Docs

Introduction


Welcome to Horizon

The Horizon Project operates on key driving principles:

  • Edge data is growing faster than bandwidth to central points (Cloud)
  • Edge data is perishable and loses value over time
  • Real-time Edge analytics and cognitive models are required for local (autonomous) policy-based decisions
  • Edge analytics and cognitive models need to change as central analysis identifies gaps, optimizations, and new opportunities
  • Everything is an Edge: from baremetal machines in cloud data centers to single-board computers (SBCs) all may operate on Horizon

“It's estimated, that by 2020, the average Internet user will generate about 1.5GB of traffic per day—as compared with the smart hospital, which will generate 3,000GB, the autonomous car (4,000GB per day each), airplanes (40,000GB per day), and the smart factory (1,000,000GB per day). And that's all data that needs to be processed and analyzed.” -- Intel CEO Brian Krzanich, Intel Vision for a Fully-Connected World [1]

Connected devices already outnumber humans on planet earth. As connected devices become more ubiquitous and more capable, many human-machine interactions will be replaced by multiple machines autonomously interacting with their environments and with each other, completely transforming the human experience.

                                           Each inflection point in the history of computing has triggered an explosion in the number of computing devices. [2]

"Each inflection point in the history of computing has triggered an explosion in the number of computing devices.... Fueled by ubiquitous connectivity and the availability of billions of IP addresses with IPv6, the number of connected devices is forecasted to surpass 25 billion in 2020, up from 2.5 billion in 2009 and 10 billion today." Intel Vision for a Fully-Connected World, Intel CEO Brian Krzanich [1]

Real-time machine decisions require highly responsive reliable compute at the edge, not just low-latency high-bandwidth connectivity.  Horizon's Core Technologies form a compute system that enables autonomous machine-to-machine interaction. Horizon is the basis for extending Cloud to the Edge and provides a system to drive Insights at the source of the data. 

 

Horizon enables a "Programmable Edge" for Analytics and Cognitive Applications

Horizon enables a "programmable edge", and the ability to gain insight from data using analytics and cognitive applications.  Data filters, container microservices, and Machine learning models can be pushed to the edge to run compute at the source of the data, and stream high value data back to the cloud at lower bandwidth.

Horizon creates a decentralized, autonomous, edge compute fabric that enables any Edge (connected device) to directly connect to any other Edge to facilitate the processing (filtering, analytics, etc…) of streaming data -- beginning right at the source.  Processing the data at its source creates higher value data streams. That higher value data can then be used locally for immediate and autonomous actions with sub-millisecond response times (hundreds or thousands of times faster than cloud-based solutions).

Horizon can enhance traditional cloud applications where remote devices centrally aggregate and analyze Edge data or remotely initiate Edge actions.  By migrating these solutions to Horizon, vendors can gain access to an open ecosystem, and can add features to cooperate with devices from other vendors.  By moving to Horizon, vendors can also automate their firmware upgrade procedures, and make their devices more responsive in the field.

Horizon is designed to encourage cross-vendor collaboration in an open, decentralized ecosystem.  Horizon facilitates automated, vendor independent machine-machine decision-making and cooperation. Horizon welcomes new devices, provides introductions to other devices, and when desired it can help both parties to leverage the mutual trust of a third party to safely connect and cooperate.

 

Sources: 

1  Intel Vision for a Fully-Connected World, Intel CEO Brian Krzanich

2  Image and quote, Device Democracy, IBM Institute for Business Value