Architecting the Intelligent Edge

JUNE 18, 2019 • BLOG POST • VISHAL LALL

IN THIS ARTICLE

  • For elements that characterize Intelligent Edge architecture are multi-protocol connectivity, at-scale management, embedded security and autonomous operations
  • In addition to digital experiences, edge use cases also enable business outcomes that cannot be achieved via traditional computing paradigms, including that of the cloud
  • Edge use cases range across various industries, including professional sports venues, healthcare and manufacturing

New approach that integrates connectivity, compute, analytics and security at the edge to enable unparalleled digital experiences and superior business outcomes

Over the past few years, many applications and workloads have migrated to centralized clouds that offer flexible infrastructure coupled with ease of use. In a world where everything is now hyper-connected, a growing amount of data from an increasing number of connected devices is spawning new use cases, closer to where the data is generated – at the edge. Latency and cost tradeoffs are challenging the rationale and value proposition of simply using cloud computing to execute these use cases.  Moreover, emerging technologies at the edge – such as AR, VR and IoT – all have requirements not served by the current cloud computing model.

The edge use cases are architected to deliver superior digital experiences such as personalized recommendations in a shopping mall or a seamless check-in process in a hotel.  In addition to digital experiences, these edge use cases also enable business outcomes that cannot be achieved via traditional computing paradigms including that of the cloud. These use cases integrate connectivity, compute, analytics and security to connect people and things at the edge, to apps and data in the cloud, and to enable control and action locally. These aren’t just promising, futuristic possibilities – they’re available today.

Changing the status quo

The technologies that couple edge connectivity and processing with cloud computing are already reshaping various industries. For example, to deliver a rich, physical/digital retail experience, a wide variety of devices such as cameras and sensors are being seamlessly connected, video streams are being analyzed for image and processing, and the results are then compared against large databases in real-time in the store to enable a superior shopping experience.  In addition, retailers are now starting to leverage artificial intelligence, machine learning and IoT-enabled presence-aware sensors to provide personalized, in-store recommendations that further contribute to a superior customer experience.

Similarly, hospitals are integrating connectivity with location-aware mobile engagement to improve digital patient experiences in areas such as visit planning and wayfinding.  Medical devices, such as infusion pumps and meters, are also increasingly connected, leading to use cases such as asset inventory and device analytics. Hospitals are also experimenting with machine learning-powered analytics to support diagnosis and treatment in radiology and other disciplines.  As an example, a recent Stanford study found that a new algorithm could read chest X-rays for 14 pathologies and perform as well as radiologists in most cases.  As these use cases enter the clinic, they will require local edge inference and processing, and strong security protocols.

Factories provide another example of the Intelligent Edge. Most factories currently run on operational technology (OT) networks that are typically segregated from IT networks.  As the manufacturing industry is starting to see the benefits of smart analytics to enhance production, we are seeing an initial convergence of OT and IT elements in their networks.  This trend will be further accelerated if the promise of 5G comes true and if it is successful in replacing elements of existing time-sensitive networks in factories. Manufacturing edge use cases include predictive analytics that leverage computation power to run AI inference algorithms that predict machine failures and alert operators, and/or take closed-loop autonomous actions such as shutting down at-risk equipment in advance of failures.

The new architecture

A new architecture is needed for the edge as the existing approaches don’t work for a variety of reasons. In particular, a distributed device network, many connectivity protocols, lower power envelopes, new security threats, and smaller device footprints all require a completely different approach.  Four elements characterize the new intelligent edge architecture.

Current connectivity solutions were designed for people as the primary user base. The Edge consists of a wide variety of heterogeneous devices – from microcontroller-based units to embedded devices in things – that connect via various protocols such as Ethernet, Wi-Fi, Bluetooth 5, Thread, and Zigbee.  Edge networks need to be purpose-built for these devices and must support multi-protocol connectivity, thus providing a consistent and converged experience across numerous devices, environments and use cases.

Furthermore, the distributed and heterogeneous nature of the edge network requires a completely new approach to onboarding, authentication, monitoring and management of devices.  As such, at-scale management of devices in the network is a key capability.  This capability takes on additional relevance as, in contrast to a datacenter environment, there are no staff on-site to fix software and hardware issues. Data and application orchestration are incremental capabilities needed in this management platform.

Security is a significant concern at the edge as traditional security models of protecting apps and data such as firewalls and data loss prevention (DLP) don’t work in a distributed environment. Moreover, the broad array of device architectures introduces considerable new challenges as compared with that of current cloud and mobile environments, which are much more hardened and controlled. Hence, embedded security is a critical capability as the need for protection from data theft and device takeover grows exponentially at the edge, requiring security capabilities that are enabled by AI.


Finally, autonomous operations are essential at the edge for three primary reasons. First, in many cases, connectivity at the edge may be inconsistent, and in some cases, not present at all, so cloud computing doesn’t work effectively and consistently. Second, business-critical systems require very high levels of uptime, requiring local processing. Third, applications at the edge have latency requirements to enable the correct experience and outcome which can’t be achieved by working off the cloud. 

HPE at the Edge

At Hewlett Packard Enterprise, we continue to expand our edge portfolio to enable new experiences and business outcomes for our customers.

At Tottenham Hotspur, we built the most technologically advanced stadium in the world with an unrivaled digital experience for fans. At the heart of the stadium operation, we enabled secure connectivity which is the essential foundation for the fan experience, providing a fully immersive environment that is at the heart of the stadium’s vision. Big screen displays, audio, responsive retail outlets, digital signage, turnstile access systems and more, are all connected via a multi-protocol network to local compute nodes. And the club is already seeing the impact of this technology, as it was one of the contributors to Tottenham gaining the largest annual profit of any football club in history this year.

At Boston Children’s Hospital, the renowned healthcare provider implemented a location-based mobile app to help visitors navigate its facility using a familiar blue dot and indoor maps across its six campuses. The mobile app allowed the institution to communicate with visitors, patients and employees in an interactive, digital way. In the first six months, the app was downloaded by more than 4,500 patients, and 65 percent of those reported that they felt their experience at the hospital had improved.

Texmark Chemicals deployed industrial IoT capabilities to create the refinery of the future. They modernized their infrastructure with predictive and advanced video analytics, and new safety and security capabilities, and they enabled the connected worker and implemented full lifecycle asset management.  The new IIoT solution makes it easier for the company to plan inspections and maintenance. Improved maintenance planning is estimated to reduce associated costs by at least 50%.

These new digital experiences not only deliver personalized experiences for the fans, patients and employees, but also significant business outcomes and economic value for the organizations.  The action is at the edge – and it is happening – today.

Tune in to Discover Las Vegas to learn more about our exciting Intelligent Edge announcements.

RELATED NEWS