Abstracting the Data Center: A look at the DCOS Platform

Published on Friday, 28 March 2014

It’s time to take a step back and look at the data center model that’s impacting today’s business, . It’s time to see just how far this platform has come and exactly where it’s going. It’s time to say hello to the truly agnostic data center. Almost every new technology is being pushed through some type of data center model.

Inside of your current data center model – what do you have under the hood?

  • Storage, Networking, Compute
  • Power, Cooling, Environmental Controls
  • Rack and Cable Management
  • Building and Infrastructure Security

Although some of these underlying components have stayed the same. Requirements from the workloads that live on top have drastically evolved. Through it all, we’ve also seen an evolution of the physical aspect of the data center. We’re creating powerful multi-tenant, high-density platforms capable of handling users and the new data-on-demand generation. With all of these new technologies and demands, the modern data center has truly become a distributed node infrastructure.

So here is the real challenge: how do you control and manage it all? How do you control practically every aspect that is critical to data center functionality? Most of all, how do you do it on a distributed plane?

Data Center Abstraction

Data center abstraction is an emerging field where all physical and logical components are abstracted and inserted into a powerful management layer. This new model is sometimes referred to as the software-defined data center. However, today we’re focusing on the management layer. The data center of the future will be truly agnostic where all resources become presented to a powerful management layer, which can then be controlled anywhere and anytime. This is the data center operating system.

One example of this data center operating model is provided by IO and its IO.OS environment, which helps control many of the absolutely critical components, from chip to chiller. The great part is that this DCOS layer has visibility into every critical aspect that a data center has to present.

This conversation takes us far beyond standard DCIM. We’re now looking at an open data center and open cloud architecture. So what makes up a solid data center operating system? What has IO.OS done to really help organizations regain control of their global data center footprint? Let’s examine what it takes to create a powerful DCOS framework.

The Control Layer: Let’s start at the top. The control panel and management layer of a DCOS platform incorporates an easy-to-follow yet very granular interface. There is direct visibility into everything that is residing within your data center. This includes energy management, controlling QoS, monitoring the current state of all VMs, and even creating sensor setpoints throughout your data center. Here’s the important piece: you will have visibility into you entire data center and cloud environment.

The Integration Layer: What if you have outside cloud instances? What if you have big data engines? What if you need visibility into some resources that are “outside” of your data center? A big piece of the DCOS framework revolves around creating a more open infrastructure. Whether this means integrating with a big data engine or applying key APIs to allow communications between applications and resources, your DCOS model must help extend your infrastructure. This means incorporating front and back-end resources and pushing critical data to the control layer of the DCOS. Now, imagine integrating with automation, logging, and other critical systems that were once islands within a data center.

The Proactive Layer: Imagine changing the temperature settings within a specific section or rack within your data center based directly on pre-set thresholds. How about modifying environmental and resource variables based on compliance with set application requirements? Automation and maintaining an intuitive data center infrastructure are key pieces in the DCOS model. Proactively staying ahead of service impacts across your entire data center (both logical and physical) allows administrators to focus on efficiencies, not fires. You basically have an intelligent proactive DCOS layer which allows you to make granular system adjustments on the fly. Ultimately, this allows your data center to change at the speed of business. Which, in today’s world, is pretty fast. Here’s the reality: it’s not even about “real-time” any more. A solid data center operating system provides the proactive element, as well as an intuitive structure, where metrics can be gathered and system changes can be planned around existing and future demands.

The Visual Layer: Immerse yourself inside of your data center with visuals, metrics and a point-of-view that’s never really been experienced before. Platforms like IO.OS leverage both physical and virtual sensors within the IT and support spaces to provide detailed, intuitive, measurement data for power delivery, energy recovery, IT equipment performance, environmental sub-systems, and even the applications running in a particular rack. Do you have a converged infrastructure? Maybe you have a highly distributed data center plane with numerous nodes. Or, maybe you’ve got a heterogeneous model containing a number of resources. Abstract it all, present it to your DCOS, and visualize your open-structure data center model. Capacity planning, resource management, workload optimization, and even user controls are now introduced directly into the data center management plane.

The Security Layer: It’s not just security. It’s about all-encompassing, data center security revolving around physical and logical components. A DCOS model must incorporate all critical resources into it security structure. This means creating security-based roles, complying with regulations and governance, creating secure data repositories, and continuously monitoring the entire data center against physical and logical advanced persistent threats (APTs). This can be DDoS or a physical breach of a data center. The point is that you have continuous, proactive, security management throughout your entire distributed data center plane.

The Virtual Data Center Layer: Find trends, analyze metrics, see critical reports, and utilize multiple layers to review everything that makes up your data center. All of this can be done from your fingertips regardless of the location. A virtualization-aware data center running DCOS immediately presents a number of new capabilities to support emerging technologies and business needs. By having control at your fingertips, literally, the time it takes for you to respond – or make intelligent decisions – is cut dramatically. IO.OS Mobility, for example, leverages the power of HTML5 and provides you with a consolidated view of your entire infrastructure, current status, measurements, historical trend graphs, warnings and alarm-filtering capabilities. Now, you have direct data center control and proactive visibility securely from any device, anytime and anywhere.

The data center model is changing rapidly. There is a drive to support more users, more technologies and a lot more cloud services. The only way to keep up is to evolve your data center and the entire control model. Resource utilization and optimizing your overall infrastructure not only create cost savings, but help keep your users and organization running optimally.

As you look at your data center model, make sure to understand just how well it’s being managed and how much intelligent visibility you really have. Having a proactive and intuitive platform control can truly allow the administrator to take their data center to the next level. Whether it’s dynamic resource provisioning, or catching problems before they arise, the future data center platform will have a powerful logical layer where both physical and logical resources are management. The idea is to create a truly scalable data center environment capable of meeting your user and business demands both today and in the future.

source : www.datacenterknowledge.com/