Edge data centers: the next big thing in cloud computing

The humming banks of servers housed in data centers are the beating heart of our interconnected, data-driven world. In the early days of Web 1.0, data centers were sprawling, centralized giants. They are still around and still vital. But all our clicking, tapping, swiping, and streaming – coupled with an explosion of the “internet of things” (IoT) – has led to a massive surge in global internet traffic, now routinely measured in zettabytes. (In case you haven’t been paying attention: one zettabyte = one trillion gigabytes.)

Such staggering volumes of data have created a need to move from mega to micro – from a centralized to a more diversified IT ecosystem, which means lots of local data centers around the edges of far-flung global networks. This movement is already well underway, with edge data centers projected to soon handle 75% of all enterprise data generated worldwide.

Building, running, and maintaining edge computing solutions requires people and parts on site to keep everything running smoothly. Enter the need for a sound logistics strategy and state-of-the-art supply chains that are robust and resilient.

Edge data centers bring computer processing closer to users, sensors, and other IoT data sources – at the edges of far-flung global networks and away from core data centers. Deploying processing and storage facilities closer to where data is generated minimizes the need for continuous, long-distance communication between clients and servers, improving performance while delivering more robust data privacy and security.

With their smaller footprint and greater agility, edge data centers are designed to process time-sensitive data faster while sending less critical information to the larger, centralized data centers that run big data analytics.

• Speed: Smaller data centers closer to the devices generating and receiving the data result in higher bandwidth and faster response times – translating to better performance for customers and users. This is especially critical for highly time-sensitive use cases like autonomous fleet operations.
  
• Privacy: Processing data closer to its points of origin and end use makes it easier to comply with location-specific data privacy regulations. This is particularly important in areas where privacy laws are strict or still rapidly evolving.
• Security: Centralized data center architectures are much more susceptible to disruption – both technical outages and cyberattacks – than diversified architectures with more edge computing capabilities in place. Lower volumes of data transmission to and from the remote, centralized cloud make it easier to manage cybersecurity and pinpoint any problems that do arise.
• Access: “Dark zones” with limited broadband internet access continue to hold back the economic progress of rural or underdeveloped regions. But edge data centers can tap into increasingly available 5G networks without internet access, bringing the power of AI and IoT applications to these areas.
• Cost: Edge data centers can be tailored more precisely to regional needs, resulting in less over-resourcing and better long-term cost-efficiency. The up-front costs of establishing local nodes can be steep, but the enhanced speed, performance, privacy, and security make edge computing a sound long-term investment.
• Sustainability: Large, centralized data centers commonly over-resource at times of reduced demand. Edge data centers can be scaled more readily up or down in response to local demand. This requires considerably less energy for operations and cooling, improving the carbon footprint of the companies that operate them – and indirectly of those that use their services.

No gain without pain, of course. The many benefits of edge computing come with a few challenges:

• Up-front investments: The initial investment of time and money to establish multiple edge data centers, each with targeted microservices, must be considered from the outset.
• Monitoring: By definition, decentralization diversifies the task of monitoring IT infrastructure, making it more complex.
• Evolving technology: Edge computing is still in the early stages compared to established cloud computing models. Early adopters must always be prepared to adapt to changing standards and threats.
• Need for international expertise: Managing diversified networks and storage systems is complicated, requiring specialized IT expertise at multiple geographical locations simultaneously. This can be incredibly challenging in developing markets with sometimes opaque regulatory requirements. An efficient cross-border service logistics solution is essential here.

As a logistics company, we’re preparing to meet the growing supply chain needs of edge data centers. But we’re also keeping tabs on how the technology will impact logistics operations. Self-driving trucks moving in convoys – also known as autonomous platooning – will rely on edge computing for ultra-high-speed vehicle-to-vehicle communication. IoT sensors on the lead truck will detect road conditions and set the pace, streaming the data to an edge data center via the cloud for communication to the other vehicles, with the other trucks processing that information and providing a response alongside data from their sensors. The bandwidth and lightning-fast data transmission needed for this won’t be possible without edge data computing solutions.

At DHL, we regard edge computing as a trend likely to have a moderate impact within the next five years. However, deciding to establish edge data centers at local nodes is just the first step. Many other considerations are critical to delivering the kind of performance that makes edge computing a sound long-term investment, such as the speed of deployment, the resilience and reliability of the centers, and the ongoing availability of replacement parts alongside support and maintenance personnel.

Robust logistics and end-to-end supply chain management will be central to success. Operations and supply chain management will need to be integrated. And edge data center supply chains must have real-time data analytics to manage risks effectively and enable quick contingency planning. Logistics companies who know the local conditions and can also navigate across international spheres will have … well … the edge.