May 2, 2018 | Edward Layeux
As we look up to the evening sky and across the map of stars we can only imagine what lies beyond. The universe as we know it has no physical boundaries or edges to gaze out from, but our imagination allows us to view the possibilities. Similarly, modern technology exists without boundaries, allowing us to inter-connect globally and immerse ourselves in realities that ignite our imagination. Unlike the universe, perceived edges exist along our global technology networks and are poised to power our world like never before. This is known as Edge Computing.
Edge Computing refers to a distributed and decentralized architecture where data analytics and processing are computed in real-time closer to source origins or on-device. This decentralized paradigm emulates cloud services but is composed of multiple gateway servers and sensors positioned across network infrastructures and Internet of Thing (IoT) devices. These micro endpoints harness, store, and process data, creating conditions to make near instant decisions. This real time landscape allows for microsecond data analytics and processing with little to no latency and seamless integration with the cloud. Moreover, security and privacy are enhanced as identifiable data is retained locally at the edge instead of transmitting over networks.
Time Value of Data
The drive to the edge is driven by the need for speed required by the latest mobile and next generation technologies where split-second timing is crucial. Artificial intelligence (AI), immersive experiences (AR/VR/MR), conversational user interfaces (CUI), and emerging IoT smart technologies are at the forefront of the drive to the edge. These technologies are re-imagining and transforming all sectors of our society and industries and are only expected to proliferate. From autonomous vehicles to smart cities, the realization of the time value of data is critical in decision-making and reaction. Data loses value when it cannot be processed fast enough. For example, autonomous self-driving vehicles of the near future will need to make split second data-driven decisions in real-time and even a delayed response of a few milliseconds could have severe consequences.
Immersive experiences like virtual and augmented reality realize the time value of data. These experiences require meaningful insights and real-time data-driven reactions. They rely on our neurological processes in the brain to coordinate eye and head movements that occur within milliseconds of each other. Any amount of latency diminishes the immersive experience and renders it as less than life-like.
The complexity of machine learning and conversational interfaces are enhanced with the time value of data in edge computing leading to faster and more accurate real time analytics and computations. For instance, conversational interfaces are more responsive interpreting voice instructions computed locally on-device as opposed to transmitting data back and forth over networks.
The volume of information IoT devices, sensors, and next generation technologies could generate will be massive. Everything from machine learning algorithms to industrial sensors, drones, robots, streaming video creates an explosive quantity of data. Transmitting all of it over networks to centralized cloud servers is too slow and intensive for existing networks architectures.
For autonomous vehicles alone, it is estimated that the amount of data flowing between cars and the cloud could reach 10 exabytes per month by 2025, approximately 10,000 times larger than present volumes. Moreover, IDC research estimates that by 2020, the amount of digital data generated is predicted to reach 44 zettabytes, unleashing a new wave of challenges. Further, in the coming decade it is estimated that there will be approximately 100 trillion sensors added to our global economy, generating an unfathomable amount of data.
Transformation to the Edge
The transformation to the edge creates infinite possibilities and outcomes. Its impact is far-reaching across sectors like Manufacturing, Energy, Oil and Gas, Telecommunications, Chemical, Health Care, Agriculture, Finance, and Advertising among others.
Future commerce practitioners will gain unprecedented insights into consumer behavior with greater, near-instant analytics, improving online and retail experiences. For example, edge devices can be used in physical retail to monitor and harness information from customer mobile devices and online shopping experiences to tailor merchandise closer to their preferences as they browse through stores.
Industrial Internet of Things
Edge Computing is a major enabler of IIoT and a gateway to Smart Manufacturing. It provides manufacturers with the ability to process data more effectively, with deeper insights, leading to faster and more intelligent actions, increased productivity, higher cost savings, and conservation of network resources.
In automated robotic assembly lines, edge computing helps facilitate production by harnessing and analyzing data from connected machinery and devices to aid performance and preventive maintenance.
Remote industrial machinery such as oil rigs, wind turbines, mining equipment, magnetic resonance scanners, pipelines, power grids, traffic lights, and undersea blowout preventers generate massive amount of continuous data. These remote machines rely on time sensitive data for decision-making but are restricted due to proximity to centralized cloud servers or limited internet connectivity.
For instance, remote wind turbines generate vast amounts of sensor data, and if unexpected failures or disasters like fire occur, companies need real time analytics and reaction. They cannot afford to justify any amount of network latency or unavailability.
Want to learn more about IIoT? Check out our blog post, The Industrial Internet of Things: Let's Talk IIoT.
The drive to edge computing coincides with the explosion of IoT data and the emergence of 5G mobile technology. For example, AT&T have invested in building edge computing networks based on 5G with the goal of providing single-digit millisecond latency to power 5G applications.
Additionally, several organizations have transformed to edge networks:
- Dropbox global edge computing infrastructure stores and serves approximately 90% of their user’s data delivering much higher levels of performance.
- General Electric provides edge-to-cloud solutions designed to make data analytics more powerful.
- Dell is transforming infrastructure at the edge with its micro Modular Data Center (MDC).
Microsoft and Amazon are also extending software solutions out to the edge including:
- Microsoft’s Azure IoT Edge service enables edge devices to run cloud services and analytics in near real-time and communicate with sensors and other devices connected across networks, even with intermittent cloud connectivity.
- Amazon’s AWS Greengrass software similarly powers local edge devices and gateways running on local networks and is delivered through AWS Lambda.
Gartner estimated that by 2022, half of all enterprise data will be generated from the edge.
Back to the Future
Our world is changing fast and furiously, and it seems that the present, past, and future are interwoven. It was not that long ago when centralized mainframes ruled technology, until a more decentralized client server model re-shaped the landscape, only to make way for modern centralized cloud platforms. Now, the promise of the edge is pushing the barriers back to decentralized networks to meet the demands of tomorrow.
So, fasten your seatbelts and hold on because things are about to get a lot faster.
If you’d like to discuss these topics further, SMITH can help.
Whether you’re exploring an evolution in your approach to selling by moving to pure ecommerce or a hybrid approach, looking for an edge in operations optimization with augmented reality, or ready to deploy intelligent replenishment for your customers, we can help.