One of the more enticing aspects of VMware Cloud on AWS is the ability for an organization to project an IT presence anywhere in the world independent of their existing operations. From oil rigs to retail shops, edge compute is a critical component of the modern world. Managing the costs of those edge deployments without compromising on capabilities is where solutions such as vSAN differentiate from the competition.
Why vSAN for the Edge?
The issue with edge computing is entirely a problem of stepped increment scaling. The edge typically encompasses a large number of small deployments, each deployment focuses on a small amount of mission-critical and/or regulated data.
These sites, while small in size, are enormously important to the business and thus require enterprise-class availability and data protection. ROBO is often thought of as a lesser class deployment than the “core datacenter,” but this is an outdated perspective. Data is the lifeblood of the modern world, and it is increasingly generated and processed at the edge.
To further complicate this architecture, which already has to enable enterprise-class features at the lowest entry scale possible, these sites also tend to change over time. Something that starts small can rapidly change in size and/or scope. This demands a platform that can be recomposed.
vSAN is the perfect platform for this use-case, enabling customers to align the hardware to the specific requirements of each deployment while maintaining the flexibility to scale up or scale out as needed.
Edge computing and the vSAN for ROBO editions
While a standard vSAN Cluster requires at least 3-Nodes, a 2-Node variant exists specifically for these smaller sites, enabling customers to start as low as 2 hosts and scaling as needed. The entry scale 2-Node capability requires a vSAN Witness to store the witness components that act as the tie-breaking vote within the cluster. The vSAN for ROBO license editions make it easy to license and maintain these smaller deployments. This has prompted some to ask if they can combine the two and use VMware Cloud on AWS to host the witness for a ROBO or Edge deployment?
vSAN Witness Appliance in the cloud
The answer is a qualified yes. The vSAN witness appliance is an OVA, VMware Cloud on AWS supports OVA deployment. There are no supportability considerations from the VMC perspective, the witness appliance is just another supported vSphere VM.
There are however the existing requirements around the Witness Network throughput and latency. Entry scale 2-node cluster deployments require the witness network Round Trip Time (RTT) of no greater than 500 milliseconds (250 ms one-way), and at least 2Mbps throughput for every 1000 components being stored in vSAN. VMware Cloud on AWS can easily satisfy these requirements for much of the world. That coverage will only continue to improve as we expand the service availability within AWS.
Customers looking to modernize their edge infrastructure should take a hard look at VMware HCI and the capabilities provided by vSAN. Organizations who require a regional point of presence to consolidate and manage a broader deployment will find VMware Cloud on AWS the ideal platform. Together they empower customers to recompose their infrastructure globally as quickly as the market demands.