Containers and virtual machines
Triton Compute Service provides three classes of compute instances: hardware virtual machines, infrastructure containers, and Docker containers.
Triton virtual machines (VMs) are hardware virtual machines (HVMs) wrapped in highly secure zones, supporting a variety of images including Linux, FreeBSD, Windows, and Docker Engine.
Each HVM is implemented using a hypervisor: either KVM or bhyve. KVM was the first HVM supported by Triton and remains a strong choice for most VMs, particularly if you are already using KVM in Triton and are satisfied with the performance and features. Bhyve is a newer HVM which aims to improve performance and scalability, while growing the feature set. Because bhyve is relatively new, not all images are supported on bhyve. In general, Linux HVM images created by Joyent after August, 2018 may be used with either KVM or bhyve. For Windows, there are separate HVM images for KVM and bhyve due to the different boot loaders. Bhyve is a good choice if the image you need is available for bhyve. Note that KVM instances cannot be resized, whereas bhyve instances can be resized if the instance uses flexible disk space. Flexible disk space needs to be set up and enabled by an administrator.
Leverage container-native Linux on Triton infrastructure containers to extend the benefits of containerization to legacy applications and stateful services, such as databases. Containerizing an application in infrastructure containers is easy because they offer all the services of a typical unix host and behave similarly to hardware virtual machines.
Infrastructure containers enjoy their own virtual NICs, filesystems, and all the resource and security isolation that you'd expect of a Virtual Machine, but with the elastic performance and bursting that's only possible with containers.
Triton is the best place to run Docker. While other Docker run-time environments require complex VM host cluster management and provide only limited networking and security capabilities, Triton transforms an entire data center into an easy to manage, elastic Docker host, while delivering enterprise grade networking and security to each Docker container.
It can be difficult to decide which type of infrastructure to use for an application. Below are some example scenarios in which you may decide on using a containerized instance versus a virtual machine.
|Secure and high performance||✓||✓|
|Brand new microservice-based application||✓||✓|
|Existing application to be converted into microservices||✓||✓|
|Quickly resize instances without rebooting||✓||✗|
|Specific environment needs such as FreeBSD, Windows, and Docker Engine||✗||✓|