There are a wide range of environments where any breakdown of the surveillance system can lead to substantial financial losses or significant personal risks: prisons, central banks, and casinos, to name but a few. Here the ideal situation is obviously to have high availability systems where there is no chance that functionality or picture data will be lost, and where the surveillance system continues to operate completely unaffected even if there are several simultaneous component failures. But is this in fact a realistic proposition, or just pie in the sky?
Achieving this level of resilience and reliability presupposes a system designed without any 'single points of failure' which could be responsible for influencing the function of the whole system. This requires not only the duplication of critical elements and the intelligent mutual monitoring of all the installation's components and functions, but also a system of automatic hand-over from failed components to those held in reserve. The catch of course is that as long as this necessitates the incorporation of lots of redundant hardware, then high availability comes at a very high price, both in terms of investment and running costs. But things are changing, fast.
The arrival on the market of Microsoft's Windows Server 2008 has provided a new tool for IP video system developers - an integral virtualization solution. This enables the generation of multiple virtual machines (i.e. virtual software-only computers) within a single piece of hardware and allows IP video system manufacturers to create redundancy within their solutions by duplicating virtual servers rather than hardware ones. It means, for example, that a hardware video management server can now be used for visualising both live and recorded pictures.
With multiple virtual machines operating simultaneously on physical server computers, high availability in the system is achieved using a so-called 'failover cluster.' This is a group of independent server computers (nodes) which work together to increase the availability of applications and services. They are linked to each other by both cables and by software. If one of the cluster nodes fails, then another node in the cluster takes over using a process called 'failover.' This ensures that the video management servers in a surveillance system are always available.
In the virtual world it turns out that less really is more. Besides achieving high availability without hardware duplication, virtualization offers other benefits. Virtual machines make much more efficient use of hardware; they take up less space and use less power than real ones, so it therefore follows that it is now possible to make devices which are much more compact and much more energy efficient. And, since units using less power produce less waste heat and thus need less cooling, complete system installations consume considerably less energy. Current evidence suggests as much as 30~40 % less.
At Geutebruck we have exploited this IT technology in building a new high availability, high end, IP video solution. In essence, we virtualised our own high-end GeViScope video management servers on clustered, modular server hardware and combined this with high availability level 6 RAID systems. The resulting modular server with high availability hardware is designed for 24/7 operation in digital surveillance systems and delivers high end server performance. We launched it at the CeBit exhibition in Germany in March 2009, and our first installation was only a matter of weeks later in a large casino in Switzerland. You can read more about our high availability solutions at Geutebrueck.