Edge VMs are an infrastructural evolution of traditional virtual machines in which the storage and compute resources that power the virtual machine are located geographically closer to the end user (i.e. at the Internet’s edge).
To get a better understanding of edge VMs and their functions, let’s first review what virtual machines are.
As we explain in our definition of VMs, a virtual machine (VM) is a software-defined computer with its own operating system that runs on a host server with a different underlying operating system. Many VMs can share the resources of a single physical machine. All these VMs operate independently and can fulfil the same various functions of any OS within an IT infrastructure.
The purpose of edge computing is to deliver computing capability to the very edge of the network. With the development of edge computing, an increasing demand for virtual machines at the edge has arisen. Along with containerized applications, edge VMs promise a much faster, more efficient network architecture.
Though similar, edge VMs differ from cloud VMs in their placement in the network. Edge VMs are located as close as possible to the end user while cloud VMs are usually housed in datacenters outside of town. Real estate cost was always a big factor for vast datacenters filled with rows of machines. The problem is that these server farms may very well be thousands of miles away from the end user, resulting in high latency.
First, a developer creates a workload and houses it in a point of presence (PoP) close to end users. For sites in distant locations, the developer can easily replicate the workloads in multiple PoPs. In this way, it functions like a content delivery network (CDN). VMs at the network’s edge decrease latency for end users and greatly improve application performance.
Second, if multiple workloads and PoPs are used, the developer can use an anycast IP to route traffic to the closest VM/PoP to decrease latency. One snowy week in Canada, techie Andre Toonk wrote about his experiments with StackPath edge computing and anycast in a blog post on Medium. “It’s cool to be able to deploy your workloads around the world with just a single click,” he wrote.
Third, like any other virtual machine, an edge VM is a computer file that behaves like a computer. As Microsoft points out in their definition of a virtual machine, it’s like a computer within a computer. What distinguishes an edge VM from other types of virtual machines is its proximity to the user. As the Canadian blogger discovered, dispersing these software instances to the edge is pretty easy.
Fourth, in the case of StackPath, edge VMs have the added benefit of a private network backbone. Based on the average total response times for 1,000 requests, StackPath’s private backbone is 21% faster than the public Internet.
Suppose your favorite cafe is in the center of town but you live in the suburbs. It’s worth the trip because you love the food, the atmosphere, and the service, but the trip still costs you a higher amount of time and resources. Finally, the owner sees the growing demand and decides to open another cafe on the edge of town—in your neighborhood, in fact. This saves you a lot of fuel, time, and aggravation. Similarly, edge VMs offer computer processing closer to the user, saving all the costs related to round-trip data transmissions to a central server.
StackPath’s containers and VMs at the network edge reduce latency and increase efficiency in the delivery of application services. This results in high performance networking that takes advantage of other StackPath services, such as a content delivery network (CDN), a web application firewall (WAF), and a global domain name service (DNS).
One reason for the improved efficiency is the density of the VMs. Unlike hardware that uses a certain physical footprint for each machine, VMs existing in software have no real physical dimensions. You can pack hundreds of VMs and containers in one box, a virtual world within a single machine. This fits nicely into today’s hyperconverged infrastructure where there is no longer the need to reserve so much data center space for endless racks of servers. The reduced space requirements allow edge VM providers to easily put their VM hosts in densely populated areas, even city centers.