Unikernels vs Containers: An In-Depth Benchmarking Study in the Context of Microservice Applications

Abstract

Unikernels are a relatively recent way to create and quickly deploy extremely small virtual machines that do not require as much functional and operational software overhead as containers or virtual machines by leaving out unnecessary parts. This paradigm aims to replace bulky virtual machines on one hand, and to open up new classes of hardware for virtualization and networking applications on the other. In recent years, the tool chains used to create unikernels have grown from proof of concept to platforms that can run both new and existing software written in various programming languages. This paper studies the performance (both execution time and memory footprint) of unikernels versus Docker containers in the context of REST services and heavy processing workloads, written in Java, Go, and Python. With the results of the performance evaluations, predictions can be made about which cases could benefit from the use of unikernels over containers.

Virtualization Costs: Benchmarking Containers and Virtual Machines Against Bare-Metal

“…Moreover, since unikernels support the execution of one process, they increase isolation-although unikernels run applications in kernel mode, making it unclear how they affect the exposure of the system to attacks and bugs-and provide a simpler model for performance prediction. From the software engineering point of view, while VMs and containers do not impose strong constraints on how programmers develop their applications, unikernels require applications to be written specifically for them [10].…”

“…RAMspeed RAMspeed [31] is a cache and memory benchmark for several UNIX-like operating systems. One can use it to benchmark integer, floating-point, MMX, and SSE instructions in two types of benchmarks: *mark benchmarks that pass sequential data streams to the relevant hardware units, and *mem benchmarks to measure the speed of simulated Copy, Scale, Add, and Triad tests 10 with the support for a batch run returning the average speeds. VTmark uses version 3.5.0 from http:// www.…”

“…https:// gitlab. com/ knurd 42/ kcben ch 10. Triad is a combination of Add and Scale operations: it reads data from the first memory location, scales it, then adds data from the second one and writes to a third place 11.…”

Enabling and Leveraging AI in the Intelligent Edge: A Review of Current Trends and Future Directions

“…Another recent virtualization technology is unikernels [36], which are essentially virtual machines that avoid context switches by always running in kernel mode. They also minimize their memory use and image size by only including the kernel functions required by the software they are compiled for.…”

Real-time detection and optimization algorithms for flexible resource allocation in 5G networks and beyond

“…In this work, to address the aforementioned issues, we assume a novel and flexible edge cloud environment exploiting lightweight virtualization technologies, such as unikernels [47] or containers, in the context of a video content distribution environment. The core idea in our proposal is that virtual servers (i.e., virtual machines or containers) could host individual video content and CHAPTER 2.…”