  The latest firmware revisions for some cameras force one to configure a password the first time a camera is accessed after being powered up.  That’s good.

The next level of security that might be worthwhile is to configure a VLAN just for video surveillance network.  Done correctly, this will help mitigate several risks when IP cameras don’t have good security features built-in.  This will add to administrative overhead and possibly additional equipment costs, but it’s not as high overhead as some of the next additional steps.

To additional key elements of security that are often missed are to:

- Perform logging on your network by having all devices log important events, scanning and alerting on these events.  (Especially logins, configuration changes, etc)

- Regularly audit your network infrastructure to make sure it’s configured correctly

Both of the above tasks can be large undertakings given the large number of IP networkable devices in an IP video network.

Additional layers of security that are possible, but have signficant overheard are MAC based access control takes you even further.  802.1X technology offers a strong way to control who can plug into your network.  802.1X offers perhaps the strongest layer of security for getting access to a network, it is also difficult to manage and operate.  So difficult that it’s not seen much deployment in networks. There are some companies like Cloudpath networks who offer a management tools for wireless networks.  Few cameras support this (Axis with firmware v4.4+ does however).

The best technical standards are ones that are actually used.    And which standards are more likely to be used?  How about standards that do the following:

- Proven to work because code exists that actually implements the stands and actually works

- Consists of a collection of tools of which only a few are “must have”

- Free software available to help you implement it

- Utilizes existing, widely available and understood tools

- The specification is freely available

These all lead to practical, working systems that can be quickly implemented. Free code which actually works and does something valuable is an amazingly powerful force.  Such things often become standard without any special governing body approving them.  ie: A defacto standard.   This approach is partly why the Internet is the way it is.

The last point is pretty important.  How many websites would there be if you had to pay thousands of dollars a year for the privilege to know how to create one?

These points are so important to the Internet that all Internet standards are freely available to anyone for download and the vast majority of them have freely available open source implementations.  One of the “founding beliefs” of the Internet Engineering Task Force from David Clark is: “We believe in rough consensus and running code“.