There's More Than One Digital Insurance Infrastructure

In a recent post, technology journalist Arthur Cole posed an interesting question for insurers: is today’s infrastructure ready for the telematics revolution?

The vision of telematics – and indeed, the digital insurance enterprise that is emerging – is a compelling one, and necessary to compete in today’s and tomorrow’s tech-driven economy. But will insurers’ back-end systems be able to handle it?

Telematics, for one, is a big data problem, and it takes all the essentials of big data to make it all work: processing power, storage, real-time networks, and analytics solutions, just to name a few. One approach, mentioned by Cole, is the cloud, which provides many of these capabilities on demand.

However, seeing value from the Internet of Things, which is what telematics represents, also requires coordinated management strategies, as it won’t be just one system involved, but entire networks of systems, sensors and capabilities.

Again, cloud is a good thing, but the road to digital may require various gradations of cloud. For example, data centers may need to evolve from large, centralized facilities, as are maintained by many insurers these days, but rather, to smaller, “micro” data centers that are more at the edge of things, closer to the devices and systems of which they are part as the IoT gains traction. These edge data centers may form a $6.3-billion market by 2020, according to a report from MarketsandMarkets, as reported in Data Center Dynamics.

Taking things a step further, some are referring to this distribution of infrastructure as “fog computing.” As mentioned in a post last month, an “OpenFog” architecture enables processing power can be located at any point between endpoints and central clouds. There is no “right” place for processing – it can be within small servers at the edges, or within a centralized cloud service.

As Cole explains in his piece on the telematics infrastructure challenge, all possible resources must come into play to enable development of a “top-to-bottom architecture that provides easy integration with leading telematics devices and original equipment manufacturer solutions, as well as real-time data processing and error-handling,” along with “efficient data storage and management” to the tune of at least one gigabyte per vehicle per month.

Existing data centers won’t be able to handle this alone, and cloud services may not be enough to cut it. Ultimately, a multi-layer hybrid approach needs to be considered.