Edelweiss - Fotolia
Failure in the data center is inevitable, but a new feature called Windows Server cluster sets tries to keep those disruptions to a minimum.
Clustering has long been a staple of workload resilience in the data center. Applications in a cluster operate across multiple redundant instances to share the load to keep things running, even when a node in the cluster fails. Until recently, the technology has been limited in scope, architecture
Windows Server cluster sets combine clusters into a common logical fabric that can span facilities and geographic regions. Each cluster within the fabric retains independent
How cluster sets increase resiliency
The goal of cluster sets is to support scale-out -- rather than scale-up -- capability. Consider a typical cluster: If the storage or memory resources start to run low, the administrator must scale up the cluster by adding more resources.
However, that isn't always a simple proposition. For example, adding storage may affect rebuild times, and it may be impossible to add more memory on hosts with full dual in-line memory module slots. Adding more nodes or redesigning an entirely new cluster is often expensive and undesirable.
Windows Server cluster sets combine multiple clusters into a fabric, so member clusters can share resources, such as storage and memory. This scale-out behavior reduces costs and adds flexibility to clustering without compromising the resiliency of the individual member clusters.
For example, a cluster that uses storage technology such as Storage Spaces Direct can survive the loss of two nodes. Thus, a cluster with four nodes will fail if three nodes fail. Yet, a cluster with eight nodes will also fail if three nodes fail. By comparison, in Windows Server 2019's cluster sets, two clusters with four nodes each can still function if the same three nodes crash.
Combining clusters into a greater fabric doesn't necessarily improve the hardiness of individual member clusters, but establishing a fabric of clusters and sharing resources can provide better resilience across the overall fabric. The fault boundary does not extend beyond the member cluster where the fault occurs.
Cluster sets break previous hardware
It is possible to create one cluster of older hardware and another cluster of newer hardware, and then join both clusters using Windows Server 2019 cluster sets. This can help organizations looking to repurpose or extend the lifecycle of existing
At a higher level, a cluster set can include clusters located in different physical areas to implement cloud-like availability sets or availability zones.
Cluster sets ease VM management
Windows Server cluster sets can simplify cluster management. Once an administrator establishes the cluster fabric with a unified namespace, they can use centralized cluster management to select the cluster and node best suited to handle a new VM based on the requirements and each cluster's available resources. Conversely, admins can tag VMs with specific needs to utilize predetermined clusters.
IT pros can move VMs between member clusters, while the individual cluster will determine which node should receive the VM. This behavior can benefit administrators with tasks such as transferring all VMs off an aged cluster before decommissioning it.
Dig Deeper on Microsoft Windows Server 2019
Related Q&A from Stephen J. Bigelow
Learn how load balancing in the cloud differs from a traditional network traffic distribution, and explore services available from AWS, Google and ... Continue Reading
Access management is critical to securing the cloud. Understand the differences between AWS IAM roles and users to properly restrict access to AWS ... Continue Reading
Containers have rapidly come into focus as a popular option for deploying applications, but they have limitations and are fundamentally different ... Continue Reading