VMware vSAN HCI Mesh in vSAN 7 Update 2
The Scale-out architecture of VMware vSAN enables powerful non-disruptive scale-up or scale-out capabilities. You can non-disruptively expand capacity and performance by adding hosts to a cluster (scale-out) or just grow capacity by adding disks to a host (scale-up). As application workloads organically grow, this enables performance to be right-sized at each expansion interval. Despite this some challenges persisted for using HCI to handle some scaling events:
- An application refactor requires significant storage to be added for logs.
- A new line of business for analytics may consume excessive compute, potentially stranding storage assets.
- Compute workloads are scaling significantly while storage needs remain static.
- Purchasing and refresh cycles for storage and compute are distinct and sometimes years apart.
Historically, when these scaling events happen it could cause an existing cluster to run out of storage or compute and potentially isolate the lower demanded resource. While vMotion enables “Shared nothing migration” between clusters, this added extra time and manual effort to re-balance cluster usage. HCI mesh helps solve these challenges by bringing together disaggregated compute and storage workloads from disparate clusters.
What’s New with HCI Mesh in vSAN 7 Update 2?
To solve these challenges, VMware has updated vSAN HCI Mesh, in vSAN 7 Update 2 to support HCI Mesh compute clusters. HCI mesh previously had a requirement for all participating clusters to be providing vSAN storage. Now, compute clusters can seamlessly consume remote vSAN datastores. Scaling has also increased now up to 128 remote hosts, or 5 clusters can be connected to a vSAN cluster.
In addition, new improvements to VMware Storage Based Policy Management (SPBM) include the ability to select vSAN cluster data services. As virtual machines now have multiple choices per cluster potentially, this allows deterministic placement of workloads at the storage layer.
Going back to our earlier examples of scaling, here are a couple scenarios where HCI Mesh could help:
If you are looking to move to HCI but have an existing blade infrastructure with an additional two years of service life, you could deploy a vSAN cluster and export storage to their existing blades. This would allow you to fully utilize their compute asset for its service life.
If you have a requirement to add an isolated cluster for a small 2-3 DMZ or test dev environment but want to leverage RAID 6 you could deploy compute-only clusters for these special requirements, while consuming vSAN storage from the primary cluster.
If you are running an E-commerce web server farm that needs an extremely high ratio of compute to storage you may have found it difficult to size for HCI. If you need 10 nodes of storage, and 50 nodes of compute HCI mesh will allow you to deploy separate compute-only clusters.
How does VMware vSAN HCI Mesh work?
HCI Mesh is a unique, software-based approach for the disaggregation of compute and storage resources. HCI Mesh brings together multiple independent clusters for a native, cross-cluster architecture that disaggregates storage and compute resources and enables efficient utilization of capacity. Simply, HCI Mesh allows one or more clusters to remotely mount datastores from other vSAN clusters within vCenter inventory. This approach maintains the essence and simplicity of HCI by not fundamentally changing the existing HCI model or requiring specialized hardware. Now, a cluster with excess compute can mount excess storage from a remote vSAN cluster.
Requirements to Setup HCI Mesh Compute Cluster
Setting up a cluster is quick and easy. Simply configure vSAN VMkernel ports on the hosts that will need to connect to the cluster and enable vSAN on the cluster. You will be prompted to configure a cluster as a compute cluster. vSAN Enterprise or Enterprise Plus licensing is required to share storage from a cluster, but only vSphere licensing is required to mount a vSAN datastore from a remote HCI Mesh compute cluster.
Cross Cluster Management
HCI Mesh can help accommodate the requirements of the new vSphere Lifecycle Manager. vLCM was introduced in vSphere 7 as VMware’s next-generation framework for software and firmware management of vSphere and vSAN hosts. vLCM strives for homogeneous configurations since it uses a single desired image for an entire cluster. Since organizations may be limited to which clusters may utilize the new hardware, they can deploy the new servers in a cluster of the same manufacturer, but use the resource in another cluster through HCI Mesh.