It seems in the technology industry we’re always on the lookout for the “killer application” to take something that’s new and make it mainstream. The best example of this was, of course, Lotus 1-2-3 and its role in the Windows interface becoming the de facto standard it is today. After that, the floodgates opened and Windows took off like a rocket.
Not everything has a “killer app” associated with it, though. Sometimes the industry just needs a handful of “deadly” applications to drive adoption. Think of the deadly apps as being apps that have killerish qualities but which don’t have the breadth to be universally deadly.
For network virtualization, I don’t believe there’s a single killer app that will drive adoption. Instead we should see a number of use cases across different verticals that can be considered killerish. One of these use cases is IP video surveillance.
Video surveillance has evolved greatly over the past few years. What once was a proprietary technology that was deployed on a dedicated network has shifted to IP based endpoints that run on the corporate data network. Also, the responsibility for video surveillance and other physical security has shifted to the CIO as organizations try and gather greater intelligence across the entire security infrastructure. It’s a technology that was only about $4 billion in revenue in 2012, and which grew to almost $6 billion in 2013; in fact, ZK Research forecasts project growth to over $19 billion in 2017.
Obviously the network plays a critical role in IP video surveillance, since any sort of dropped packets, latency, or jitter can create holes in coverage which then can limit the usefulness of the technology. ZK Research recently ran an IP video surveillance survey to understand the role the network plays. One of the questions we asked was, “What are your primary challenges associated with IP video surveillance operations?” The biggest challenge mentioned in responses was delay in pulling up video footage (33%). Blurriness of video accounted for 19% and network resiliency was another 16%. In fact, 75% of the challenges are network related. Only 22% of the respondents cited camera or non-network related issues.
I think it’s fair to say that traditional networks were not really designed for the demands of IP video surveillance. Legacy networks are hampered with many limitations such as poor performance, complexity of running multiple protocols, slow recovery time, and security challenges associated with using VLANs for segmenting traffic.
Network virtualization can alleviate many of these problems through the following:
- Simplification of the network. Network virtualization leverages an underlying network fabric that is typically more flat and much more simple than a traditional multi-tier network. This means fewer boxes, fewer touch points, and a reduced need to continually tweak the network. When it comes to a demanding application like IP video surveillance, simpler is better.
- Minimized number of network protocols. Legacy networks often require numerous protocols to operate, including. but not limited to, Multi-Protocol Label Switching (MPLS), PIM, STP, Border Gateway Protocol (BGP), and Open Shortest Path First (OSPF). Finding a way to eliminate as many of these as possible will ensure quick network recovery and efficient bandwidth use. One way to meet the challenges of IP video surveillance is to leverage a protocol like Shortest Path Bridging (SPB), which is widely used by many of the network virtualization vendors today. Traditional networks are built on a concept of “stacked protocols”, where each layer in the stack was developed to add new functionality to the existing Ethernet standards. However, each of these protocols runs independently and must be managed separately, causing network instability. SPB is an industry-standard protocol that can consolidate all of these functions and create a more efficient network.
- Higher levels of security. MPLS and VLANs are OK if the goal is simply to separate traffic for performance purposes. However, IP video surveillance requires high levels of security, which is why the technology has been deployed on a private, closed network. Network virtualization can give the performance benefits of MPLS and VLANs but also the security of a dedicated network.
- Predictable performance. Multicast sessions are very CPU intensive. The rapid initiation of video streams can create CPU spikes and cause erratic performance on the network. Networks that support IP video surveillance must eliminate these spikes to ensure all corporate applications perform consistently. A virtual network can protect the IP video streams from general corporate traffic and vice versa.
Better camera technology coupled with a greater awareness of the value and need for IP video surveillance has caused the technology to grow by double digits over the past few years. I certainly expect this to continue for at least five more years, offering providers an excellent opportunity for vendors to try grab share.