StackPath has two products that allow execution of custom code and applications on the global StackPath platform - Edge Computing Containers/VMs and Serverless Scripting.
Serverless Scripting allows execution of isolated scripts in response to user requests. This allows low latency execution of complex logic as geographically close to the user as possible. Being able to intercept a request, perform some processing, personalize the response, and return the response within milliseconds allows for improved response times, lower costs, and better user experiences.
Serverless is a great way to implement certain types of functionality, but what if the script also needs to perform callbacks to a central API, make database lookups, or process image data through a custom library? These still involve making calls back to separate infrastructure, perhaps running in a central public cloud region, which means more latency, bandwidth costs, and slower response times.
This is where Edge Computing comes in.
Edge Computing Containers/VMs
StackPath Edge Computing is a fully managed environment that allows container and VM workloads to run in any of our global locations. With no clusters to manage, the workload requirements are specified (CPU/RAM sizing) along with:
- Where the instances should be deployed
- How many instances should run globally
- What disk resources are needed
The container/VM image is then uploaded and StackPath does the rest.
StackPath will ensure that the number of instances specified per location are always running, manage the reliability and redundancy of the deployment, and deal with all security patches. You need only to focus on what's running inside the containers or VMs.
Combining Edge Engine and Edge Computing
Workloads are deployed on the same platform that all StackPath products run on. With free, low latency connectivity between other instances and StackPath services within the same location, applications can take advantage of sophisticated request flows.
For example, content can be hosted on our CDN, requests can then be intercepted with Serverless scripts, heavy processing handed off to a manipulation library running on containers, which then returns a customized response back to the script and finally to the user, all within the same edge location and completed within milliseconds. Eliminate unnecessary network hops back to centralized infrastructure and save on bandwidth costs at the same time as improving response times.