Virtual Memory Streaming (VMS)

Whether you’re a service provider, a SaaS vendor, or an enterprise, running a cloud is not an easy task. You're constantly trying to scale up your apps, improve the performance of your infrastructure and lower your costs. Why not let Virtual Memory Streaming handle that for you?

Virtual Memory Streaming (VMS) is a game changing scale-out virtualization technology that allows you to get the best out of your cloud virtualization layer - rapid scale, high performance, and low TCO. VMS is compatible with most commonly used open source cloud hypervisors such as KVM and Xen, as well as integrates out of the box with OpenStack platform. Unlike other performance enhancing solutions, VMS is not a generic storage based virtualization play. Instead, it optimizes your entire virtualization experience providing faster boot times for your virtual machines, generous runtime memory savings and a superior scale-out platform for your apps.

VMS Features

How does it benefit you?

Launch using live images Using live images, VMS speeds up the time it takes to launch your virtual machines by 10-100x and cuts peak IOPS by 90%.
On-demand memory streaming VMS dynamically allocates and streams memory on demand. This reduces the memory footprint of your virtual machines and increases the density of virtual machines per server by 2-5x, without any performance loss.
Painless active directory integration VMS allows you to easily integrate your virtual machines with your existing enterprise infrastructure such as active directory, without a single reboot.
Easy integration with Openstack and Xen Server environments VMS is available as an easy-to-install package for Openstack and Xen Server. It is compatible with Xen and KVM hypervisors, as well as different Openstack storage mediums such as Ceph RADOS block device and Logical Volume Manager.

Virtual Memory Streaming under the hood

VMS brings together several advanced hypervisor memory management techniques including granular page sharing, dynamic memory footprint management, cloning, read caching, and a unique virtual machine cloning capability. These techniques work together to provide rapid scaling, high performance and lower TCO for your virtualization layer.
With VMS, virtual machines do not have to go through a performance intensive boot process. Instead, they launched from pre-booted live images and are ready to serve requests within seconds. The launched virtual machines have a have a low memory footprint, resulting in dense memory oversubscription.

It all starts with a single live image

Using VMS, you can create a live image of a running virtual machine instance. The entire disk and memory state of the running instance including the operating system and applications is preserved in the live image. The live image can be stored in a variety of different storage backends, such as distributed object stores as well as centralized network file systems.

Creating a new live image snapshot

For OpenStack, VMS is implemented as an OpenStack extension. A lightweight service that runs on each compute host and exposes an API primitive for nova called live-image-create.

From a single live image to multiple running instances

When a new virtual machine is launched, VMS streams the live image state to the new instance making it immediately available. The new instance starts up from where the snapshot left off. Common elements are de-duplicated and shared between virtual machines to reduce IO and storage needs. The new instance is automatically configured by VMS on the fly - when the new clone comes up, we change its network identity and hostname. VMS also seamlessly joins your new virtual machine to your active directory without a reboot.
Instantiating machines this way allows the simultaneous launch of hundreds of instances instantaneously. And, in addition to doubling the virtual machine density, VMS slashes boot IOPS by 90+%, significantly cutting storage and network requirements and the budgets to back them.

Launching virtual machines from live images

The VMS OpenStack architecture represents a radical rethink of machine instantiation — it increases scalability and performance as well as lowers cost. Having pioneered this technology, we’ve made it available as a software extension to virtualization platforms based on leading hypervisors such as KVM and XenServer and the OpenStack cloud framework. It is simple to install, affordable, proven, and fully supported. Request a free trial today.

Learn More at Your Own Pace.

Read Our Blog Join a Webinar