How can you run your business better if you don’t actually know what’s going on? Traditional video surveillance is primarily a security application and very reactive – when something bad has happened, you spend hours reviewing the footage and then take action. With Big Data technologies, there is an opportunity to move beyond security to monitoring and analyzing the business in order to better understand customer behavior, optimize operations, manage risk, and innovate.
Hector Martinez, Business Development Lead for Defense, Safety, and Security at EMC discusses the state of the video surveillance industry and how Big Data has the potential to transform the industry from just a security application to a business recommendation engine.
1. What are the top use cases for Video Surveillance today?
The most common use case we see across all industries is physical security – protection of facilities, theft, and safety. Beyond that we are seeing specific industries with requirements for long term archival to deploy analysis-based applications. This is where Big Data becomes critical in supporting these modern applications for video surveillance.
2. What are the key challenges with Video Surveillance solutions today?
The vendors of Video Surveillance solutions have traditionally been tape-based systems, lacking modern back end technologies such as scale-out storage. As a result, there is a short shelf life for this rich video content, serving a single purpose of physical security. To add to this, physical security solutions are typically managed by the Security Department having little interaction with IT to drive better decisions around the cost and use of these valuable assets.
3. Explain how Big Data can overcome these challenges?
The good news is that heightened security concerns around the world are causing organizations to expand their use of video surveillance and other physical security technologies, forcing Security Departments and IT to converge and innovate. And Big Data storage, Hadoop , MapReduce are driving these conversions to do more with video surveillance. For example, we are working with a large retailer in Australia that is forward thinking and is looking to replace their taped based Video Surveillance for Loss Prevention with a Big Data platform that centralizes video content across all stores. Through new features such as real time access, instant event search, and archiving, other departments have visibility to what is going across retail stores. Operations can uncover theft for more efficiently, Marketing can also optimize in-store promotions, and HR can better understand organizational behavior.
4. What role does EMC play in modernizing video surveillance systems with Big Data?
EMC is working with video surveillance vendors to provide seamless integration of Big Data storage with their systems. For example, EMC was one of the first vendors to certify purpose built high performance storage with Genetec’s Security Center, a unified IP based security platform. The EMC Reference Architecture with Genetec below ensures optimal configuration and tuning, enabling users to view real time video streams while also receiving policy based and anomaly based alerts. Although you can use EMC VNX and EMC Isilon storage for high-throughput solutions, alternate storage solutions include EMC Symmetrix VMAX.
In addition to Big Data storage, Big Data analytics plays a key role in modernizing video surveillance systems. EMC Greenplum provides the analytics platform to discover anomalies and take action. Customers can take thousands of hours of past video recordings from their video surveillance systems and throw it into Greenplum HD (Hadoop) to make predictions of future threats or issues. In fact, we have found that the EMC Big Data solution – Greenplum and Isilon provide the lowest cost solution over other Big Data storage and analytic solutions.
5. Yes, I recently met with Victor Fang on the Greenplum Data Science team on this large-scale video analytics topic. His team is working on a large-scale video/image analytics platform that leverages the Greenplum product stack – MPP Database and Hadoop. With this platform, a suite of standard and 3rd party computer vision algorithms can be integrated and scale up for analyzing Big Video/Image Data. Applicable use cases include anomaly detection (sterile zone, camera tempering, etc), trajectory analysis (traffic behaviors on highway or parking lot), and video indexing, etc.
Yes, I also have been speaking with customers that confirm simple Hadoop programs with MapReduce enhance their video surveillance solutions by decomposing a large computation into a set of smaller parallelizable computations for situational intelligence or surveillance video searches.
6. Dubai International Airport is an EMC Video Surveillance customer. Where can I learn more about this story and other customer successes?
We have a comprehensive set of resources on emc.com to get started and learn about what other customers are doing with EMC Video Surveillance solutions.