The Difference Between Cloud Compute and Edge Compute

What’s the difference between cloud compute and edge compute?

Cloud compute is the delivery of compute power (CPU, Storage, I/O resources, DBs, etc) through a cloud services platform with a pay-per-consumption-based business model.

With cloud compute, you don’t need to invest large amounts in hardware and in managing that hardware. Instead, you can provision the exact types and sizes of the resources you need and only pay for what you use.

With edge compute, the business model stays the same but the performance increases due to the physical locations of the compute nodes (i.e. edge containers and edge VMs) being located closer to the end users.

How cloud compute works

There are many cloud compute providers that allow you to deploy containers and VMs. That’s been around for years. These allow you to run applications, databases, and different types of software without the need to manage the server itself. You pay for consumption, memory usage, or CPU time and everything else is covered. Your container or VM is deployed on a server, delivering the logic it was intended to.

The common thing for most of these providers is that they use a centralized location for the deployment of containers or VMs.

For example, in the United States there are three centralized locations/regions:

• West Region
• Center Region
• East Region

You usually choose one of these for your container and VM needs, which means that every request that requires some compute logic will travel from wherever the request was made to the compute location you are using. This is not very performant, especially when requests are coming in from the other side of the country (or the world).

How edge compute is different

Edge compute allows anyone to deploy containers and VMs at different edge locations (StackPath currently has 45). Instead of using a single location for all your compute needs, you can deploy your containers or VMs wherever there is an edge location.

For all these locations you will have a single ANYCAST IP address you can use and the IP will automatically GEO load balance each request to the nearest edge location. For example, users from Texas will hit a Dallas “edge compute POP” and users from Chicago will hit a Chicago compute POP.

True edge compute service providers locate their POPs in the most densely populated locations in the world that are also close to internet exchanges (IXs) in city hubs. For example, the Dallas POP could be located 200 miles north of Dallas or in the city center. The latter is better.

This new approach to compute dramatically reduces latency compared to using a centralized location, which means faster apps and happier users.

Still have questions about the difference between cloud compute and edge compute? Send them to us in this Twitter thread.