Scalable Virtual Machine Multiplexing

Опубликовано 6 сентября 2016, 17:55

Recent years have seen a significant surge in the interest and use of virtualization, driven by better hardware support, increasing power costs and low resource utilizations. While virtualization has many applications --- for instance, to support legacy software, to improve fault containment and to provide strong resource guarantees --- by far the most common is server consolidation. By aggregating multiple services on a single physical machine, organizations can reduce costs and increase the utilization of their infrastructure. Critical to server consolidation is the ability to efficiently multiplex several virtual machines (VMs) on a single physical machine. An inevitable consequence of multiplexing is resource partitioning: individual VMs can only use a fraction of the actual physical resources. This limits not only the resources available for a single VM, but also the total number of VMs that the hardware can support. This is a fundamental barrier---the aggregate resources available to the VMs are bounded by the capacity of the underlying hardware. This talk will discuss mechanisms to increase the perceived resource capacity of individual VMs as well as the total number of VMs. I will present ``time dilation'', a technique that allows the perceived aggregate capacity of the VMs to exceed the capacity of the underlying hardware, enabling some interesting applications such as predicting protocol performance in resource rich environments. I'll then show how to leverage time dilation to build a framework and methodology called DieCast, to accurately test large systems using a much smaller infrastructure. Finally, I will describe Difference Engine, a system that exploits fine grained similarities in memory among VMs to extract twice as much memory savings as the current state of the art, thus freeing memory for additional VMs.