The Distributed Cloud, Black Boxes, and Airline Disasters

The tragedy surrounding Malaysian Air flight 370 brings renewed urgency to a concept described at HP’s recent Industry Analyst Summit in Boston. Martin Fink, an HP Executive Vice President and the company’s Chief Technology Officer, presented on the topic of “How the Cloud Will Evolve.” Fink sees what he calls a “distributed mesh cloud” developing, and he expects that the cloud of the future will have multiple functions: translator, coordinator, orchestrator, arbitrator, aggregator, replicator, anonymizer, border guard, and learning engine (his favorite – more on this below). Yet challenges that the cloud faces include the sheer magnitude of data, legacy databases, hardware constraints, and time-consuming data integration. In addition, the need for real-time insight is hampered by insufficient resources as well as limited bandwidth, networking, and computing power. The lack of boundaries is another cloud challenge. Departments, companies, geographic areas, sovereign states, and others are concerned about privacy as well as defending themselves from attacks that might come from parts of the cloud.

Fink used an airline example to highlight this. What if the cloud acted as a learning engine, centralizing data from a mesh of network-connected aircraft? This type of central learning cloud could allow an airliner in flight to transmit information about an issue it is having with weather, mechanical, or some other problem. This could prove to be very valuable data for other planes in flight, yet there are legitimate concerns about who could access it. Are you willing to allow other entities (including competitive airlines) to see your flight data? Data privacy aside, in an emergency situation the ability to access data could have immediate and potentially life-saving consequences. Why should that information only be accessible through the black box on the plane? In regard to Malaysian Air flight 370, it is amazing that in this day and age, days might pass before the aircraft is located.

Of course, this type of use of the cloud has huge data storage, transmission, power consumption, networking, and computing requirements. To make the distributed cloud work, Fink stressed that you would need better networks and storage (particularly ones that operate using a minimum of energy), which is why HP says it is so focused on technologies like photonics, memristors, and Moonshot (HP’s software-defined server). Perhaps in the future innovative use of the cloud could mean that we won’t be so dependent on a black box that goes down and is inaccessible after an airline disaster.

Of more direct impact to graphic arts users is how the cloud can be leveraged for all types of communications. For more information on that topic, see InfoTrends’ recent study entitled “Cloud-based Workflows & Infrastructures for Graphic Communications.”