Successful edge computing deployments require scalable management tools and infrastructure that run with minimal...
IT interference. To that end, organizations should consider investing in micro data centers, data processing software and centralized infrastructure management. Edge computing brings opportunities for data processing in remote and branch office deployments, smart factories, autonomous cars and unmanned sites, such as wind turbines and scientific data gathering in Antarctica.
There are predictions that edge computing will be bigger than cloud, which is quite bold given the millions of physical servers that make up public cloud infrastructure. According to research firm Million Insights, this market is expected to grow to $3.24 billion by 2025.
Micro data centers support data generation and local data processing. Organizations can use them to act immediately on data or to reduce the cost of data transfer, which is what makes them ideal for edge computing.
Micro data center basics
Organizations use edge computing for a host of reasons. No matter the size or type of deployment, central management is key.
Central management helps admins monitor individual micro data centers and run them without any on-site IT staff; only commissioning and decommissioning should require someone travel to the installation location. Admins should have the right software and take time to set up proper workflow filters to easily identify micro data center information within the overall IT ecosystem.
Security is an integral part of any data center, but the unique challenge with micro data centers is that they may be in unsecured locations -- such as a roadside communications cabinet -- so a secure physical enclosure is essential.
Before installing micro data centers, admins should research if they must add remote management functions to their data center infrastructure management software and what hardware access methods they need at any remote locations.
Where can admins use micro data centers?
One reason organizations use micro data centers is for large-scale data center operators that must reach more markets and underserved locations. Telecommunications companies are a classic example; they often move services -- such as video on demand -- to new locations, which require scaling up infrastructure to meet increased processing needs.
Another use for micro data centers is to bring IT capabilities alongside operational tasks, such as factory equipment maintenance. A micro data center can link sensors in a factory to machine learning software and enable just-in-time manufacturing maintenance.
Of course, each organization has different edge computing requirements, so admins must populate the rack with organization-specific servers, network, storage and management software.
Hewlett Packard Enterprise's (HPE) Edgeline Converged Edge system is a range of x86 servers designed for a factory or another data collection location that some consider a micro data center. Admins can manage HPE Edgeline servers with the same tools as their other HPE servers, which enables common management for remote data centers. HPE also offers another product that it classifies as a micro data center, which has HPE server racks optimized for non-IT managed spaces.
Dell's modular data centers can pare down to a few infrastructure racks at each location, but they use the same scalable management framework as an on-premises data center. Dell also sells outdoor enclosure options to hold these micro data centers that are closer to telecommunications roadside cabinet sizes than enterprise IT data center racks.
APC, a Schneider Electric company, has data center infrastructure management tools and a range of hardware; its programs remotely control and measure power use, as well as environmental conditions, such as temperature and humidity. APC's micro data center racks are designed for locations without IT staff or even a controlled environment. The racks have electronic locks, air filters and active cooling mechanisms that are all centrally managed.
There are a few other visions of edge computing deployments from software and service providers. As more companies invest in edge computing infrastructure, admins will have offerings beyond micro data centers: Amazon has AWS Snowball Edge, which is a small, suitcase-sized appliance with storage and compute that can run Amazon Elastic Compute Cloud instances or Lambda functions inside a building. And there are instances of VMware's ESXi hypervisor that run on ARM CPUs, which is one potential direction for future micro data centers.