Abstract

Virtualization allows for more efficient hardware usage by allowing several instances of Operating Systems (OSs) or Virtual Machines (VMs) to run on a physical server. Containers are a subset of lightweight virtualization and reduce the overhead of virtualizing an entire OS by sharing the server's OS with the virtualized instances. Moreover, containers work closer to hardware than VMs and are similar to Linux processes, however, this limits connectivity freedom since processes do not have access to network addressing. To resemble container's communication to that of conventional networks we use network drivers. Studies show that in processing- or memory-bound scenarios, containers perform better than VMs, but in network-bound scenarios they achieve less performance. This work analyzes performance of networking implementations for Docker in different container allocations and workload scenarios.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call