Author: Hu Yoshida, Chief Technology Officer, Hitachi Vantara
One of the breakout presentations at Hitachi’s Next 2017 event in Las Vegas recently was on Fog Computing.
Sudhanshu Gaur, Director, Digital Solution Platform Lab Hitachi America started the presentation by describing how an octopus has a distributed brain. Apparently an Octopus has separate brains in each tentacle besides the brain in his head. If the head brain wants to eat something, it has to send a message to the appropriate tentacle to seize the food and bring it to its mouth. If a tentacle is severed it continues to react to stimuli for some time after.
In addition, the suckers on the tentacles have some ability to recognize its own skin so that it does not get stuck on itself. This was used as an example of fog computing where intelligence is distributed between the cloud and the edge device.
Fog computing, also known as fog networking or fogging, has been defined as a decentralized computing infrastructure in which data, compute, storage and applications are distributed in the most logical, efficient place between the data source and the cloud. Fog computing essentially extends cloud computing and services to the edge of the network, bringing the advantages and power of the cloud closer to where data is created and acted upon.
A good example of fog computing is an autonomous car. Some compute and storage resides in the edge device, the car, where it needs to make instant decisions about steering, braking and accelerating. The car may be connected to a traffic controller that manages the traffic on a roadway system, and it may be connected to a system in the cloud that monitors usage, predicts remaining driving range, and prescribes maintenance.
Peter Levine, a general partner at the venture capital firm Andreessen Horowitz, would take exception to the fog definition that fog computing extends cloud computing and services to the edge of the network. He is being widely quoted in the business media as saying that edge computing will displace the cloud. He points to edge devices like self-driving cars and drones which are really data centers that need to make immediate decisions using AI and machine learning and cannot wait for cloud computing decisions. Peter believes that his job as an investor is to recognize where the industry is headed before it happens and he sees this as a major new trend. You can see his 25 minute video at this site
I have been in this business since the mainframe days, and I have seen the pendulum swing from centralized mainframes to distributed open systems and then back to a centralized cloud; so I am not surprised to see it swing once more to a distributed edge or fog. IoT will drive compute to the edge and points in between, where it is most logical and efficient to drive business outcomes. Whenever we see these swings it is accompanied by major disruption in the industry and many technology companies that were leaders in their day, are no longer in existence - Digital, Amdahl, and yes even EMC.
Cloud, like mainframes and distributed open systems will still have a place in the computing ecosystem, but the role of the cloud will change. IoT, devices on the edge will act and react with other devices in real-time aided by AI and machine learning. Cloud will do the heavy analytics work, facilitate communication between systems, provide common repositories and data governance, work that would be too heavy for the edge at this time.
IoT is about edge computing and Hitachi is pivoting to address this with the formation of our new company, Hitachi Vantara, which integrates Hitachi Data Systems Infrastructure and cloud, Hitachi Insight Lumada IoT platform, and Pentaho analytics. Hitachi Vantara is a new company, purpose built, to address the IoT market and all points in between.