Security, connectivity, remote management and heterogeneous hardware are the main components of edge facilities. People need to understand the management and maintenance methods of edge data center. The architecture of edge computing represents a new generation of small infrastructure and new technical maintenance requirements. But even if they are decentralized, factors such as reliability, security, and edge data center maintenance are just as important as on premises data centers.
Edge data center facilities operate the same common computing hardware as data centers, including storage and network devices. Enterprises often deploy edge data centers in non-traditional locations, such as mobile communication base stations, factory workshops, and even in the field. Edge data center provides management for multiple devices on site, usually supporting power capacity of 50KW to 150KW.
For managers, a major problem is the large number of edge computing sites that need to be managed, while it is not realistic to set up a dedicated team for each edge data center. Managers can manage and maintain the on-site data center, but it is not necessary for edge deployment. Another issue is how to integrate edge data center facilities into the overall data center strategy.
These two problems lead to four major challenges of edge computing that data center managers must solve: security, connectivity, remote management, heterogeneous hardware.
1. Strengthen security settings
Security, both application and physical, is one of the biggest challenges facing edge computing facilities. Data center operators must invest in application vulnerability management and products that detect security threats and prevent network attacks.
The edge data center extends the whole attack area, because each edge site is a potential network access point. Managers must ensure that applications running on edge servers do not become a potential launch platform for attacks on other infrastructure.
As part of the maintenance of the edge data center, physical protection measures to protect equipment need to be considered. The security gateway on each edge site is a best practice, protecting the physical port can reduce the possibility of malware intrusion or other attacks.
Open ports may be plugged into malicious devices to load malware. Administrators should encrypt static data so that they do not steal or copy data from physical storage devices. Edge data centers should also support security lifecycle updates that need to be verified to prevent hackers from hijacking data or embedding malware into the system.
In a local data center, managers usually don't have to worry about locking the standby port because it's more difficult to gain physical access to the actual server.
2. Keep the connection across the edge
Servers in an on premises data center usually have strong and reliable internet connections; edge deployment is not always the case, especially because low bandwidth and intermittent connections are characteristics of edge use cases. One strategy is to establish auxiliary control channels. The other is to realize the independent management function to solve the problem of network interruption, which is part of the maintenance of edge data center.
Managers must also consider data backup and protection workflows. For distributed data centers, problems such as data protection and availability become more difficult because of the wider range of attacks. For example, lightning might hit one of the edge data centers. Management must consider the impact of site damage on data loss and adopt a failover strategy. When hardware and software are distributed, these challenges will become more challenging, because when data is lost or interrupted across different networks, all problems in traditional data centers will be magnified.
Traditional data center management mode is not suitable for edge facilities. Especially for valuable data, one strategy is to send all data to the cloud platform or other nearby access points to increase flexibility. Another strategy to reduce space and power requirements is to tier data so that only hot data remains on the edge.
3. Remote management and edge data center maintenance
The edge data center of remote operation needs to use the basic components of large data center, such as cabinet or rack, uninterruptible power supply (UPS), backup battery, and network connectivity to make these components work together. The biggest difference is that if the technician does not know the equipment or information, it is difficult to solve the problem on site.
Ivo Petrov, an expert in Tripp Lite's Data Center Department, suggests that managers develop remote access strategies to help field technicians and reduce overall maintenance time through built-in service modules and KVM on IP or console servers.
Components with self-healing function or keepalive probe for restarting components, as well as UPS power supply with automatic and manual operation bypass function can ensure operability during equipment maintenance. Another option to minimize field service is to use a console server for out of band access over a mobile communications management network.
The distributed nature of the edge data center may require more maintenance and management personnel. In order to maintain consistency and security, it is important to establish standards for monitoring and maintenance across these providers. These standards should include configuration management, homogeneity assurance and comparative performance analysis.
It is also important to build a business model that complements the management model for the edge data center. The requirements of workload, lease, service level agreement and service quality determine the business model and directly affect the design of management model.
4. Simplify heterogeneous hardware and it oversight
Another challenge, says Jason shepherd, CTO of Internet of things and edge computing at Dell technologies, is that remote edge data center devices may use more combinations of hardware, operating systems, and protocol choices than traditional data centers.
These devices can also learn from the functions of PLC, computers, routers, and server management paradigms, which traditionally have different toolsets for management. The responsibilities of these managers are also scattered between organizations and departments.
The operational technology (OT) team typically manages industrial equipment and programmable logic controllers (PLCs), while the IT team typically oversees the data infrastructure. Solving the problems of Internet of things devices on the spot may involve these two skills. Operational technology (OT) personnel on site do not necessarily have the skills to manage the computing infrastructure, which managers need to have. Information technology (it) often does not have the necessary knowledge of the actual use cases that drive the business. Shepherd explained that the workflow of an edge data center must ensure that both it and ot teams have the tools they need to do their jobs.
Data center managers should use consistent management tools. Shepherd says managers can add a mechanism so that updates are not pushed to OT in critical processes. Developers such as edgex foundry of the Linux foundation and pulse IOT center of VMware are trying to decouple infrastructure from applications that users can use as a foundation.
Source: foreign translation, if you need to reprint, please indicate the source