About edge computing: Is it right for your business?

Like most emerging IT trends, "edge computing" is not a revolution, but more of an evolution. Edge computing has its roots in content delivery and peer-to-peer networks and grid computing at the beginning of this century. However, the continuous improvement of network, computing and analytical technology capabilities and the need to handle massive data growth mean that computing will become more and more important to IT managers in enterprise organizations today.

With the expected amount of data being sent to and growing through the network, organizations are developing computing capabilities closer to the edge of the network, where it is generated. The disruption created by edge computing allows local users to generate and perform data analysis in real time. However, despite the growing momentum behind this development trend, the industry remains uncertain about how and when to deploy edge computing. Will edge technology drive industry business decisions in the next few years, or will broader deployment require further advancement and comprehensive long-term planning; which use cases will actually drive the deployment of edge computing, and what challenges still exist; and so on. A series of questions need to be answered.

[[212389]]

What is edge computing?

For its definition, the broad concept of edge computing appears to have originated nearly 20 years ago, when edge servers were a term referring to servers in a content delivery network (CDN). It has more recently emerged in the context of processing, analyzing, and applying data generated by data sources at the edge of the network, rather than transmitting the data to "core" processing units.

The term “edge” is based on processing proximity to data sources and where analytics occur, so it covers a range of possible uses. Vast amounts of data from Internet of Things (IoT) devices, mobile devices, and networks, each with different processing requirements and priorities, will continue to be generated. This shift will emphasize the importance of where computing occurs, requiring smaller, more flexible processing units closer to users. This may take the form of distributed micro data centers with local network access and interconnection points, essentially forming a distributed cloud service.

What pushes us to the edge?

The force behind the upcoming growth in data transmission is the result of the increase in computer-to-computer communications and IoT devices. As a result, a variety of sensors and processors can create and transmit large amounts of data information. At the same time, the market is also increasing investment and development in artificial intelligence and augmented reality, drones, and automated transportation systems - all of which are driving the development of edge computing.

In 2015, the Cisco Cloud Index found that approximately 90% of the data created globally had been generated in the previous two years. The index also predicts that monthly Internet Protocol (IP) traffic will continue to grow at a compound annual growth rate (CAGR) of more than 100% from 2014 to 2019. At the same time, the number of Internet users is expected to grow at a CAGR of 7% per year, and the number of connected devices will grow even faster, at a CAGR of 11.4%.

Over the same five-year time span, video data transmission is expected to grow by 80%. To put these numbers into perspective, as of 2016, there were approximately 3.4 billion people active on the Internet. YouTube users download 400 hours of new video every day, Instagram users like 2.5 million photos every minute, and Facebook users share 3 million posts every minute and like more than 4 million posts every minute. Additionally, approximately 4 million Google searches are conducted every minute, more than 200 million emails are sent every minute, more than 400,000 apps are downloaded from the Apple App Store, and 277,000 tweets are sent every minute. In financial terms, Amazon sells approximately $80,000 worth of goods and services every minute of every day!

The combination of consumer and technology factors creates a more complex set of drivers that vary widely across industries and geographies. In general, the three major drivers that are driving us toward edge networks are as follows:

• Changing consumer and business expectations, and data usage;
• Developments in emerging technologies, particularly in the areas of networking, processing, software, and protocols, make edge computing possible;
• Applications of edge processing, such as the opportunity to integrate IoT device data, provide higher network processing and transmission efficiency, and lower latency, thereby providing better customer experience and data security.

Edge processing in enterprise business

As with any other IT trend, business and technology professionals want to know how these new technological advancements will help them optimize operations. However, edge computing may not be right for all businesses, at least not yet. The adoption of edge computing and applications will ultimately depend on how well these technologies are aligned with business goals and whether the enterprise has the resources to effectively implement, manage and monetize them.

There are several industries that will particularly benefit from edge computing:

• Smart Cities. Edge computing can be widely used in smart communities or cities. With the increasing number of sensors and information sources (transportation systems, medical systems, public service utilities, and security programs), it has become less feasible to store and analyze data in a central location. Edge computing can also reduce delays in community services, such as processing for medical emergencies, law enforcement, traffic patterns, and public transportation. It also takes into account geographic precision, so that information related to a specific street, block, or suburb can be shared instantly with users in the area. These applications and technologies will ultimately determine whether the edge extends from transportation system sensors and street lights to pumps, turbines, and other traditionally unconnected utility devices in data centers. How will smart city edge networks collect and distribute information in the event of a natural disaster? How can smart cities and edge technologies be used to deliver and mitigate the impact on the supply chain of resources such as water and gasoline?
• Smart Commercial and Public Transportation. Edge computing already performs many functions for the commercial and public transportation industries. For complex aircraft such as airplanes, ships and spacecraft, the need for accelerated processing and edge computing/analysis at the edge means that only the most important information is transmitted for further analysis; most of it is stored locally. Edge technology allows traffic and environmental sensors to process and provide the most relevant information to vehicles, including self-driving cars. The second major function of edge networks is their informative nature: providing local data patterns to the wider network system to improve the efficiency and safety of transportation. Intelligent transportation systems are also a natural part of the development of smart cities.
• Smart Home. Some data center OEMs claim that every home in the U.S. will soon be its own data center, and this claim is gradually becoming a reality. However, edge computing connects smart home systems to core production centers, rather than creating independent data centers as data moves to the edge. Therefore, the role of "batch sending" and real-time connected devices in the smart home will continue to be discussed and developed.
• Self-driving cars/planes and remote-controlled machinery. Perhaps the most notable example of edge technology is the estimated 200+ CPUs required by autonomous vehicles, which are “data centers on wheels,” according to Peter Levine of venture capital firm Andreessen Horowitz. Autonomous vehicles can process live video and streaming photos, making immediate decisions based on data inputs in real time. They emphasize the need to share collaborative information through smart transportation networks. This concept can also be extended to drones in industries such as agriculture, mining, and oil and gas, which must react and process collected data in real time.
• Media and other content. CDNs already bring content closer to users, and edge computing is the logical next step to provide users with additional operational applications. They will also be involved in future content delivery, enabling providers to expand geographic coverage and maximize the efficiency of delivery networks, especially as more value-added and interactive services are introduced.
• Manufacturing and Industry 4.0. Robotics, artificial intelligence, and machine learning have already been adopted by many industrial companies, all of which are optimal use cases for edge computing. The guiding principle of edge computing in manufacturing is to simplify production into a standard process from demand to production, delivery, and consumption. This aspect will require the exact type of collaboration between data sources in various locations that edge computing provides. The Industrial Internet of Things (IIoT) continues to improve efficiency and reduce costs through predictive maintenance, improved safety, and other operational efficiencies. The edge is essentially the data supply chain of Industry 4.0.

Develop an edge strategy

Some companies have adopted an "all-in-one" approach to the Internet of Things and the edge, although they should be mindful that some industries that have been slow to adopt are still busy with cloud computing. The trend from centralized and cloud computing to distributed edge computing requires careful consideration to determine the best balance between convergence and disaggregation for each use case.

For enterprises that are adopting edge computing, developing a comprehensive strategy will include five main components:

1. Identify goals and requirements, including business objectives/drivers and brand, customer and return on investment requirements.
2. Map the network topology from the edge back to the core.
3. Define the systems, protocols, and procedures that make up edge processing, abstraction, and communication capabilities.
4. Define the network that links edge processing units to data sources and back to core processing facilities.
5. Develop monitoring, maintenance, and security strategies for edge computing systems.

Notes on edge operations

Establishing edge computing capabilities in an enterprise is only the first step in implementing this technology. Long-term success will depend on the comprehensive development of operational operations, security and maintenance requirements, as well as integration with IT and communications providers. As the demand for edge computing continues to grow, the roles of all parties must be clearly defined and data must be prioritized in the network. Despite the obvious benefits of edge computing, there are still some factors that should be considered.

Security and privacy are the main challenges that need to be addressed. How data is generated and processed, as well as ownership of data, must be defined. Concepts of criminal or civil liability may become complex. For example, who will be held responsible when a self-driving car crashes due to programming flaws? What impact will data protection requirements have? What about edge innovation, imitation or regulation?

Analysts also wonder when the growing number of data-producing devices will create a network that reaches a tipping point and overloads. Software-defined networks, 5G technology, and edge-to-edge communications will continue to evolve. Interference and bandwidth demands in the radio environment will place demanding demands on transmission networks. Traditional macro networks will need to be supplemented by adding small cells.

Resilience must also be considered when discussing edge computing. It raises a variety of questions: What is the impact of a unit, source, or unit failure in an edge processing unit? How will it affect connected units and shared programs? How can systems that rely entirely on fluid networks protect themselves from cyberattacks? How do capacity and latency coexist, and how should networks be prioritized? The answer seems to be that resiliency will depend on what is being delivered, where it is regulated, the end-user experience, and cost/revenue metrics.

As edge computing continues to grow, the shared languages ​​that support multiple edge systems, multiple platforms, and runtimes will grow and need to improve. The protocols that make systems and networks work together will also need to evolve. Industry leaders are already developing IoT gateways and routers that can support edge computing, and software like Apache Spark is responding to the evolution of the industry's needs by clustering and running various patterns, writing to disk when needed. The combination of these inflection points and those small development milestones will produce significant changes.

Understanding and preparing for edge computing will require consideration of many aspects, including mobility and security. Holistic planning can help ensure that requirements are met for a system that can identify and communicate with edge devices and users.