Developing and Deploying Microservices with Toro Unikernel
In a cloud-based architecture, microservices are deployed on VMs which are guests that run a General Purpose OS and the microservice runs as a user application. In this context, Toro is a dedicated un …
| Talk Title | Developing and Deploying Microservices with Toro Unikernel |
| Speakers | Matias Vara Larsen (Software Engineer, Huawei) |
| Conference | Open Source Summit + ELC Europe |
| Conf Tag | |
| Location | Lyon, France |
| Date | Oct 27-Nov 1, 2019 |
| URL | Talk Page |
| Slides | Talk Slides |
| Video | |
In a cloud-based architecture, microservices are deployed on VMs which are guests that run a General Purpose OS and the microservice runs as a user application. In this context, Toro is a dedicated unikernel for microservices that has several benefits against a General Purpose OS like faster instantiation, reduced memory footprint, and simple disk image management. In Toro, a microservice boots up in 150ms, consumes less than 4 MB of RAM and the image takes about 140 Kb of disk. The same image can then be used to launch VMs in a modern hypervisor like HyperV or KVM, or hosted in a cloud provider like AWS or GCE. This talk illustrates Toro by using as a running example a simple microservice that hosts a static web page. This example allows us to show the benefits of Toro in term of CPU usage, memory usage, and disk usage. Also, we show how easy is to change a network driver or a filesystem thus allowing the tweaking of a microservice depending on the target.
