Image of wireless distributed computing network connecting multiple buildings

The new world of distributed computing from edge to cloud

Distributed computing isn’t a new term, but its definition has changed immensely over the years. Originally associated with the earliest days of the internet (see ARPANET), multiple architectures have evolved to include cloud-based services and infrastructure, the Internet of Things (IoT), and most recently, edge computing. Given the latest advances in edge computing and its ability to address new and business-critical enterprise use cases, it’s time for businesses to prepare for a distributed computing future.

Just as there are various cloud computing models, such as public, private, and hybrid, there are also multiple types of distributed computing, with edge computing among them. Edge computing, one of the more recently developed distributed compute architectures, is essentially the organized amalgamation of multiple disciplines including cloud computing, pervasive connectivity and mobility, the IoT and powerful analytics. Edge computing, powered by 5G and advanced Wi-Fi, provides resources such as compute intelligence, storage, and networking closer to where the data gets created, which enables a host of new services and applications that would not be possible if that data had to travel all the way to a data center.

Is Edge Computing Distributed Computing?

Effectively, edge computing moves traffic and services from a centralized cloud to the edge of the network and closer to the customer. Instead of sending all data to a centralized public cloud for processing, the network edge analyzes, processes, and stores the data. Collecting and processing data closer to the customer reduces latency and brings real-time performance to high-bandwidth applications. A distributed computing landscape showcases a world where data gets created, analyzed and acted upon anywhere from the end computing device to a locally operated data center.

Edge computing could mean processing data at the end-user computing device such as a smartphone or an automobile. Edge computing can also mean processing data at the edge of an enterprise’s premises, such as inside a factory. Additionally, edge computing describes a data center infrastructure managed or hosted very close to the enterprise by either communication service providers or cloud computing providers. This local data center could be a Multi-Access Edge Computing (MEC) network, using advanced wireless technologies such as 4G, 5G and the latest Wi-Fi to enable cloud computing and IT services at the edge of any network. A modern distributed computing definition includes all of these network options.

The new world of distributed computing provides:

  • Connected devices with high-performance processing. Today, most newly manufactured devices contain a processor, such as a Central Processing Unit (CPU), and the option for wireless connectivity. Device manufacturers also have access to a more comprehensive range of high-performance processing options such as Graphical Processing Units (GPUs), Vision Processing Units (VPUs) and Neural Processing units (NPUs). Additionally, many organizations are also retrofitting their existing equipment with sensors, processors and connectivity to enable real-time data gathering and analytics.

 

  • High-speed connectivity everywhere. While the market anxiously awaits the widespread availability of 5G, an edge computing strategy leverages multiple wireless technologies such as 4G, 5G, and advanced WLAN with 802.11e. It includes using various spectrum bands from 1-60Ghz. The availability of a broader range of spectrum choices provides enterprises and service providers with the opportunity to make appropriate network selections based on the required latency, speed, and the number of simultaneously connected devices.

 

  • The option of ultra-fast and ultra-reliable connections. For some use cases, high-speed connectivity isn’t enough. Organizations may also need low latency and ultra-high reliability (as close to 100% uptime as possible). For example, robotic control and video-driven machine-to-human interaction require both attributes in addition to high throughput.

 

  • Flexible, scalable service. Distributed computing is cloud-aware, virtualized and programmable to support modern application development. It’s software-defined, which makes it scalable and flexible. Edge computing devices work with the cloud to attain the proper processing in the correct location. This type of computing requires an application-aware orchestration layer to map the necessary resources for a workload’s specific requirements.

 

  • Intelligence and automation. Distributed computing uses artificial intelligence to understand where and how to process workloads. AI models analyze network traffic information to improve visibility, enable proactive troubleshooting and deliver rapid insight to resolve network issues. The use of AI allows both the distributed endpoints and the cloud computing infrastructure to be more effective.

 

Distributed computing promises new enterprise flexibility and benefits

Organizations realize it’s not practical to process all data in one area, whether it’s at the edge or in the cloud. Edge computing doesn’t replace the cloud, but it extends computing to wherever it’s needed by placing the required computing infrastructure closer to the endpoint that is running a workload. It eliminates the need to move data back and forth between far-away servers and connected devices. For example, a company may want to perform a specific task, such as detecting anomalies in a sensor reading on the machine where the sensor resides. Meanwhile, data science and information technology (IT) groups can deliver use cases, such as large volume historical data processing and AI model creation, in the cloud. With the introduction of 5G, edge computing and enhanced processors, enterprises can:

 

  • Deploy private wireless networks if desired. While public networks continue to improve service, companies also have the option to use 5G and Wi-Fi to build a dedicated private wireless network for their operational needs. Companies may own and manage the network or purchase it as a service from a third party, such as a mobile operator. These networks can support a range of applications, including deterministic communications and compute for mission-critical use cases. 5G can support low latency, high-availability services that require end-to-end quality-of-service, such as machine remote control, process automation, predictive asset maintenance, and CCTV monitoring. For example, the U.S. Department of Veterans Affairs (VA) Puget Sound Health Care System is working with AT&T to pilot a variety of healthcare use cases using 5G and multi-access edge computing (MEC ) technologies The VA is evaluating how 5G can support mobile-to-mobile connectivity across and between medical devices, allowing the tracking of people and assets within the facility. Additionally, it’s evaluating improved medical procedures and training using emerging technologies such as augmented reality (AR) and virtual reality (VR).

 

  • Deliver real-time actionable insight and support new use cases. Increasingly companies need to process and act on data immediately or in near real-time to deliver insight and data-rich experiences to their employees and customers. New Internet of Things use cases such as remote monitoring and control of assets also require acting on data immediately. On-device processing coupled with high-speed connectivity will allow companies to infuse analytics into applications regardless of where the data is collected. Distributed computing enables mission-critical use cases that need deterministic data flows, such as automation in manufacturing, telemedicine in healthcare, and real-time video analytics in retail.

 

  • Minimize costs, compliance, and security risks. Latency isn’t the only reason companies want to process more data closer to where it originates. The price for transporting and analyzing data in the cloud increases as companies connect more assets. Processing data at the edge will help companies control costs avoid unplanned cloud computing expenditures. .  Local data processing also allows companies to keep sensitive data on-premises to minimize data security and privacy risks while meeting specific regulatory compliance guidelines.

 

Businesses must prepare for a new connected future.

With the enablers now in place, business leaders recognize that a distributed computing environment can provide new value to their organization. However, there’s no “one-size-fits-all” approach to where and how a company should process its data. Each application has its requirements. IT leaders must categorize their workloads and define what functions to execute at the edge versus in the cloud based on throughput, latency and reliability requirements. In addition to performance requirements, companies may want to process certain information at the edge, such as analyzing patient or customer data records, ensuring data security, privacy, and regulatory compliance.

 

Many organizations also spend too much time designing a theoretical proof of concept trial instead of focusing on a small, well-scoped use case that can provide immediate value if it’s successful. For example, many IoT projects failed because companies didn’t match technology purchases to a business outcome.  While a company should define a comprehensive wireless strategy, it can gain early insight into its overall requirements by testing how various wireless technologies would solve a well-defined problem. Today, leading organizations are looking at how distributed computing can support use cases such as predictive maintenance, visually detecting manufacturing defects and using augmented reality for field repair. For example, Audi wanted to create smart factories and reduce the need for customized hardware and software solutions to handle individual use cases. An Audi factory that assembles 1,000 vehicles daily — each requiring 5,000 very precise welds — is now using a predictive quality-control solution to boost weld inspections by 100 times. Audi worked with Intel to create predictive analytics and modeling systems that improve quality control, increase factory efficiency, and turns data into valuable insights.

 

The world of distributed computing will only continue to improve as wireless technology becomes cheaper, faster and more reliable. Distributed computing holds untold promise for organizations of all sizes. The key to success lies in embracing technology today and defining a plan that allows the organization to evolve and add solutions as the market matures.

 

Regardless of the architecture employed, what’s clear is that distributed computing represents an immense opportunity for the enterprise to build new services, improve productivity, and ultimately compete more effectively.