City Wide Surveillance RFP Examined

In this note, we examine a city wide surveillance project of over 100 IP cameras across more than 20 locations. This is an especially interesting RFP to consider as the document is simultaneously grossly under specifies performance yet hard specs specific models of hardware that may not meet the actual needs of the city. Members, review the city wide surveillance RFP [link no longer available].

Summary of key points:

• Of the approximately 140 cameras, ~90 are SD and ~50 are MP.
• The City wants access control integration, specifically, "Installation shall require the cameras that are mounted at access controlled doors be integrated into the output of Presidio access system to automatically video record any card read."
• VMS specified is Aimetis Symphony Standard edition. Accepted alternative is Ganz VNS NVR software.
• Camera line specified is Ganz including IP and MP and 8MP, multi-imager cameras. Sanyo is listed as an accepted equivalent including an Immervision lens.
• No functional or performance specifications are offered for either cameras nor VMS.
• One site has an existing Pelco DX8100 that the customer wants to keep.
• Wireless links with encoder/decoder pairs are specified for certain locations.
• Specification deviations can be offered but they must be made 'line for line' with clear notation about any deficiencies relative to the specification

Our observations and concerns:

• Resolution Choice: As a general rule, the city is using MP outdoors and SD IP indoors. This seems to be a good rule of thumb as the outdoor cameras tend to cover larger areas, needing greater resolution. The fire departments are using all MP which may be because of larger indoor areas.
• VMS Strange 'Equivalents': The Ganz VMS is listed as an equivalent to Aimetis. First, it's pretty clear that the Ganz VMS is an OEM of Luxriot. Why they do not simply spec Luxriot is strange. Secondly, Luxriot has a lot less functionalities than even Aimetis standard (and a significantly lower cost). Since their is no functional specification, it's hard to know what the customer actually values. However, comparing these 'equivalents' will be difficult given the gap in functionalities and price.
• Vague Access/Video Integration: The specification is not clear about how the integration should be accomplished. It appears that a hardwire connection between a door and camera can be used, perhaps to trigger an alarm input to the camera. The problem with such 'integrations' is that finding video associated with access control events will be problematic - especially knowing what card was read and what happened (was access denied, accepted, was the door forced open, etc.). Worse, Luxiot/Ganz VMS has limited to no software based access control integration. Aimetis does have software based access integration but only in the higher Professional version that was not specified. Also, the City's access control system (Synergistics Presidio) is not listed in Aimetis's 3rd party system integration list.
• Enterprise Management Issues: With so many cameras at so many different sites, a single sign on to each server will be key in managing/controlling access. Unfortunately, the Ganz/Luxriot VMS does not support this and Aimetis would require the Professional version to accomplish this.
• Old OEM'ed 8MP Camera: The City is specifying an OEM of a legacy Arecont Vision 8MP multi-imager camera. This camera is MJPEG only, is likely to consume 30+ Mb/s outdoors and will use a significant amount of storage (that might cost the city hundreds of dollars more per camera). Furthermore, this camera is not day/night which is going to cause problems at night. Plus, Arecont recently announced a far better 'next generation' version that is smaller, cheaper and better.
• Specifying 'Deviation' Problems: Since there are no functional specifications, it is going to be hard to approve deviations. Deviations need to show deficiencies but not to a 'real' spec but to specific products already included.
• Encoder/Decoder Over Wireless: Often large projects will use an encoder/decoder pair to transmit legacy analog video over an IP wireless link. We think this is a bad approach. First, it's expensive as the encoder/decoder pair will be costly (over $1,000 is common, especially specifying Axis gear). Secondly, quality reduction is significant because the video is encoded, decoded and then re-encoded into a DVR. Third, if PTZs are used with this setup, latency problems can become significant. In such scenarios, savings are likely to be minimal and performance is certainly better by replacing the old analog cameras.
• Use of Analytics: Perhaps Aimetis's key differentiator are their built-in video analytics. However, their video analytics are only available in the Enterprise license version. In this project, it is only specified for use with a small number of PTZs. The biggest challenge is the lack of specification as to how the analytics will work and what performance the customer wants. Without this, bidders may underbid, be unprepared or unable to deliver the expected performance. Worse, since this specification is silent on the matter, the City may not have grounds to hold the bidder to the expected but unstated performance level.
• Designed by Reps: Reading this document, I get the feeling that the bid was 'designed' by local manufacturer representatives. Selecting these brands and specifying in such a vague manner makes little sense and leaves significant holes.