What an amazing ride it has been so far; Datrium has been everything I hoped it would be, from technology and engineering to sales and management teams. This is my first Datrium “Beyond Marketing” series post, and I am covering the release of DVX 3.0.
The first part of this release takes Datrium where no hyperconvergence vendor has been before, offering support for multi-hypervisors (vSphere, Red Hat Virtualization and CentOS) and Linux bare-metal workloads, and support for bare-metal containers with granular data management.
- Red Hat Virtualization (RHV) support
- Linux Bare-Metal (RHEL and CentOS) support
- Docker Persistent Volumes (Virtualized and Bare-Metal)
- Full Data Services for Containers
Do you know Datrium Open-Convergence?
From an architectural perspective, the best way to describe this game changing tech is to visualize all active data, both VMs and Containers, serviced with data locality and using internal flash (SSD and NVMe) on each server. At the same time, a protection copy of the data is hosted in clustered data nodes with distributed erasure coding. Each server runs the DVX hyperdriver software responsible for IO processing and enterprise data services.
One of the advantages of the architecture is that servers are stateless, and losing any given amount of servers doesn’t impact data protection, availability, or SLAs. On the other hand, data nodes are highly available and protected with active/standby controllers, mirrored NVRAM, and hot-plug drives.
Lastly, when applications move between servers or when a failover happens, the DVX software instantly uploads the data to the target server. The DVX software uses other servers as the source before pulling data from the data cluster, guaranteeing flash to flash performance whenever possible. Nevertheless, because of the native global deduplication is it likely that most fingerprinted data is readily available on the target server.
For official information on features and time frame refer to the official Datrium Press Release (here).
Red Hat Virtualization (RHV) Partnership & Certification
Datrium customers now can deploy Red Hat Virtualization (RHV) and inherently get all Datrium data service benefits, including Flash and NVMe IO acceleration and end-to-end blanket encryption.
Red Hat is the world’s leading provider of open source solutions and has been named a Visionary in the 2016 Gartner’s Magic Quadrant for x86 Server Virtualization Infrastructure.
Besides enabling the use of data services, one of the biggest benefits of Datrium’s multi-hypervisor implementation is the ability use of the same DVX system for supporting concurrently RHV and VMware vSphere deployments.
Datrium is now certified by Red Hat and providing support for RHV we are providing choice to customers, but also paving a path to support the entire Red Hat stack and application partner ecosystem, including OpenStack, OpenShift, and CloudForms, providing a unified and consistent set of management capabilities across:
- Red Hat Virtualization, VMware vRealize, and Microsoft Hyper-V(*).
- Private cloud platforms based on OpenStack®.
- Public cloud platforms like Amazon Web Services and Microsoft Azure.
Whilst Datrium works independent from CloudForms, it does enable multiple virtualization platforms to run across the same DVX system, eliminating silos and complexity, and in some cases enabling easy workload migration between hypervisors.
An interesting fact about RHV is that it has record-setting SPECvirt_sc2013 benchmark results, including highest overall performance and the highest number of well-performing VMs on a single server.
(*) Microsoft Hyper-V is not supported at this point in time.
Download Red Hat Virtualization Runs on Datrium Solution Brief (here)
Linux Bare-Metal (RHEL and CentOS)
DVX 3.0 enables customers to deploy the DVX hyperdriver on Linux bare-metal servers and inherently enjoy all enterprise data services benefits from the DVX platform, including Flash and NVMe IO acceleration and end-to-end encryption.
Linux administrators see datastores as local NFS mounts, and the mounts are backed by the DVX hyperdriver (manually installed in each server with 3.0 release) responsible for enabling IO acceleration and data services.
With this release, Datrium provides support for KVM and Containers, but other use-cases may be supported in upcoming releases, including Splunk, SAP, Hadoop and more.
Containers Persistent Volumes (Bare-Metal and Virtualized)
Containers are ephemeral, and files and services running inside a Container will not exist outside its lifetime. However, many applications require the ability to persist user session activity, making some aspects of the application stateful. Enterprises want persistent storage for Containers, and they also want to use the same infrastructure to manage dockerized and traditional workloads during the application lifecycle, development and production.
Datrium native Container implementation, via a Docker Volume Plugin, enables customers to seamlessly implement Continuous Integration/Delivery and Micro-services solutions as part of the delivery infrastructure, while still leveraging Datrium data services, such as deduplication, compression, erasure coding, encryption, replication, snaps, clones, etc.
No more choosing between bare-metal and hypervisor
Containers and VMs used together provide a great deal of flexibility in deploying and managing apps.
Organizations usually start their Containers journey running apps in VMs for the added flexibility provided by virtualization stacks. However as soon the application lifecycle and methodology are fully defined, organizations move their production Containers environment to bare-metal to harvest additional performance, reducing the (9-15%) CPU overhead created by the virtualization stack.
Datrium supports Docker persistent volumes for both virtualized and bare-metal deployments, while still providing IO optimization, acceleration and data services, including end-to-end encryption, snaps, replication and more. Using Datrium’s approach to Containers the development lifecycle is streamlined and automated much more easily because the drift between environments (Dev, QA, Staging, Pre-Prod, and Prod) is minimal.
Image courtesy of Docker Website
Data Services and Protection for Containers
Albeit some may argue that Containers should remain ephemeral, in my experience working with enterprises, there is a clear need for maintaining persistence across sessions for some applications and datasets, but also there is an enormous need to protect data in persistent volumes.
With Datrium persistent volumes may be cloned on one server can be immediately used on another, between both virtual and bare-metal deployments.
A significant challenge with Containers, however, is that it represents an order of magnitude more objects to manage than virtual machines, especially when persistent volumes are implemented. DVX 3.0 addresses this challenge with a combination of powerful search capabilities, the ability to create logical groups of Containers (called a Protection Group) aligned to applications, and assignment of protection policies to those groups for instant recovery, archive, DR and more.
In other words, all data services typically used with virtual machine workloads, such as snaps, cloning replication, and blanket encryption are now also available for Containers at the granular Container level, and the Datrium GUI makes it easy to understand and monitor.
Download the Docker Containers on Datrium DVX solution Brief (here)
I will soon release the second part of this blog series. Keep tuned!
[Update] You can find Part 2 of the announcement here.
This article was first published by Andre Leibovici (@andreleibovici) at myvirtualcloud.net