Data Centre Intelligent Automation and Management Innovation of the Year
What, exactly, is novel about the product / process or innovation?
Adaptable Redundant Power (ARP) is a recently patented critical power control system designed for data centres.
The ARP control system manages the power system configuration to enable the simultaneous existence of different virtual redundant IT load power configurations (e.g., 2N, N+1 and N) on the same power infrastructure, and can seamlessly change the configuration to match the IT load service level agreement (SLA) in a cloud environment.
ARP also enables the data centre to access what is normally referred to as trapped power or stranded capacity, using a switching technique called Inherent Power, thereby reducing cost and overprovisioning.
ARP can also be deployed as a disaster prevention tool to execute IT load prioritization to prevent a total data centre failure, in the event of multiple component failures.
How does the product or process break with conventional ideas or processes in its field?
Conventional data centre power systems cannot change their configuration. This means, once they have been built, no matter what the system topology or capacity is, the power system reliability and availability levels cannot be altered without physically rewiring a part of, or the whole power system.
ARP creates a “virtual” pool of power, made up of the combined parallel resources from multiple UPSs. The power can be intelligently and automatically switched and allocated to particular racks or IT loads to match the IT workload service level agreement (SLA).
How does it go beyond marginal improvements on something that already exists?
ARP is fundamentally different to existing power systems because it has the following attributes:
• Increases the amount of available power by accessing stranded capacity, therefore reducing capital cost, operating cost and overprovisioning.
• Enables the power system SLA to dynamically match the IT SLA via virtual redundancy.
• Provides the ability to automatically react to simultaneous multiple component failures.
How do customers benefit from the product / process or innovation?
End users generally operate cloud native applications managed by a data centre orchestration system based on Kubernetes*. Examples of companies that do this are AWS, Microsoft, Google, Oracle and IBM. The orchestration system decides when, where and how IT workloads are deployed in the data centre. IT workloads can have variable service levels. Typical examples of different IT SLAs are:
99.995% SQL database
99.99% Load balancer
99.95% k8s clusters (with AZ)
ARP functionally improves the alignment between IT and the power system by enabling end users to match the power system SLA to the IT workload SLA.
Also, because ARP accesses what would normally be stranded capacity, it reduces CAPEX and effectively increases the realisable capacity of the power system.
In addition, ARP protects the most critical IT workloads in the event of an N-1 scenario. For example, if a data centre had four generators configured as N+1, and two generators were to fail when the site is running close to full design capacity, the ARP control system will automatically disconnect the lower criticality IT workloads to protect the high priority workloads, thereby preventing a total data centre failure.
* Kubernetes is an open-source system for automating deployment, scaling, and management of containerized applications. It was originally designed by Google and is now maintained by the Cloud Native Computing Foundation. It aims to provide a "platform for automating deployment.