Server Based Video Analytics Financial PerformanceBy John Honovich, Published on Aug 02, 2009
In a video analytics survey article [link no longer available], an IMS analyst states, "Everyone thought edge [solutions] would be a winner. Server might be stronger than thought."
Putting the edge vs server technology debate to the side, the logistical benefits of server have always been obvious - it's cheaper to upgrade your DVR or add a server than to migrate to smart IP cameras. This is especially important in a recession. Last year, we examined those advantages of centralized analytics.
The other obvious factor is the simple lack of support from the Top 10 camera manufacturers. Only Bosch offers broad support for video analytics (I don't include Sony and Panasonic whose technology is clearly closer to VMD than analytics). Unless the large manufacturers are willing to invest in, release and promote smart cameras, it's almost mathematically impossible for edge to dominate.
This does not mean I am opposed to edge based analytics. I think there are many advantages for them. However, as a practical matter, product options in today's market make it unlikely for edge to dominate.
The next big trend to look for in server based video analytics is the use of GPUs. Though there are no products currently in production, many companies globally are evaluating such use. I do not know enough to have an opinion on future success but it is highly likely that in the next few years, it emerges, at least, as a marketing trend.