The proliferation of cloud has created new types of service models aimed at very direct solutions.

The proliferation of cloud has created new types of service models aimed at very direct solutions.

Five Great Ways to Optimize Your Data Center

2 comments

dreamstime-server-install

How can data center managers get more from their data center model? Bill Kleyman looks at five areas where you can optimize.

It’s a new year and once again data center professionals are looking at new ways to make their infrastructure even more efficient. At this point, the modern data center has become the central hub for almost all technologies. Whether it’s content delivery, data distribution, or just overall business support, the data center is absolutely critical for today’s business to thrive.

The proliferation of cloud computing and IT consumerization has all driven a pretty big boom in data center demand. Because of this, data center administrators are constantly trying to find ways to improve performance, enhance infrastructure density, and increase multi-tenancy capabilities. As more converged systems make their way into the data center, infrastructure optimization will be critical to maintain a high level of service.

So what are some great ways that you can optimize your data center? How can data center managers get more from their data center model? Let’s look at a few ways.

  • Software-defined technologies (SDx) and virtualization. The capabilities of the modern hypervisor are light-years ahead of what was possible just a few years ago. We are now able to directly integrate with key APIs to reduce resource hops and drastically improve workload performance. Even more so are the new technologies around software-defined networking, storage, security – and the data center! Our ability to abstract so many critical layers is a huge reason why we can build in many more data center efficiencies. New levels of network virtualization allow administrators to create vast network environments capable of spanning numerous data center across many countries. We are no longer bound by hardware requirements. Software-defined technologies are capable of delivering many kinds of data center efficiencies at numerous different levels.
  • Utilize cloud computing. There is a very real reason that hybrid cloud platforms are gaining so much popularity. One great way to increase data center efficiency is to integrate a cloud model. Many organizations are expanding their data center to the cloud for many different reasons. Competition in the data center space has created new offerings, better pricing, and a lot more available resources. This means that creating intelligent links between private and public cloud environments has become much easier. Data center control can now span many different cloud models where administrators don’t actually have to worry about physical infrastructure. Rather, they concern themselves with the workloads running directly on top. This type of infrastructure optimization can see a part of an environment remain in-house – while other piece is directly extended into the cloud. The beauty here is that by using cloud automation, software-defined technologies, and improved distributed infrastructure management – the cloud can be a powerful tool to optimize your data center.
  • Optimize power usage. There’s a good reason that data centers are being built in some interesting climates. Now that bandwidth is much better and data centers can reach further – organizations are warming to the idea of deploying data centers in some pretty remote locations. For example, Iceland’s power grid draws entirely upon hydro-electric and geothermal power, ensuring a totally “green” power supply. Plus, power in Iceland is available at 4.5 cents per kilowatt hour, with lower pricing available for bulk purchases. That’s an attractive pitch when you consider the potential for fluctuations in power pricing in countries like the U.S., Britain and Germany. Outside of building a new data center, there are other direct ways to optimize power utilizing. Examine your existing environment and inspect lost power areas. For example, what wattage are your servers taking when sitting idle? Do you have good power management systems which can dynamically provision/de-provision resources as needed? Directly in line with power usage are optimizations around the overall infrastructure as well. A well monitored and maintained data center with good power regulation will see direct PUE optimizations.
  • Improve cooling and other data center environmental variables. Data center environmentals are never easy to control. Cold or hot spots can create energy drains and poor airflow circulation. In creating data center environmental optimizations, there are several key factors to examine. This includes rack placement, server density, flooring (slab vs raised), aisle control, and much more. Furthermore, using predicative trend analysis systems helps with both current and potential future demands. Optimizing a data center also means reducing operational costs. The less money you need to spend running the environment the more you can have to spend on critical infrastructure improvements.
  • Create management transparency. The vast amount of data center distribution and the increased utilization of cloud computing has created a new challenge for the modern data center. As one of the most important pieces around data center optimization, clear management is absolutely critical. We now have concepts around data center virtualization and even the new data center operating system (DCOS). These management platforms take DCIM, automation, cloud control, and many other data center services to a whole new level. Basically, these new management systems place all critical components under one management environment. Now, we have much more data distribution and vast data centers sharing resources over the wire. One of the best ways to optimize your data center is by knowing exactly what is being run on physical systems. From there, you’re able to make proactive decisions around resource allocation and where there needs to be improvements. The future data center will only continue to support more users. All of this will continue to drive the need for optimization-ready data center management platforms.

As cloud computing and other such technologies become more prevalent, core data center technologies will need to evolve to support all of these new demands. Just like anything in technology, the data center infrastructure will evolve. Along with that will come new ways to keep your environment up and running optimally.

Remember, there are a lot more than five ways to create a better data center. Always take the end-user into consideration and create redundancy whenever possible. One major aspect of data center optimization is making sure your environment stays up in the first place. With that in mind, always take a proactive approach to data center planning and deployments. Building in infrastructure redundancy and DR planning are also critical pieces around maintain an optimally run data center environment. In optimizing your data center, you develop  a more resilient platform that will save you money and improve the overall performance of your business.

About the Author

Bill Kleyman is a veteran, enthusiastic technologist with experience in data center design, management and deployment. His architecture work includes virtualization and cloud deployments as well as business network design and implementation. Currently, Bill works as the National Director of Strategy and Innovation at MTM Technologies, a Stamford, CT based consulting firm.

Add Your Comments

  • (will not be published)

2 Comments

  1. Software Defined 'x' (storage & data center) and hybrid cloud architectures are high-value areas to optimize the modern data center. These technologies and architectures will also transition your data center from a reactive disaster recovery-based environment into a highly available environment.

  2. I would think that the overall goal would be to decrease IT payments for both capital costs and ongoing functional expenses Thanks for the interesting read Bill.