Cisco MediaNet for IP Video Surveillance Examined

Now more than 4 years from Cisco's first foray into the video surveillance market, the industry waits for Cisco's big move or key advantage to be revealed (beyond their obvious channel strength). Indeed, many of their surveillance offerings have been re-brands or 'me-too' offerings (see Cisco-Pelco and Cisco PSIM/Analytic sources).

At ASIS 2010, an announcement that received little recognition has the potential to have a significant impact. Cisco announced that its IP Surveillance Cameras would start shipping with their MediaNet [link no longer available] video optimization offering built in. Cisco markets MediaNet as a solution to the demanding requirements of running various applications over a single IP network including data, telephony and video (such as surveillance).

We think this is interesting because (1) it leverages a clear technological advantage that Cisco has over IP video surveillance providers - IP networking and (2) it addresses a core need - running IP cameras over the IT network.

Inside the Pro section, we dig into MediaNet, examining its capabilities, strength and weaknesses.

[premium_content]

Recommendations

The adoption of medianet for IP video surveillance will depend on two key factors:

• The use of enterprise class Cisco networking gear
• The comfort level of running IP video on the corporate LAN

The best fit for Cisco's medianet/IP video surveillance are for larger organizations already employing Cisco enterprise class equipment, and require guarantees that additional IP video will not adversely impact their business critical network applications.  For these organizations the incremental cost of a medianet will be minimal, the cost savings of not having to deploy a secondary IP video surveillance network will be substantial, and the reduced labor cost for camera installation and access layer provisioning will be modest.

Organizations that do not already employ Cisco enterprise class networking equipment may find the cost of a Cisco based medianet unjustifiably expensive. These organizations may elect to over-provision bandwidth, or employ their vendor's brand of 'medianet' using open standards or proprietary solutions.

Determining whether an organization meets these two criteria will be critical. We believe that many larger (even mid-size) organizations will find Cisco's medianet to be a valuable play. Secondly, given Cisco's strong reputation with IT groups, we believe Cisco will make a strong case for medianet to organizations fitting the profile. On the other hand, smaller organizations will likely dismiss the medianet strategies out of hand as being too expensive and complicated to implement.

Key Short Term Barrier: As of Nov 15, 2010, MediaNet is only available for Cisco 4000 series HD cameras. Cisco will need to expand this to their other cameras and potentially add more cameras to offer a sufficient product lineup to motivate uptake and provide a complete solution.

Key Findings

• Medianet provides the best value for networks of large scale such as service providers and large enterprises
• Networks already supporting Cisco IP Telephony and converged networking are strong potential fits for Cisco medianet / IP Video Surveillance
• Larger non-Cisco networks will have a more challenging migration cost/benefit analyses, as there are compelling business reasons to migrate or adopt Cisco medianet/IP video surveillance
• Medianet is an architecture enabled by both existing and developing networking features and standards and proprietary Cisco products
• Medianet 2.1 includes access layer features that can function and provide modest benefits in smaller scale networks deploying Cisco video surveillance systems, but at a steep cost
• For Medianet access layer features, minimum hardware requirements call for enterprise class switches (e.g. 3560 Catalyst) and thus requires a significant premium expenditure over SMB products
• Currently only two IP cameras are supported, restricting the flexibility of the system's design

Medianet Overview

Medianet encompasses a whole discourse sponsored by Cisco on how to best deploy networking technologies to manage the exploding use of video across enterprise [link no longer available] and provider networks [link no longer available]. Such networks are much more likely to have competing IP video applications running concurrently on their networks:

• Telepresence
• Digital Signage
• IPTV
• VoD
• Video Conferencing
• IP Video Surveillance

Medianet is an evolving strategy to address the challenges of efficient delivery and reliability of simultaneous, often competing IP video applications by the implementation of general networking features and application specific products. Currently, many of the medianet best practices can be 'turned on' in existing Cisco and non-Cisco networking gear. These medianet technologies are simply, rehashed for IP video, open networking standards such as high availability (e.g. STP/RSTP, and HSRP/VRRP), IP multicast (IGMP, PIM), and QoS (e.g., DSCP, 802.1p) among others.

Some proprietary Cisco technologies include Video Stream for wireless transmission, Cisco PfR (performance routing) and Cisco WAAS (WAN accerleration). Medianet includes other technologies available in the Cisco MXE 3000 that recognize end-point client device profiles and adapt, transcode and filter video content based on client connectivity, processor, etc. 

While much of a medianet architecture pertains to already existing 'intelligent' networking features, there are some application specific products and technologies relating to end-points (e.g., Cisco 4000 series IP cameras) and switches (e.g. 3560/3750 Catalyst switches). These are known as access layer Medianet features - see below for discussion.

Medianet Access Layer

In Medianet 2.1 Cisco has introduced application specific products that provide auto configuration or plug-n-play features such as Auto Smartports, location information, auto-QoS, and DSCP marking. For IP video surveillance applications only the Cisco 4000 series IP cameras are fully optimized to benefit from these features:

• CIVS-IPC-4300 - Box, 30fps at 1080p, H.264/MJPEG, (online $1000)
• CIVS-IPC-4500 - Box, 30fps at 1080p, H.264/MJPEG, analytic DSP (online $1400)

In terms of networking devices, the auto configuration features require one of the following switches at the access layer:

• Cisco Catalyst 3750
• Cisco Catalyst 3750-E
• Cisco Catalyst 3750-X
• Cisco Catalyst 3560
• Cisco Catalyst 3560-E
• Cisco Catalyst 3560-X

The default LAN base image will provide the necessary software features, but the IOS software must be 12.2(55) SE or later. Alternatively, the access layer features can also be implemented using select 2800, 2900, 3800, and 3900 series ISR router chassis equipped with approved switch modules. For details on compatible hardware/IOS see Cisco's datasheet covering the compatible hardware/IOS.

For networks dedicated to IP video surveillance systems the only benefits that would be achieved are those at the access layer, namely auto-configuration or plug-n-play features. The utility of other medianet features such as QoS are significantly diminished in dedicated surveillance networks, as there are no competing voice or video applications. Moreover, a significant cost is entailed on the purchase of higher end conforming Cisco access layer switches just to save a modest degree of labor.

The cost of the lowest priced conforming 24-port PoE switch retails at $3795 ($2000 online). A typical SMB switch with 24-port PoE can be found well under $1000 online (e.g., ~$500 for a Cisco SRW 24-port PoE switch). Also the two supported Cisco 4000 series IP cameras max out at 1080p starting at roughly $1000 online. Such cost premiums and product limitations make the Cisco video surveillance/medianet option less attractive for a dedicated surveillance network.

Medianet for IP Video Surveillance

First plug in the Cisco 4000 series IP camera to any switch port. CDP (Cisco Discovery Protocol) recognizes the end-point type as an IP security camera and provisions a VLAN, IP address and location information to the camera. The port is auto configured with spanning tree portfast to expedite video transmission, and the camera's preferential DSCP marking is observed or honored by the switch. The camera then automatically registers with the Cisco VMS and relays it's location information to the VMS as well.

According to Cisco this eliminates any pre-configuration of the IP camera, and minimizes technical skill required to install IP cameras. The switch provides all connectivity configurations to the IP camera and allows the camera to automatically find its VMS. The camera can then be accessed via the VMS to optimize video streaming and camera settings. Moreover, the installer doesn't need to perform anything beyond attaching an Ethernet cable between camera and switch.

To get a better sense of the cost considerations let's compare a large pre-existing Cisco network adopting Cisco video surveillance and medianet against one electing to go with a non-Cisco video surveillance product. We'll estimate approximately $100/hour for an installer's labor rate. Assuming an average of 30 minutes (or 1/2 hour) per camera (15 minutes to configure camera and 15 minutes for switch port) our cost differential is $50 per camera for the non-Cisco product. Also, any future moves, adds, and changes will entail additional labor costs. For a 100 camera system this represents $5000 and $25,000 for a 500 camera system.

See a Cisco video demonstrating auto provisioning:

Medianet Network Architecture Considerations

The prospect of integrating bandwidth intensive applications such as video surveillance into a corporate network has traditionally been a dicey topic for IT decision makers, concerned about the potential malignant impact it could have on their networks and careers. However, Cisco's declaration of commitment to medianets will build confidence in the IT community that such a diversity of applications (even extremely bandwidth hungry ones) can 'live' harmoniously together all on one really 'intelligent' network.

There are essentially two types of networks potentially suited for a Cisco medianet/IP video surveillance adoption. Both categories will tend to already have converged networking technologies in place such as QoS and IP telephony.

• Large organizations with existing enterprise class Cisco networks
• Large organizations with non-Cisco vendor enterprise class networks

The first case organizations will typically have the necessary higher end Cisco access layer switches in place, especially if they are supporting an IP telephony environment. Thus, the incremental cost to medianet will be minimal. Medianet adoption is essentially a promise by Cisco to provide long term leadership and product development to address the concern of video applications menacing other business critical applications on the corporate LAN. This will be a difficult value proposition for IT groups to dismiss.

The second case network type, having little to no Cisco gear, may elect to address the growing challenge of managing IP video applications across the network with a mix of their vendor's proprietary solutions (if available) and various open standards, choose to over-provision network bandwidth, or even elect to build an entirely independent network just for IP video surveillance.

Because of the risks and costs involved in these options, these organizations would be wise in considering the value proposition of Cisco's medianet strategies. Obviously there will be CAPEX investments in access layer gear, and potentially distribution and core equipment down the road, but the prospect of having an industry leader such as Cisco join forces to ensure a productive and high performance network (despite the introduction of lots of video) will be hard to ignore.

In our estimation, any large network from essentially any vertical market will qualify as a potential Cisco medianet/IP video surveillance adopter or beneficiary. The degree to which adoption will cost and the extent to which the organization may benefit will of course vary according to certain factors. However, the number of these candidate organizations is fairly high and less of a minority than might have initially been presumed.