Datrium
The Open Convergence Revolution

Powered by the synergy of multiple domain experts including former Data Domain CTOs and principal engineers from VMWare, Datrium marks the confluence of virtualization and storage to herald a new era in data center management. In an interview with CIO Applications, Craig Nunes, VP of Marketing, and Rex Walters, VP of Products, speak on Datrium’s successful journey over the last couple of years in delivering value to clients.

What spurred the idea of Datrium and how has the journey been so far?

Nunes: Fundamentally, our founders came together to address converged infrastructure differently than conventional market players. Customers were looking to keep pace with business, which the legacy array-based converged infrastructure did not support; a more agile infrastructure approach was needed. Although hyper-converged Infrastructure (HCI) had debuted a couple of years prior to Datrium being formed, it could not meet the data centers’ requirement for converged infrastructure for scalable tier 1, low latency, and mission-critical workloads. HCI failed to provide the scale, performance or the architecture to meet the requirements of multi-cloud enterprises. This is precisely the area we focused on, called Open Convergence and can be thought of as the combination of tier 1 HCI, scale-out backup, and cloud DR.

As for the journey, although Datrium started with virtualized infrastructure and VMWare, we recognized a need to incorporate Linux, KVM and container-based infrastructure to enable greater application mobility between clouds. Enterprises—predominantly large SaaS businesses or companies with large test and development organizations—have begun moving towards container-based infrastructure. Unfortunately they have had to set up an entirely new infrastructure silo to address the shift to containers, fracturing their infrastructure strategy. We cater to the needs of such clients because we bridge their current virtualization environment, either Vmware or Linux KVM, with their emerging bare-metal container environment within a single converged system.

What are the challenges that clients face in container-management?

Nunes: From a business outcome perspective, the challenges are three-fold. Firstly, SaaS-based software development organizations must deliver high performance, which directly affects their service levels and revenue. Secondly, any time that development teams need to spend setting up infrastructure or waiting on infrastructure to perform certain tasks directly impacts there ability to produce new software. Thirdly, the pace of business growth and the desire to modernize is heavily constrained by size and scope of budgets. Multiple silos are associated with today’s virtualized primary workloads, backup and cloud infrastructure that draw from a limited set of funds and force other tradeoffs. We bring 5-10x better performance than traditional approaches, eliminate administration time and software development time with container-specific data services, and can do so for a fraction of the cost versus arrays and hyperconverged infrastructure.

Walters: At the product level, there are three requirements as well. First, enterprises are trying to keep the data close to running applications, through Solid State Devices (SSD) for application performance reasons. Second, safeguarding the data is equally important, which necessitates equipping the data with mobility to move along with their respective workloads. Building a resilient architecture to keep the traditional storage environments intact and deliver scalable performance and resilience is a challenge. The third requirement is to have a scalable distributed namespace spanning all the functional applications.

We present a global pool of data to ensure its availability to all running applications and still provide the statelessness, where the data isn’t lost when applications break down; if data moves to a server different from where the application is hosted, we can still push it back. That’s fundamental while implementing a distributed process in any environment.

What is the value proposition you deliver to your clients from your container management solution?

Nunes: We consolidate virtualized and container-based infrastructure on a single platform—Datrium DVX. Further, we converge that with our clients’ backup and secondary storage infrastructure as well as cloud DR requirements, providing one system and one pane of glass to manage it all. Organizations need not worry about what portion of the infrastructure should be dedicated to container-based software development/production, or about underutilized silos in traditional virtualization or container-based applications. It is a single composable system serving the whole spectrum of needs.

What methodology do you follow to deploy solutions for clients?

Nunes: Our system, Datrium DVX, is composed of two tiers—a server-powered performance tier, essentially scale-out servers running our software called ‘compute nodes’ which adds performance to the system as you scale out. The protection tier is composed of data nodes that provide capacity and persistence for customer data. In deploying DVX, it is as simple as plugging it in and powering it on, and in 10 or 15 minutes you are serving and protecting data.

In terms of scalability, you can start with a single compute node and data node, and can scale up to 128 compute nodes and 1.7 petabytes of data nodes for an incredible level of scale and performance.

Which are the primary industries that you cater to with your solutions?

Nunes: We cater to government, education, financial services, energy, media/entertainment, manufacturing, healthcare sectors and cloud service providers. From a container perspective, we have seen demand from SaaS companies; education and technology companies have also been early adopters of our solutions. Most of our clients are running a mixed of performance-intensive workloads on a variety of server/compute node configurations and infrastructure variations.

While traditional IT operations may have lots of VMware installed, our platform allows a mix of container-based hosts and elevates them to first class citizens from a data management perspective. The ease of use, massive scalability and data protection on-prem or in the cloud extends to the whole environment, whether it is a Linux-based container farm or traditional IT running on VMware.

What is Datrium’s roadmap for the foreseeable future?

Nunes: Today, the market adoption of our platform has reached new heights. We are just seven quarters in the market but have already several hundred deployments of our product at client locations. This year, we are looking at doubling the U.S. sales and have recently announced new operations in both Canada and Japan. Our alliance and technology partners have observed our work and have grown from half-a-dozen to more than 30 in the last six months. The market has begun to experience the value we bring with more organizations jumping on board everyday.